Download - Transition from GSM to UMTS

Lucent Technologies - Proprietary This document contains proprietary information

of Lucent Technologies and is not to be disclosed or used except in accordance with applicable agreements

Copyright 2000 Lucent Technologies Unpublished and Not for Publication

All Rights Reserved

GSM to UMTS Transition RF Engineering Guideline EG: GSMUTR

401-380-373 Issue 1.1 July 2000

Copyright ©2000 by Lucent Technologies. All Rights Reserved.

This material is protected by the copyright laws of the United States and other countries. It may not be reproduced, distributed, or altered in any fashion by any entity (either internal or external to Lucent Technologies), except in accordance with applicable agreements, contracts, or licensing, without the express written consent of the Customer Training and Information Products organisation and the business management owner of the material.

For permission to reproduce or distribute, please contact:

The Manager, RF Systems & Capacity Engineering Group 01793 883275 (domestic) (44) 1793 883275 (international)

Notice

Every effort was made to ensure that the information in this information product was complete and accurate at the time of printing. However, information is subject to change.

Contents

Issue 1.1 - July 2000 Lucent Technologies – Proprietary

See Notice on first page

iii

1. ABOUT THIS DOCUMENT 1

1.1. Purpose 1

1.2. Contents 2

1.3. Scope 3

1.4. Audience 3

2. INTRODUCTION TO THE UMTS AIR INTERFACE 5

2.1. Background 5

Frequency allocation 5 Standards 5 UMTS summary 7

2.2. Band plan 8

Satellite allocation 8 Terrestrial allocation 8

2.3. UTRAN air interface attributes 11

2.4. Channel mapping on the air interface 13

Access stratum 13 Logical channels 14 Transport channels 15 Physical channels 17

GSM to UMTS Transition – RF Engineering Guideline

iv Lucent Technologies – Proprietary


Issue 1.1 - July 2000

2.5. Channel spreading, coding and modulation 23

Uplink 23 Downlink 27 Synchronisation codes 29

2.6. Physical channel frame structure 29

Uplink time slot fields 29 Downlink time slot fields 31

2.7. Speech coding 34

Transcoder Free Operation 35 GSM Full Rate codec 36 GSM Half Rate codec 36 GSM Enhanced Full Rate codec 36 Adaptive Multi-Rate codec types 37

2.8. Codec speech quality 44

Fixed rate codecs 44 Disadvantages 45 Adaptive Multi Rate codecs 46

3. MOBILE HANDOVER 49

3.1. Handover types 49

3.2. Cell sets 51

3.3. Preparation for UTRAN to UTRAN handover 52

3.4. Preparation for UTRAN to GSM handover 53

Silence Duration parameters 54

3.5. Handover execution 58

UTRAN soft handover 58 UTRAN to GSM handover 58

3.6. GSM to UTRAN handover 59

GSM to UMTS Transition - RF Engineering Guideline



v

4. SUBSCRIBER SERVICES 61

4.1. Coding and interleaving for subscriber services 62

4.2. Services multiplexing 62

4.3. Rate matching 63

Uplink 63 Downlink 63

4.4. Control channel coding and interleaving 64

Dedicated Control Channel 64 Downlink Common Control Channels 64

4.5. Channel mapping examples 64

8kbs-1 bearer - speech 64 144kbs-1 bearer - data 65 384kbs-1 bearer - data 65 480kbs-1 bearer - data 66

5. LUCENT EQUIPMENT 67

5.1. Node-B (BTS) 68

Distributed Milli-cell 69 Microcell (Ultra-small cell) 70 Milli-cell 70 BTS traffic capacity 75 BTS further reference 77

5.2. BTS Antennas 78

No existing network 78 Existing single band network 78 Existing dual band network 80 Dual band and tri band GSM/UTRAN diplexers 84 Broadband power divider 87 Broadband indoor antennas 87 Antenna feeder 88 Masthead amplifier 89


vi Lucent Technologies – Proprietary



Repeater 90 Active/smart/adaptive antennas 91 General antenna comments 92 Further reference – antennas and BTS accessories 92

5.3. Radio Network Controller 93

RNC characteristics 96 RNC further reference 99

5.4. Radio Resource Control software 99

Radio resource allocation 99 Radio Resource Allocation functions 104 Radio Access Bearer parameters 107 Physical channel related parameters 107 Reverse outer loop power control 110 Further reference – power control system 111 Power control parameters 111 Radio Resource Control software – further reference 112

5.5. Handover 112

Measurement reporting 113 Measurement messages 116 Measurement performance 117 Soft (and softer) handover algorithm 117 Hard handover algorithm 122 UTRAN – GSM handover algorithm 122 UTRAN – GSM GPRS handover algorithm 122 Handover control software – further reference 122

5.6. Lucent equipment capacity 123

Hardware (Lucent Network Release 1.0) 123 Software (Lucent Release 0.1) 123

6. RF NETWORK COVERAGE AND CAPACITY DESIGN 125

6.1. Frequency planning 126

Frequency planning criteria 126 Example UTRAN band assignment – United Kingdom 127




vii

6.2. Code assignment parameters 129

Primary Synchronisation Channel Code 129 Secondary Synchronisation Channel Code 129 Scrambling code 130 Code assignment summary 130

6.3. Air interface link power budget 131

Background 131 Effect of coding scheme on link power budget 131 Link power budget elements 133 Example link power budgets 145 Margins for fading and building attenuation 148

6.4. Estimating coverage and traffic capacity 149

Fixed cell loading 150 Adaptive cell loading 152 Estimating base station numbers 153 Land-use classification 159 UMTS Dimensioning tool – further reference 159

6.5. Airpro coverage and traffic distribution prediction software 160

Introduction 160 Principal features 161 Airpro default values 173

6.6. BTS and antenna settings optimisation software 174

Introduction 174 Applications 175 Cellular radio standards 175 Optimising strategy 175 RF model 176 Results 177 Comparison trials 177

6.7. Inter-system boundary 178

6.8. Further reference 178


viii Lucent Technologies – Proprietary



7. RF NETWORK PRACTICAL IMPLEMENTATION 179

7.1. Use of existing sites 179

7.2. EMC at existing GSM sites 180

Transmitter isolation 180 Receiver performance degradation 181 Inter-system isolation criteria 183 Estimating inter-system isolation 183 Antenna coupling 184 Separate UTRAN and GSM antenna systems 185 Single UTRAN and GSM antenna system 190 UTRAN BTS spurious emissions 191 Worked example of co-siting GSM and UTRAN 193

7.3. Use of repeaters 198

Introduction 198 Design 199 Uniform range extension 200 Cascaded range extension 200 Repeater gain and composite noise factor 200 Donor cell shrinkage 201 Median repeater link budget calculation 202 Median repeater link budget adjustment 203 Repeater donor cell antennas 203 Summary 204

7.4. Use of microcells 205

Embedded and non-embedded microcells 206 Microcell problems 207 Co-channel macrocells and microcells 207 Macrocells and microcells on different channels 208 Dual layer UTRAN 208

7.5. Masthead amplifiers 209

Without masthead amplifier 210 With masthead amplifier 212 Summary 214




ix

7.6. Practical antenna considerations 214

Intermodulation products 214 Front-to-back ratio 215 Variable electrical down-tilt 215

7.7. Transmit diversity 217

UTRAN implementation 218 Effect of UTRAN implementation 219

7.8. Coverage areas for different services 219

7.9. Survey test equipment 220

7.10. Site selection and design optimisation 222

RF network engineering requirements 222

8. DEPLOYMENT – WORKED EXAMPLE 225

8.1. Scenario 225

Coverage area 226 Existing base station sites 226

8.2. Background 229

8.3. Estimating coverage and capacity 230

Coverage model 230 Traffic model 232 System model 233 Analysis results 237

8.4. Coverage prediction with Airpro (CE 5) 241

Existing 900MHz coverage 241 Existing 1800MHz coverage 244 Predicted UTRAN coverage 247 Results 260 Conclusion 260

8.5. EMC when using a common antenna 261


x Lucent Technologies – Proprietary



900MHz receiver - in-band power 261 1800MHz receiver - in-band power 262 UTRAN receiver - in-band power 262 GSM 900MHz receiver – out-of-band power 263 GSM 1800MHz receiver – out-of-band power 264 UTRAN receiver – out-of-band power 265 Conclusion 266

APPENDIX A SILENCE DURATION PARAMETERS 267

ACRONYMS 269


About this Document

1

Issue 1.1 - July 2000 Lucent Technologies – Proprietary See Notice on first page

1

1. About this Document

1.1. Purpose

This document describes the principal Radio Frequency (RF) engineering implications when cellular radio networks make the transition from the GSM air interface standard to the Universal Mobile Telecommunications System (UMTS) standard.

In particular it discusses the following areas:

• Air interface

• Handover

• Subscriber services

• Lucent equipment

• RF network coverage and capacity

• RF network practical implementation

• Future developments

Owing to the wide range of environments in which UMTS systems will be deployed, and the differing operational and commercial priorities of network operators, the choice of implementation technique will be territory and customer specific. As a result, this document can only provide an outline guide to the main issues involved in the network design and deployment


2 Lucent Technologies – Proprietary See Notice on first page


process. It is not a substitute for the detailed analysis and design that will be required on a site by site basis.

1.2. Contents

• Chapter 2 - Air Interface

This chapter provides a brief introduction to the main characteristics of the radio link, including the function of transport channels and their mapping to the physical channels transmitted over the air.

It describes the concept of channel codes, which are used to implement Code Division Multiple Access (CDMA) to the shared RF channel (frequency), and identifies the main codes applied to the air interface. It also describes spreading factor (and associated de-spreading gain) and bearer data rate concepts.

As this document deals with the physical layer of the link, protocol aspects are not covered.

• Chapter 3 - Handover

This chapter describes the main UMTS handover types, with particular emphasis on inter system handover with a neighbouring or overlaid GSM network.

• Chapter 4 - Subscriber Services

This chapter describes the basic subscriber services, from the viewpoint of their associated bearer data rates and spreading factors.

Later chapters use these concepts when looking at the effect of the anticipated mix of services on RF network capacity and design.

• Chapter 5 - Lucent Equipment

This chapter describes the network elements that form the base station system. It describes the main attributes of Lucent’s equipment, and equipment capacity and interdependencies as they relate to RF network capacity and design.

• Chapter 6 - RF Network Coverage and Capacity Design

This chapter describes the network design process from the RF perspective, including radio link power budget calculation, estimated coverage techniques, and traffic modelling. It also describes Lucent software tools and techniques for network design.

• Chapter 7 - RF Network Practical Implementation

This chapter discusses some of the practical issues involved in upgrading an existing network from GSM to UMTS.



3

Areas considered include inter system interference, equipment that can be shared between systems, and the extent to which a UMTS service can be provided from an existing infrastructure of GSM base station sites.

• Chapter 8 - Deployment Worked Example

This chapter describes a worked example of UMTS deployment. Where possible, the data used is based on actual traffic projections and infrastructure configurations provided by existing network operators.

• Appendix A - Silence Duration Parameters

Appendix A reproduces an extract from ETSI UMTS xx.15 v1.0.0 1999-02 regarding the “Definition and Setting of Silence Duration Parameters”.

1.3. Scope

This document covers the radio engineering implications of the transition from the use of the GSM air interface to that of UMTS, for the provision of cellular radio services. It is primarily intended to cover circumstances where existing GSM network operators wish to upgrade their network, in whole or part, to provide a UMTS service.

In view of the range of this subject, many topics can only be covered in outline. However, sufficient information is included to identify the major issues involved and to identify areas for further planning and investigation.

This document describes the features and facilities available in Lucent Network Release 1.0 software.

1.4. Audience

This document concentrates on the World Radio Conference (WRC) Region 1 (Europe, Middle-East and Africa) as this is where UMTS networks will be first established. However, many of the subjects covered will be relevant to other WRC regions.

It is intended for use by the following groups:

• Engineers undertaking first-time design of GSM to UMTS upgrades

• Technicians who want an appreciation of UMTS design considerations

• Sales or support staff

• Managers or administrative staff who want a high-level understanding of UMTS design processes




A working knowledge of the following topics will be useful:

• GSM air interface (such as power budget and the basic structure of control and traffic channels)

• CDMA concepts


Introduction to the UMTS Air Interface 2


5

2. Introduction to the UMTS Air Interface

This chapter describes the main features of the UMTS radio link. It concentrates on the system aspects that have most impact on the design and implementation of a customer network, rather than on the design of individual network elements.

2.1. Background

Frequency allocation

The World Radio Conference (WRC) has allocated a frequency band around 2GHz for use by Public Land Mobile Networks (PLMNs) to support third generation public mobile phone and data services. These are expected gradually to replace the existing second generation networks (largely based on GSM and IS95).

The International Telecommunications Union – Radio (ITU-R) has overall responsibility for defining the third generation system, known as the International Mobile Telecommunications 2000 (IMT-2000).

The IMT-2000 system will be an integrated system that allows terminals to access both satellite and terrestrial based stations. Bands 1885MHz to 2025MHz and 2110MHz to 2200MHz are allocated to the terrestrial component. Bands 1980MHz to 2021MHz and 2170MHz to 2200MHz are allocated to the satellite component.

Standards

The ITU-R has produced high-level documents covering the performance, service type, and inter-working requirements for IMT-2000.




Various international standards bodies such as the European Telecommunications Standards Institute (ETSI) are responsible for the detailed technical specifications of the equipment required to provide an IMT-2000 compatible service. A number of different standards are likely to emerge; but they are expected to have sufficient inter-working capability to allow an integrated IMT-2000 service for subscribers.

IMT-2000 networks will support five interface standards:

• IMT-DS Frequency Division Duplex (FDD) version of UMTS (discussed in this document)

• IMT-MC US CDMA 2000 standard

• IMT-TC Time Division Duplex (TDD) version of UMTS

• IMT-SC GSM EDGE (IS-136) standard

• IMT-FT DECT standard

The four Technical Specification Groups (TSGs) of the ETSI-supported 3rd Generation Partnership Project (3GPP) have approved the detailed specification parts of their submission to the ITU-R for the IMT-2000 radio interface standard. This is a terrestrial radio interface specification known as the Universal Terrestrial Radio Access Network (UTRAN). The UTRAN is based on a Wide-band Code Division Multiple Access (WCDMA) air interface.

The ITU-R intend to approve the detailed specification of the IMT-2000 radio interface in May 2000, based on submissions received from the international standards bodies.



7

UMTS summary

A UMTS network can consist of one or more access networks, using different radio access systems, linked to the same core network. Together they form a single UMTS network. The term UMTS covers all the network elements in both the access network and the core network.

The UTRAN specified by ETSI is one such UMTS access network. It supports wide area terrestrial mobile telecommunications services, using the Frequency Division Duplex (FDD) IMT-2000 bands. This document concentrates primarily on the ETSI FDD UTRAN.

The UTRAN comprises the network elements that correspond to the Base Station Subsystem (BSS) in a GSM network:

• Base Transceiver Station (BTS)

• Base Station Controller (BSC)

In the UTRAN, the equivalent of the GSM BTS is referred to as either the BTS or the ‘Node B’. The equivalent of the GSM BSC is referred to as the Radio Network Controller (RNC). The UTRAN BTS and RNC together form the Radio Network System (RNS).

Owing to differences in the radio standards currently used, and consequently those used to provide a IMT-2000 service, this document concentrates on WRC Region 1, where the UTRAN RNS will be used as the UMTS access network.




2.2. Band plan

The WRC of 1992 allocated 230MHz of the 2GHz spectrum to IMT-2000 services with the intention of providing a uniform band plan for all three WRC regions.

The allocated spectrum consists of two blocks:

• 140MHz for TDD and FDD uplinks from 1885 to 2025MHz

• 90MHz for FDD downlinks from 2110 to 2200MHz

Satellite allocation

Within each block, two 30MHz sub-blocks with a 190MHz duplex separation are allocated to satellite operation:

• 1980MHz to 2010MHz for uplinks

• 2170MHz to 2200MHz for downlinks

Terrestrial allocation

FDD systems

• 1920 to 1980MHz Uplink Mobile Transmit (60MHz band, duplex separation 190MHz)

• 2110 to 2170MHz Downlink Base Transmit (60MHz band, duplex spacing 190MHz)

TDD systems

• 1885 to 1920MHz (35MHz band)

• 2010 to 2025MHz (15MHz band)



9

The IMT-2000 band plan and that of the existing major PLMN systems is illustrated in the following figure, according to their primary geographic area of use.

Figure 1 IMT-2000 band plan compared to existing PLMN systems

In many areas, part of the IMT-2000 band is already in use by other PLMN systems. The lower TDD IMT-2000 band will have to coexist with the Digital European/Enhanced Cordless Telephone/Telecommunications (DECT) and Personal Handy-phone System (PHS) systems.

DECT

DECT is widely used in the EU, Africa, Asia, and Caribbean to provide cordless telephony, wireless Private Branch Exchanges (PBX), and Wireless Local Loop (WLL) services. DECT uses dynamic channel allocation, and except where used for WLL applications, is usually deployed on a self-provision basis. This means there is limited scope for frequency co-ordination with a wide area IMT-2000 network. The output power of most DECT systems is limited to 250mW peak, 10mW average. For indoor systems, interference to outdoor IMT-2000 systems is further reduced by building attenuation.




PHS

PHS was designed in Japan. In recent years subscriber numbers have declined sharply, and several networks have closed. PHS can be considered as functionally similar to DECT. It is rarely used outside SE Asia, Australia and New Zealand, owing to problems arising from its partial use of DECT frequencies. Similar problems to those found with DECT are likely to be experienced when coexistence with TDD IMT-2000 is attempted.

PCS

The Personal Communication System (PCS) 1900 is currently being deployed as a second generation cellular system in the Americas. Some areas are adopting a CDMA air interface based on the IS95 standard (J-STD-008), and others the GSM air interface. Whichever air interface is used, there is a significant overlap with the IMT-2000 TDD and FDD uplink band.

In view of this, the American Federal Communications Commission (FCC) has allocated a different band for third generation cellular systems. The FCC has adopted licensing and service rules governing operation for 30MHz of the 700MHz spectrum auctioned in the spring of 2000.

It has established two license bands, one of 20MHz (two paired 10MHz bands) and one of 10MHz (two paired 5MHz bands) that can be used for advanced wireless services, including third generation broadband wireless access. Spectrum is made up in part from 746-764MHz and 776-794MHz (TV channels 60-62 and 64-66).

The FCC will auction the licenses in six “Economic Area Groupings” across America and will allow interested parties to bid for both license bands in one area. The remaining 6MHz will be used as guard band comprising 4MHz (two paired 2MHz bands) and 2MHz (two paired 1MHz bands). FCC plans to invite comments on technical and operational issues regarding these frequencies.

In many parts of America, third generation cellular services may not be able to use these frequencies until 2007 owing to the use of television channels 60 to 69.

GSM 1800 / DCS 1800

There is no direct overlap between the GSM 1800 (also known as DCS 1800) downlink band, and the IMT-2000 TDD and FDD uplink bands. However, in view of the possibility of co-siting base station equipment, and sharing antenna systems, particular care is required to ensure adequate isolation in order to avoid interference. This topic is covered in Chapter 7 RF Network Practical Implementation.



11

2.3. UTRAN air interface attributes

This document concentrates on the ETSI UTRAN system. There are two versions of UTRAN, one that uses TDD mode and one that uses FDD mode. This document deals mainly with the terrestrial FDD version of UTRAN. This is the standard that forms the basis for the wide area deployment of 3rd generation PLMNs.




The principal air interface attributes of the FDD and the TDD UTRANs are:

Feature Terrestrial FDD UTRAN TDD UTRAN

Uplink frequency /MHz 1920-1980 1885-1920 & 2010-2025

Downlink frequency /MHz 2110-2170 1885-1920 & 2010-2025

Channel bandwidth /MHz 5 5

Carrier raster /MHz 0.2 0.2

Duplex separation /MHz 130 (min) but variable N/A

Frequency stability /ppm 0.05 base, 0.1 mobile 0.05 base, 0.1 mobile

Chip rate /Mcs-1 3.84 3.84

Spreading factor 4 to 256 1 to 16

BSS separation codes Gold code 10ms, 38400 chips, length 241 -1

Scrambling code of length 16 chips

Modulation QPSK QPSK

Modulation filter Root raised cosine roll-off factor 0.22

Root raised cosine roll-off factor 0.22

Power control Fast closed loop & slow quality loop

Open loop & slow closed loop

Power control steps /dB 0.25 to 1.5 0.25 to 1.5

Minimum output power /dBm -50 -50

Power control dynamic range /dB 80 uplink, 30 downlink 80 uplink, 30 downlink

Power control sample rate /kHz 1.6

Channel coding & interleaving for services tolerating BER > 10-6

Convolutional, rate 1/2 or 1/3

Convolutional, rate 1/2 or 1/3

Channel coding & interleaving for services requiring BER < 10-6

Turbo coding Turbo coding

Modulation symbol rate /M symbol s-1 0.016 to 1.024 0.256 to 4.096

Radio super-frame length /ms 720 240

Radio frame length /ms 10 10

Radio slot per frame 15 of 666.7us 2,560 chips

15 of 666.7us 2,560 chips

Channel allocation Network controlled Dynamic



13

Feature Terrestrial FDD UTRAN TDD UTRAN

Handover control Mobile assisted measurement of signal

level & timing. GSM measurements

supported.

Probing for ODMA

Base output power class 1 macro, 2 micro, 3 pico

Mobile output power /dBm [class number]

+21 [4] +33 [1], +27 [2], +24 [3],

+10 [5], 0 [6]

+21 [4] +33 [1], +27 [2], +24 [3], +10

[5], 0 [6]

Table 1 Air interface attributes

2.4. Channel mapping on the air interface

This section summarises the mapping of logical channels and transport channels to the physical channels transmitted over the air interface.

Access stratum

The access stratum on the air interface is divided into three layers:

• Layer 1

• Layer 2

• Layer 3

Layer 1

Layer 1 is the physical layer. Signalling and traffic data is borne on the air interface by physical channels. The physical channels are defined by code set and frequency in FDD mode and by code, timeslot, and frequency in TDD mode.

Layer 2

Layer 2 is divided into two sub-layers:

• Medium Access Control (MAC) layer (lower layer).

The MAC layer is responsible for the random access procedures, physical link control, error protection, ciphering, multiplexing, and channel mapping to the physical layer (Layer 1)

• Radio Link Control (RLC) layer (upper layer).




The RLC layer is responsible for logical link control, and acknowledgement /unacknowledgement of data transfer

Layer 3

Layer 3 is the Radio Resource Control (RRC) layer. The RRC layer is responsible for coordination and control of bearers, monitoring processes, power control, measurement reporting, paging, and broadcast control functions.

In order to define a process for each different type of information, sets of logical channels are mapped onto transport channels, and ultimately physical channels are defined.

Logical channels are defined between the RLC and the MAC. Transport channels are defined between the MAC and the physical layer (Layer 1).

Logical channels

The following logical channels are used to transfer signalling information:

• Broadcast Control Channel (BCCH) -downlink

The BCCH is a downlink broadcast channel which carries system information. There are two types: BCCH-Constant (BCCH-C) and BCCH-Variable (BCCH-V), the data on which may be constantly updated.

• Paging Control Channel (PCCH) - downlink

The PCCH is a downlink channel which carries paging messages. It is used when the network does not know the location cell of the mobile or the mobile is in ‘sleep mode’.

• Common Control Channel (CCCH)

• The CCCH is a bi-directional channel which carries data when the mobile has no RRC connection to the network.

• Dedicated Control Channel (DCCH)

The DCCH is a bi-directional channel which carries point-to-point dedicated control data between the network and a mobile. It is used when a dedicated connection has been established through RRC connection set up procedures.

• ODMA Common Control Channel (OCCCH)

The OCCCH is a bi-directional channel which carries control data directly between mobiles. It is used when the mobile has no RRC connection with the network.

• ODMA Dedicated Control Channel (ODCCH)



15

The ODCCH is a bi-directional point-to-point channel which carries dedicated control data directly between mobiles. It is used when a dedicated connection has been established through the RRC connection set-up procedures.

• Shared Channel Control Channels (SHCCH) – TDD mode only

The SHCCH is used in TDD mode only. It is a bi-directional channel which carries control data for the uplink and downlink shared channels.

• Synchronisation Control Channel (SCCH) – TDD mode only

The SCCH is used in TDD mode only. It is a downlink channel which carries the location and structure of the BCCH.

The following logical channels are used to transfer user data:

• Dedicated Traffic Channel (DTCH)

The DTCH is a bi-directional dedicated point-to-point channel which carries user data between the network and the mobile.

• ODMA Dedicated Traffic Channel (ODTCH)

The ODTCH is a dedicated point-to-point channel which carries user data directly between mobiles, used as a relay link.

• Common Traffic Channel (CTCH)

The CTCH is a uni-directional point-to-multipoint channel which carries user data for a specified group of mobiles.

Transport channels

The information is transferred from the MAC layer and mapped onto physical channels via a set of transport channels.

There are two types of transport channel:

• Common transport channels

• Dedicated transport channels

Common transport channels

• Broadcast Channel (BCH) – downlink

The BCH is transmitted from the base station to all mobiles in the cell coverage area and broadcasts system configuration information.




• Paging Channel (PCH) – downlink

The PCH is also broadcast from the base station to all mobiles in the cell coverage area. It pages mobiles when they are in idle mode and only their Location Area, not their specific cell, is known.

• Forward Access Channel (FACH) – downlink

The FACH is transmitted from the base station to mobiles and carries relatively small amounts of control data and very short packets of user data, without the use of power control.

• Random Access Channel (RACH) – uplink

The RACH is a contention-based uplink channel used for initial access, non-real time dedicated control or very short packets of traffic data.

• ONMA Random Access Channel (ORACH)

The ORACH performs a similar function to the RACH when a relay link is used.

• Common Packet Channel (CPCH) – FDD mode only

The CPCH is a contention-based channel carrying bursty traffic data in shared mode using fast power control.

• Downlink Shared Channel (DSCH)

The DSCH is a downlink channel shared between several mobiles for carrying control or traffic data.

• DSCH Control Channel

The DSCH Control Channel is a downlink channel used in conjunction with the DSCH for signalling of DSCH resource allocation.

• Broadcast Channel (BCH)

The BCH is a downlink broadcast channel carrying system information for the whole cell.

• Synchronisation Channel (SCH) – TDD mode only

The SCH is a downlink TDD mode channel carrying synchronisation data for the whole cell. Note: This SCH has no connection with the physical channel SCH used in FDD mode.

• Uplink Shared Channel (USCH) – TDD mode only

The USCH is an uplink TDD mode channel shared by several mobiles for carrying control or traffic data.



17

Dedicated transport channels

• Dedicated Channel (DCH) – duplex channel pair

For the duration of a call, one DCH is transmitted from the mobile to the base station and one DCH is transmitted from the base station to the mobile. They form a duplex circuit that can be used to carry a number of different types of user data (or logical channels).

• Fast Uplink Signalling Channel (FAUSCH)

The FAUSCH is an uplink channel used to allocate dedicated channels in conjunction with the FACH.

• ODMA Dedicated Channel (ODCH)

The ODCH is dedicated to one mobile when used for relay links.

With the exception of the FAUSCH, each transport channel has an associated transport format. This is defined by a combination of encoding, interleaving, bit rate, and mapping onto physical channels. Some transport channels may use variable formats.

Physical channels

The transport channels are mapped to the physical channels transmitted over the air interface. This mapping process is flexible and for some logical channels there are several options depending on their function and the type of information being transferred.

Different transport and physical channels are used in the uplink and downlink directions. Depending on the type and data rate of the transport channel, coding and multiplexing/ demultiplexing may be applied prior to the data being carried by the physical channel.

The channels carrying broadcast information are directly mapped. That is, BCCH to BCH and PCCH to PCH.

Mapping for the other control and traffic channels is more flexible. For example, the downlink DCCH can be mapped onto either the FACH or the DSCH depending on the information requirements. In the uplink, the DCCH may take information from the CPCH, FAUSCH, RACH, USCH or DCH. The logical channel DTCH has access to a similar range of transport channels. The CCCH can only use the RACH and FACH for bi-directional communication.

Downlink mapping of transport to physical channels

Three common control channels are used in the downlink: BCH, PCH and FACH. Each have a specific coding process and are mapped onto a physical channel that is specific to common control.

For DCH coding, multiplexing and mapping to physical channels is similar to that for the uplink.




If a mobile is in soft handover (establishes a connection with a new cell while still communicating with its current serving cell) the same data may be received by the mobile from multiple base stations, each using different physical channels.

Prior to DCH transmission in the physical layer, information is added at Layer 1. During a soft handover the physical layer information bits received from the target base stations will be different from those received from the serving base stations. Also, the composition of the CCTrCH may not be the same for all base stations involved in the soft handover.

Transmit Power Control (TPC) bits and Transport Format Combination Indicator (TFCI) bits are added to the DCH (as Layer 1 information) before it is transmitted over the physical channel. During soft handover, the TPC bits received from the neighbour base stations will differ from those received from the serving base station. The TFCI bits are identical.

Figure 2 Mapping of downlink transport and physical channels

OVSF Code 4

OVSF Code 3

Code (and mux/divide if CCTrCH exceeds capacity of one Physical Channel)



OVSF Code 1

OVSF Code 2

DPCH

Sec. CCPCH

Prim. CCPCH

TPC and TFCI added prior to transmission for UE power control



DCH

BCH

PCH

FACH

(comprisingDCCH,DTCH)

Base Station

Code

DPCH

DCH


OVSF Code n

DPCH

DCH


Code



19

Uplink mapping of transport to physical channels

Two common control channnels are used in the uplink: RACH and FAUSCH. Each one has specific coding mechanisms and is mapped onto a physical channel.

The DCH channels (DCCH and DTCH) are coded and multiplexed to form a Coded Composite Transport Channel (CCTrCH). Depending on its data rate the CCTrCH is then mapped onto one physical channel or demultiplexed onto several physical channels. The physical channel used for the CCTrCH carries data from the DCH transport channel only. The DCH data is not multiplexed with data from other transport channels.

Transmit Power Control (TPC) bits and Transport Format Combination Indicator (TFCI) bits are added to the DCH (as Layer 1 information) before it is transmitted over the physical channel.

Figure 3 Mapping of uplink transport and physical channels

Physical channel types

The physical channels are defined by their basic resource characteristics, in terms of code and frequency plans, modulation and transmission. A number of different physical channel types are used in both the uplink and downlink.

Downlink physical channels

Downlink physical channels are grouped into four types:

• Synchronisation Channel (SCH)

The SCH transmits the synchronisation codes used by the mobile to synchronise to a base station. It comprises a primary and a secondary channel, which are transmitted simultaneously.

OVSF Code 1

OVSF Code 2

OVSF Code 3

OVSF Code “n”

DPCH

DPCH

DPCH

DCH

DCH

DCH

RACH

(comprising DCCH and DTCH)

Base Station

Code

Divide -If CCTrCH exceedscapacity of onePhysical Channel

Code& Mux.

Code

Transmit Power Control (TPC)Transmit Format Combination Indicators (TFCI)added prior to transmission for BTS power control

PRACH




• Common Pilot Channel (CPICH)

The CPICH provides the phase reference for the downlink common channels, and is implemented as either a Primary or Secondary Pilot Channel.

• Common Control Physical Channel (CCPCH)

The CCPCH is the bearer for the Broadcast Channel (BCH), the Paging Channel (PCH), or the Forward Associated Control Channel (FACCH). If the CCPCH carries the BCH it is the Primary CCPCH, if it carries either the PCH or FACCH it is the Secondary CCPCH.

• Dedicated Physical Channel (DPCH)

The DPCH carries the downlink Dedicated Channel (DCH) transport channel together with Layer 1 data comprising the Channel Associated Pilot, Transmit Power Control (TPC) bits, and Transmit Format Combination Indicator (TFCI) bits. The DPCH can be considered to be formed from the Dedicated Physical Data Channel (DPDCH) carrying the Dedicated Channel (DCH) and the Dedicated Physical Control Channel (DPCCH) carrying the Layer 1 data, time multiplexed together to form the DPCH

The downlink physical channel arrangement is shown in the following diagram:



21

Figure 4 Downlink physical channels

Base Station

Dedicated PhysicalChannel (DPCH)

DTCHDCCH

(Dedicated Physical Control Channel (DPCCH))

(Channel Associated Pilot, TPC, TFCI)

(DCH Transport Channel)

(Dedicated Physical Data Channel (DPDCH))

Common ControlPhysical Channel(CCPCH)

Primary CCPCH

(Paging (PCH) and Forward Access (FACH) channels)

(BCH at 32 kb/s)

Secondary CCPCH

SynchronisationChannel (SCH) (Layer 1 Data - chip rate, framing, group of Golay codes)

(Sync. code, short Golay code, already chipped. BTS specific)

Primary SCH

Secondary SCH

Common PilotChannel (CPICH)

Primary CPICH

Secondary CPICH




Uplink physical channels

The uplink physical channels are:

• Physical Random Access Channel (PRACH)

The PRACH carries the Random Access Channel (RACH) and the FAUSCH.

• Dedicated Physical Channel (DPCH)

The DPCH carries the uplink Dedicated Channel (DCH) transport channel, together with Layer 1 data comprising the Channel Associated Pilot, Transmit Power Control (TPC) bits, Transmit Format Combination Indicator (TFCI), and Feedback Information (FBI) bits.

The DPCH can be considered to be formed from the Dedicated Physical Data Channel (DPDCH) carrying the Dedicated Channel (DCH) and the Dedicated Physical Control Channel (DPCCH) carrying the Layer 1 data. These are fed separately to the I & Q ports of the mobile’s QPSK modulator to form the DPCH.

The uplink physical channel arrangement is shown in the following diagram:

Figure 5 Uplink physical channels

Base Station

Physical Random Access Channel (PRACH)

Random AccessChannel (RACH)

Dedicated PhysicalChannel (DPCH)

DTCHDCCH

(Dedicated Physical Control Channel (DPCCH))

(Channel Associated Pilot, TPC, TFCI)

(DCH Transport Channel)

(Dedicated Physical Data Channel (DPDCH))



23

2.5. Channel spreading, coding and modulation

This section describes channel spreading, coding and modulation concepts in the uplink and downlink.

Uplink

Spreading Dedicated Physical Data and Control Channels

Dual-channel Binary Phase Shift Keying (BPSK) modulation is used to spread the Dedicated Physical Data Channels (DPDCH) and Dedicated Physical Control Channels (DPCCH). The uplink DPDCH and DPCCH are mapped to the I & Q branches of the modulator respectively.

The I & Q branches are spread to the chip rate with two Orthogonal Variable Spreading Factor (OVSF) codes and then scrambled using a mobile specific scrambling code cscramb.

This process is illustrated in the following diagram:

Figure 6 Uplink spreading and modulation

When more than one code is used for transmission, additional uplink DPDCHs may be transmitted on either the I or Q branches. For each branch, each additional DPDCH is assigned its own channel code. Uplink DPDCH channels on different branches may share a common channel code.

Spreading & Modulation for Uplink DPDCH/DPCCH

++

DPDCH

DPCCH

OrthogonalVariableSpreadingFactor (OVSF)Codes

OVSFCodes

Cscramb

cos ( t)ω

sin ( t)ω

Realp(t)

p(t)Imag

I

Q

I+jQ




Spreading Physical Random Access Channels

The message component of the Physical Random Access Channel (PRACH) is spread and modulated in a similar manner to that used for the uplink dedicated physical channels. The uplink DPDCH is replaced with the data part and the DPCCH is replaced with the control part. The scrambling code used for the message component is chosen based on the base station-specific preamble code in use.

Channel codes

Orthogonal Variable Spreading Factor (OVSF) codes are used as channel codes, which ensure that a number of mobiles can share the same RF channel (frequency) without causing unacceptable interference. These codes allow Code Division Multiple Access (CDMA) to the shared RF channel (frequency).

The OVSF codes are illustrated in the following code tree:

Figure 7 OVSF code tree for channel codes

The code tree defines the code length used to provide the specified spreading factor. The higher user data rate services use shorter codes and hence lower spreading factors (and associated de-spreading gain).

A given mobile cannot use all channel codes simultaneously. A channel code can only be used by a mobile if no other code on the path from the specific code to the root of the code tree, or in the sub-tree below the specific code, is used by that mobile. Thus the number of available channel codes is not fixed, but depends on the data rate and associated spreading factor of each physical channel used.



25

For each call, the mobile is allocated at least one uplink channel code, for an uplink DPCCH. Usually, at least one further uplink channel code is allocated for an uplink DPDCH. Additional uplink channel codes may be allocated if the mobile needs more DPDCHs. All channel codes used for the DPDCH must be orthogonal to the channel code used for the DPCCH.

As each mobile using the same RF channel uses a unique uplink scrambling code, no co-ordination of the allocation of uplink channel codes to mobiles is needed. They are allocated in a predefined order that exploits the design of the scrambling codes used by the mobile transmitter.

The mobile and the network may negotiate the number and length (spreading factor) of the channel codes needed for the call, and the network allocates the necessary codes.

Scrambling codes

To allow identification during inter-cell handover, each mobile is assigned a unique code which is not repeated at other cells.

The uplink uses either short or long scrambling codes, depending on the capabilities of the base station receiver. In both cases complex scrambling is used, in which each code allocation consists of a pair of I & Q codes.

Short scrambling codes are used in cells that use a sophisticated receiver with a multiple user detector and interference canceller.

With short codes, the cross correlation between different physical channels and users does not vary with time as it does with long codes. Consequently the cross correlation matrices used in the advanced receiver have to be updated less frequently, reducing the complexity of the receiver design. The base station informs the mobile of its code allocation using the Access Grant message (although it is possible to change the code allocation during a call).

Owing to their better interference averaging properties, long codes are used if the base station does not support multiple user detection.

Each long code maps to an indicated short code, and the Access Grant message informs the mobile whether it is to use the indicated short code or the corresponding long code.

Random access codes

The base station broadcasts a cell specific spreading code for the preamble part of the Random Access Message. Additional codes may be broadcast if the traffic loading is high. These preamble spreading codes must be co-ordinated between cells to avoid interference. A real-valued 256 chip orthogonal Gold code is used, and all 256 codes may be used by the system.

A preamble signature code is used that carries one of 16 different orthogonal complex signatures of length 16 based on a set of orthogonal Gold codes of length 16. The base station broadcasts the signatures that may be used in a cell.




The preamble signature specifies one of the 16 nodes in the code tree that correspond to channel codes of length 16. The sub-tree below the specified node is used to spread the message part of the Random Access Message.

In addition to spreading, the message part is also scrambled with a 10ms complex code. This code is cell specific, and is associated with the spreading code used for the preamble part. Although the scrambling code is the same for each access slot, the scrambling codes do not collide in different access slots, as the slots are time shifted by 1.25ms.

Modulation

Quadrature Phase Shift Keying (QPSK) is used with a chip rate of 3.84M chips/s.

To reduce the linearity requirements on the power amplifier and hence improve its efficiency, the scrambling codes are designed so that N-1 out of N consecutive chips produce +/- 900 rotations of the I&Q multiplexed data and control channels. The remaining 1 out of N chips produces 0,+/- 900, or 1800 rotation.

In addition to scrambling codes design, compatible uplink channel codes must be chosen. This limits the phase transitions of the baseband signal, prior to input to a pulse-shaping filter which ensures the resulting modulated signal is constrained within the RF spectrum channel mask (that is, adjacent channel interference is limited).



27

Downlink

Spreading Dedicated Physical Channel and Common Control Physical Channel

QPSK data modulation is used to spread Dedicated Physical Channels (DPCH) and Common Control Physical Channels (CCPCH) in the downlink. Pairs of baseband data bits are taken in serial form and applied in parallel to the I & Q branches of the spreading and modulation chain. The I & Q branches are then spread to the chip rate with the same channel code cch (that is, real spreading) and then scrambled using the same cell specific scrambling code cscramb.

This is illustrated in the following diagram:

Figure 8 Downlink spreading and modulation

Each physical channel uses a different channel code, but the same scrambling code.

In addition to the DPCH and CCPCH the Synchronisation Channel (SCH) is also multiplexed onto the downlink transmission.

The SCH is transmitted intermittently, one code word per slot, and multiplexed onto the downlink after the DPCH and CCPCH have been scrambled. The SCH is therefore not orthogonal to the other downlink channels.

This is illustrated in the following diagram:

Spreading & Modulation for Downlink DPCH & CCPCHs

++

DedicatedPhysicalChannel(DPCH)andCommonControlPhysicalChannel(CCPCH)

OrthogonalVariableSpreadingFactor (OVSF)Code (c )ch

OVSFCode (c )ch

Cscramb

cos ( t)ω

sin ( t)ω

Realp(t)

p(t)Imag

I

Q

I+jQSerialto ParallelConversion




Figure 9 Downlink multiplexing of synchronisation channel

Channel codes

The OVSF codes used in the uplink are also used as downlink channel codes, to ensure orthogonality between the channels operating at different data rates and spreading factors.

The code assignment restrictions that apply to the uplink on a per-mobile basis also apply on the downlink on a per-cell basis.

The channel code used for the Primary CCPCH is predefined for all cells within the network. The BCH broadcasts the channel code used for the Secondary CCPCH.

The channel codes used for the downlink Dedicated Physical Channels are allocated by the network. The base station tells the mobile which channel codes to receive by, in the downlink Access Grant message response to the uplink Random Access request. This channel code set may be changed during a call (for example, after a handover or change of service type). The change of channel code set is negotiated over the Dedicated Control Channel (DCCH).



29

Scrambling codes

A total of 512 scrambling codes are available. They are divided into 32 sub-groups of 16 codes to aid rapid cell searching. A scrambling code is assigned to a cell at initial deployment, and the mobile is advised of the code during its cell search process.

As a solution to cell congestion, the facility to assign more than one scrambling code to a cell may be adopted in the future.

Synchronisation codes

Synchronisation is required for the mobile to decode the downlink data when first receiving a base station, and in the uplink direction for the base station to decode the mobile’s random access requests.

Primary and secondary codes are orthogonal Golay codes of length 256. They are formed by combining pairs of m-sequence codes and transmitted at the system chip rate.

Modulation

As on the uplink, Quadrature Phase Shift Keying (QPSK) with a chip rate of 3.84M chips/s is used on the downlink, with similar pulse shaping and filtering.

2.6. Physical channel frame structure

A 10ms frame structure is used for all information transmitted on both uplink and downlink physical channels. The 10ms frame structure is divided into 15 time slots, each of 666.7µs duration.

72 of these 10ms frames form one super-frame of 720ms duration.

The data rate carried by each 10ms time slot depends on the type of transport channel carried and the type of service provided.

Uplink time slot fields

Each uplink 666.7µs radio time slot may contain a number of fields, the data capacity of which varies according to the format that is used.

When the time slot carries the DPDCH, it has one data field per time slot (Ndata), which contains the number of data bits shown below:




Slot Format Channel Bit Rate /kbs-1

Channel Symbol Rate /k symbols s-1

Spreading Factor

Bits per Frame

Bits per Slot Ndata Number of Data Bits

0 15 15 256 150 10 10

1 30 30 128 300 20 20

2 60 60 64 600 40 40

3 120 120 32 1200 80 80

4 240 240 16 2400 160 160

5 480 480 8 4800 320 320

6 960 960 4 9600 640 640

Table 2 Formats for DPDCH data fields in uplink radio time slot

When the time slot carries the DPCCH, it has four data fields per time slot

• Pilot (Npilot )

• Transport Format Combination Indication (TFCI) (NTFCI )

• Feedback Information (FBI) (NFBI)

• Transmit Power Control (TPC) (NTPC)

The bit assignment according to slot format is shown below:

Slot Format

Channel Bit Rate

/kbs-1

Channel Symbol Rate /k symbol

s s-1

Spreading Factor

Bits per Frame

Bits per Slot

Npilot Number of Pilot

Data Bits

NTPC Number of TPC Data Bits

NTFCI Number of TFCI

Data Bits

NFBI Number of FBI Data Bits

0 15 15 256 150 10 6 2 2 0

1 15 15 256 150 10 8 2 0 0

2 15 15 256 150 10 5 2 2 1

3 15 15 256 150 10 7 2 0 1

4 15 15 256 150 10 6 2 0 2

5 15 15 256 150 10 5 1 2 2

Table 3 Formats for DPCCH data fields in uplink radio time slot



31

Downlink time slot fields

The downlink Secondary Common Control Physical Channel (CCPCH) has three data fields per time slot:


• Data (Ndata)

• Pilot (Npilot )

Slot Format

Channel Bit Rate

/kbs-1


s-1

Spreading Factor

Bits per Frame

Bits per Slot

Ndata Number of Data

Bits


Data Bits


Data Bits

0 30 15 256 300 20 12 8 0

1 30 15 256 300 20 10 8 2

2 60 30 128 600 40 32 8 0

3 60 30 128 600 40 30 8 2

4 120 60 64 1200 80 64 8 8

5 240 120 32 2400 160 144 8 8

6 480 240 16 4800 320 296 16 8

7 960 480 8 9600 640 616 16 8

8 1920 960 4 19200 1280 1256 16 8

Table 4 Secondary CCPCH data fields in downlink radio time slot – with pilot

Slot Format

Channel Bit Rate

/kbs-1


s-1

Spreading Factor

Bits per Frame

Bits per Slot

Ndata Number of Data

Bits


Data Bits


Data Bits

0 30 15 256 300 20 20 0 0

1 30 15 256 300 20 18 0 2

2 60 30 128 600 40 40 0 0

3 60 30 128 600 40 38 0 2

4 120 60 64 1200 80 72 0 8

5 240 120 32 2400 160 152 0 8

6 480 240 16 4800 320 312 0 8

7 960 480 8 9600 640 632 0 8

8 1920 960 4 19200 1280 1272 0 8

Table 5 Secondary CCPCH data fields in downlink radio time slot – without pilot

The downlink time slots carry both the Dedicated Physical Control Channel (DPCCH) and Dedicated Physical Channel (DDPCH) in an interleaved manner, and have five data fields per time slot:





DPCCH

• Data Field 1 (Ndata1)

DPDCH

• Transmit Power Control (NTPC )

DPCCH

• Data Field 2 (Ndata2)

DPDCH

• Pilot (Npilot )

DPCCH



33

Slot Format

Ch. Bit Rate kbs-1

Ch. Symbol Rate ksyms-1

SF Bits per Frame DPD CH

Bits per Frame DPDC CH

Bits per Frame Total

Bits per Slot

DPD CH Bits per Slot Ndata1

DPD CH Bits per Slot Ndata2

DPDC CH Bits per Slot NTFCI

DPDC CH Bits per Slot NTPCI

DPDC CH Bits per Slot NpliotI

0 15 7.5 512 60 90 150 10 2 2 0 2 4

1 15 7.5 512 30 120 150 10 0 2 2 2 4

2 30 15 256 240 60 300 20 2 14 0 2 2

3 30 15 256 210 90 300 20 0 14 2 2 2

4 30 15 256 210 90 300 20 2 12 0 2 4

5 30 15 256 180 120 300 20 0 12 2 2 4

6 30 15 256 150 150 300 20 2 8 0 2 8

7 30 15 256 120 180 300 20 0 8 2 2 8

8 60 30 128 510 90 600 40 6 28 0 2 4

9 60 30 128 480 120 600 40 4 28 2 2 4

10 60 30 128 450 150 600 40 6 24 0 2 8

11 60 30 128 420 180 600 40 4 24 2 2 8

12 120 60 64 900 300 1200 80 4 56 8 4 8

13 240 120 32 2100 300 2400 160 20 120 8 4 8

14 480 240 16 4320 480 4800 320 48 240 8 8 16

15 960 480 8 9120 480 9600 640 112 496 8 8 16

16 1920 960 4 18720 480 19200 1280 240 1008 8 8 16

Table 6 Downlink DPDCH and DPCCH data fields




2.7. Speech coding

Prior to transmission over a digital radio system, the analogue audio signal must be digitised. This function can be performed by a coder/decoder (codec). There are two types of codec:

• Waveform codecs

• Voice codecs (also known as vocoders)

Waveform codecs replicate the analogue waveform as faithfully as possible. Voice codecs extract the essential intelligibility information from speech, and reproduce it in a comprehensible manner, without regard to the accuracy with which the original analogue signal is reproduced.

Where bandwidth efficiency is not essential, such as in the landline telephone network, waveform codecs are used (for example, ITU-T G.711 1972). Where bandwidth efficiency is essential, such as in mobile radio systems, voice codecs are normally used.

As accuracy of the original analogue signal reproduction is not a measure of vocoder performance, their performance is based on subjective tests of perceived voice quality. One commonly used measure is the Mean Opinion Score (MOS). This is based on a jury marking the intelligibility and speaker recognition of pre-defined phrases spoken through the vocoder system.

The MOS value of 4.3 for the land-line standard ITU-T G.711 for Pulse Code Modulation (PCM) 64 bit is normally used as a base reference level.

In the case of residential subscribers and small businesses requiring only a few lines, G.711 circuits are not terminated at the customer premises, but are presented by an analogue ‘local loop’.

The quality degradation introduced by the local loop varies greatly, depending on its length and transmission means. For example, several miles of analogue catenary with loading coils (typically used in rural locations) can introduce significant degradation. It has been reported that a typical G.711 circuit presented by analogue local loop has a MOS of 4.0.

A wide variety of vocoder techniques are used depending on the application involved; but generally the faster the resultant data stream, the better the voice quality and resulting MOS value. When comparing MOS values, the comparison must be made when the vocoders are operating at the same Bit Error Rate (BER).



35

Currently the GSM and UMTS standards define six codecs supported by the UTRAN air interface:

• GSM Full Rate

• GSM Half Rate

• GSM Enhanced Full Rate

• GSM Adaptive Multi-Rate

- Full Rate

- Half Rate

• UMTS Adaptive Multi-Rate

The definition of the common codec list in 3GPP for GSM and UMTS follows the specifications given in ITU Q.765.5.

Transcoder Free Operation

The UMTS Technical Specifications outline the 3GPP internal codec list for both GSM and UMTS codecs to be used by the Bearer Independent Call Control protocol to set up or modify a call in Transcoder Free Operation (TrFO).

TrFO allows the transport of speech signals in the coded domain from one mobile to another through the radio access network and core network, and possibly through an additional transit network. This enables high speech quality, low transmission costs and high flexibility.

Codec type selection and resource allocation is negotiated out-of-band before and after call setup.

Possible Codec (re-)configuration, Rate Control and Discontinuous Transmission (DTX) signaling may be performed after call setup by additional in-band signaling, or by a combination of in-band and out-of-band signaling.

Up to release ´99, GSM does not support TrFO and specifies Tandem Free Operation (TFO) instead. TFO offers similar advantages to TrFO but is based on pure in-band signaling after call setup. The UMTS Technical Specifications allow interaction between TrFO and TFO. They also provide a GSM evolutionary path toward TFO.




GSM Full Rate codec

The Regular Pulse Excited Long Term Prediction (RPE-LTP) vocoder was selected for GSM in 1989. This generates an output bit rate of 13kb/s, and has a maximum MOS of 3.7. This data rate is termed the ‘Full Rate’ for a GSM system.

The RPE-LTP vocoder is formed from the following sections: LTP filter, Linear Prediction Coder (LPC), and de-emphasis filter. (The RPE-LTP is also known as the RPE-LPC.)

DTX can be independently enabled in the uplink and/or downlink, as defined by the network on a cell basis. It cannot be negotiated at call setup or during the call. The DTX scheme uses one Silence Information Descriptor (SID) frame to mark the end of a speech burst and to start comfort noise generation. Identical SID frames for comfort noise updates are sent in speech pauses about every 480 ms, aligned with the cell´s TDMA frame structure.

TFO allows the reception of GSM-FR DTX information for the downlink direction in all cases.

GSM Half Rate codec

The GSM Half Rate codec type supports one fixed Codec Mode with 5.6kbs-1. DTX may be used as in GSM Full Rate.

Owing to poor speech quality, this codec is not widely used in GSM networks.

GSM Enhanced Full Rate codec

The GSM Enhanced Full Rate codec type supports one fixed Codec Mode with 12.2kbs-1.

Again, DTX may be enabled in the uplink and/or the downlink independently, as defined by the network on a cell basis and cannot be negotiated at call setup or during the call.

The DTX scheme uses one SID frame to mark the end of a speech burst and to start comfort noise generation. It is important to note that the parameters for the start of comfort noise generation are calculated at the transmitter side from the previous eight speech frames. A DTX hangover period therefore needs to be applied at the transmitter side before sending the first SID frame.

SID frames with incremental information for comfort noise updates are sent in speech pauses approximately every 480 ms, aligned with the TDMA frame structure of the cell. The defined TFO allows reception of GSM EFR DTX information for the downlink direction in all cases.



37

Adaptive Multi-Rate codec types

Adaptive Multi Rate (AMR) is a new mobile technology that introduces a speech and channel codec able to support both GSM full rate (22.8kbs-1 gross bit rate) and half rate (11.4kbs-1 gross bit rate) channel modes. For each channel mode a number of different Codec Mode bit rates can be employed.

The Lucent GSM AMR feature is realised using a speech codec located in the Speech Transcoding Frame (STF-2000) and the mobile, and a channel codec located in the BTS-2000 and the mobile.

AMR differs from existing GSM speech codecs, in that it can adapt its data rate (speech coding) and error protection level (channel coding) in accordance with the prevailing radio channel conditions. By selecting the most appropriate channel mode (AMR Half Rate or AMR Full Rate) and the Codec Mode (combination of speech and channel bit rates), AMR is able to offer a balance between speech quality and network capacity.

The sampling rate is 8 000 samples/s leading to a bit rate for the encoded bit stream of 4.75, 5.15, 5.90, 6.70, 7.40, 7.95, 10.2 or 12.2kbs-1.

The coding scheme for the multi-rate coding modes is the Algebraic Code Excited Linear Prediction Coder (ACELP). The multi-rate ACELP coder is referred to as MR-ACELP.

The AMR Codec algorithm is applied in GSM and UMTS in three different codec types:

• Full Rate AMR

• Half Rate AMR

• UMTS AMR




The AMR operating modes are shown below.

Channel type Source coding bit rate

TCH/FS/AMR

(TCH/AFS)

12.2 kbit/s

10.2 kbit/s (GSM EFR)

7.95 kbit/s 7.40 kbit/s (IS136 EFR)

6.70 kbit/s

5.90 kbit/s 5.15 kbit/s

4.75 kbit/s

TCH/HS/AMR

(TCH/AHS)

7.95 kbit/s 7.40 kbit/s (IS136 EFR)

6.70 kbit/s

5.90 kbit/s 5.15 kbit/s

4.75 kbit/s

Table 7 AMR Codec Modes

At call set-up the network selects a suitable Codec Mode set (containing a maximum of 4 Codec Mode bit rates) appropriate for the call. The uplink and downlink must use the same Codec Mode set. However, during the call the respective links may use different Codec Mode bit rates from within the chosen Codec Mode set.

During a call, Codec Mode adaptation is controlled by the BTS. Quality measurements made by the BTS (uplink) and MS (downlink) are compared against pre-defined threshold/hysteresis values from which the BTS is able to decide whether a Codec Mode change is required. The BTS informs the MS and TRAU of any Codec Mode changes using in-band signalling. In theory Codec Mode adaptation can take place every speech frame. However, due to propagation delays and necessary filtering in the codec adaptation functions, a lower rate is recommended.



39

TRAU

SPE

SPD

BTS

CHD

CHE

Uplink Speech Data

Downlink Speech Data

MS

SPE CHE

SPD CHD

Codec Mode Indication (uplink)

Codec Mode Indication (downlink)

Sug. Mode Command (downlink)

Codec Mode Command (uplink)

CodecAdaptation

CodecAdaptation

Figure 10 Codec mode signalling for GSM AMR

In the uplink direction, the mobile indicates the Codec Mode employed on the current and next speech frames and suggests a Codec Mode to be used for the downlink. In the downlink direction the BTS uses the uplink quality measurements to generate an uplink Codec Mode command instructing the mobile to use a new Codec Mode at the earliest opportunity (earliest speech frame) as well as to indicate the Codec Mode for the downlink. This is based on either the mobile suggested mode or a BTS selected mode.

The channel mode adaptations are controlled by the BSC using the Radio Link Control (RLC) algorithms. The BSC is also responsible for the initial Codec Mode set selection. Channel mode changes occur far less frequently, no more than a few times per minute. Channel mode changes can take place in either of the following circumstances:

• At call set-up and after handover

• During a call - dynamic channel mode adaptation (AMR handover)

Both links must use the same channel mode.




AMR benefits

AMR provides improved speech quality in both FR and HR modes. Recent characterisation tests on the AMR codec conducted by ETSI have shown AMR Full Rate (AFS) to be superior to GSM EFR, particularly under bad interference conditions. AMR Half Rate (AHS) was equivalent or better than GSM FR under good radio channel conditions – down to 16dB CIR.

AMR parameters

The AMR has additional parameters, which are optional at the originating side but mandatory for the terminating side:

• Active Codec Set (ACS). Eight bits

- In FR AMR and HR AMR up to four modes may be selected by setting the corresponding bits to “1”

- In HR AMR only four out of the lower six modes can be selected

- In UMTS AMR all eight modes may be selected

- If the ACS is not specified at the originating side, all modes are supported there. The terminating side may then select freely

- If ACS is not provided, SCS and MACS (see below) also cannot be provided

• Supported Codec Set (SCS). Eight bits

- In FR AMR and UMTS AMR up to eight modes may be selected by setting the corresponding bits to “1”

- In HR AMR only the lower six modes may be selected

- If the SCS is not specified at the originating side, all modes are supported there. The terminating side may then select freely

- If SCS is not provided, MACS (see below) also cannot be provided

• Maximal number of Codec Modes (MACS). Three bits

- In FR AMR and HR AMR, one to four Codec Modes are allowed within the ACS

- In UMTS AMR, one to eight Codec Modes are allowed within the ACS

- If MACS is not specified at the originating side, then the maximum number of modes is supported there. The terminating side may then select freely

• Initial Codec Mode. Three bits.

- One of the Codec Modes within the ACS is indicated as the starting mode

- If the ICM is not specified at the originating side, the terminating side may select freely



41

- The Length Indicator field (LI) is set to 3, 4, 5 or 6 at the originating side, depending on how many parameters are specified. The terminating side returns the selected codec with a full set of parameters. Hence LI is always set to 6 by the terminating side. If any node in the path from the originating side to the terminating side does not support the parameter set offered by the originating side, it may restrict it. If necessary the missing, optional parameter octets are inserted then

GSM AMR codec types

The GSM AMR codec types comprise eight Full Rate and six Half Rate different Codec Modes (for example, 12.2 … 4.75 kbs-1).

The active Codec Mode is selected from the Active Codec Set (ACS) by the network (Codec Mode Command) with assistance by the mobile station (Codec Mode Request). This Codec Mode Adaptation, also termed Rate Control, can be performed every 40 ms by going one Codec Mode up or down within the ACS. The Codec Modes in uplink and downlink at one radio leg may be different. In Tandem Free Operation both radio legs (A and B) are considered by the “Distributed Rate Control” algorithm for the optimal selection of the active Codec Mode in each direction (uplink A and then downlink B, and uplink B and then downlink A respectively). The worst of both radio legs determines the highest allowed Codec Mode, respectively the maximum allowed rate. Besides this “Maximum Rate Control” the active Codec Mode may sometimes be frozen to a fixed mode by either of the two radio legs to allow a smooth handover procedure (“Exact Rate Control”).

All rate control commands are transmitted in-band: on the radio interface, the BTS-TRAU interface and the TRAU-TRAU interface.

The Active Codec Set is configured at call setup or reconfigured during the call. It consists of from one to four Codec Modes (MACS) at a given time, selected from the Supported Codec Set. The maximum number of MACS and the Supported Codec Set may be constrained by the network to consider resources and radio conditions.

The same Active Codec Sets are available in both the uplink and downlink. Different sets may be selected in the uplink and downlink.

At start up of Tandem Free Operation, both Active Codec Sets are taken into account to determine the common Active Codec Set. In a later phase the Supported Codec Sets and MACSs of both radio legs may be taken into account to find the optimum Common Active Codec Set. All configuration data and update protocols are transmitted in-band.

DTX scheme

The DTX scheme of the AMR Codec Type marks the end of a speech burst with a specific SID_FIRST frame. SID_FIRST does not contain comfort noise parameters. It starts the comfort noise generation with parameters calculated at receiver side from the most recently received seven speech frames. Therefore, a DTX hangover period needs to be applied at transmitter side before sending the SID_FIRST.




Absolutely coded SID_UPDATE frames follow approximately every eighth frame (160 ms) in speech pauses. SID_UPDATE frames are sent independently of the cell´s TDMA frame structure and are related only to the source signal.

Typically, in the uplink direction, an ONSET frame precedes a new speech burst. DTX on or off is defined by the network on a cell basis. The defined Tandem Free Operation allows the reception of GSM-AMR DTX information for the downlink direction in all cases.

Note: The DTX scheme of the Enhanced Full Rate codec type is not compatible with the DTX scheme of the AMR codec type in Codec Mode 12.2kbs-1, although the speech modes of the two codec types are bit identical.

UMTS AMR codec types

The UMTS AMR codec type comprises eight different Codec Modes (for example, 12.2 … 4.75kbs-1).

The active Codec Mode is selected from the Active Codec Set by the network. This is known as Codec Mode adaptation. Codec Mode adaptation can be performed every 20 ms by going to any arbitrary Codec Mode within the ACS. The Codec Modes in uplink and downlink at one radio leg may be different.

In TFO or TrFO operation, both radio legs (A and B) are considered for the optimal selection of the active Codec Mode in each direction (uplink A and then downlink B) by a Distributed Rate Control algorithm. The worst value of the two radio legs determines the highest allowed Codec Mode (maximum allowed rate). This selection process is known as maximum rate control. The active Codec Mode may sometimes be frozen to a fixed mode by any of the two radio legs to allow a smooth handover procedure (exact rate control). All rate control commands are transmitted in-band on the IU and A interface and out-of-band on the radio interface.

UMTS AMR Active Codec Set functionality is the same as for GSM AMR codec types.

At call setup the originating side sends the AMR parameter set in the Codec List. The terminating side then selects a suitable Active Codec Set from the given information and sends it back. If the terminating side does not support TrFO, a transcoder is allocated at a suitable position in the path, as close as possible to the terminating side. This transcoder can use in-band signaling to install TFO after call setup. Then, at TFO start up, both Active Codec Sets are taken into account to determine the common Active Codec Set.

In a later phase the Supported Codec Sets and MACSs of both radio legs may be taken into account to find the optimum Common Active Codec Set. All configuration data and update protocols are transmitted in-band on the TFO interface, but (possibly) out of band within the UMTS network.

For more information on TFO refer to GSM Recommendation 08.62 & Technical Specification 28.062.



43

Source Controlled Rate scheme

The Source Controlled Rate (SCR) scheme of the default AMR codec type marks the end of a speech burst with a specific SID_FIRST frame. SID_FIRST does not contain Comfort Noise parameters. It starts comfort noise generation with parameters calculated at receiver side from the latest received seven speech frames. A DTX hangover period must therefore be applied at transmitter side before sending the SID_FIRST.

Absolutely coded SID_UPDATE frames follow approximately every eighth frame (160 ms) in speech pauses. SID_UPDATE frames are sent independently of the cell´s timing structure and are related only to the source signal.

Typically, an ONSET frame does not exist in UMTS networks but may be received in TFO from the distant partner. It marks the beginning of a speech burst. SCR on or off is defined by the network on a cell or call basis. TFO and TrFO always allow the reception of AMR SCR information for the downlink.

The SCR scheme of UMTS AMR codec types is fully compatible with the DTX scheme of GSM FR AMR and HR AMR.

For compatibility with other systems, the UMTS AMR Codec application may optionally support various other DTX and rate control schemes (such as GSM-EFR, TDMA-EFR, TDMA-US1, PDC-EFR). It is not currently possible to establish TFO between UMTS and these other systems, but this is likely to be supported in a later release.

Further reference

Details of these codec types and their related procedures such as DTX and rate control are described in the respective standards documents.




2.8. Codec speech quality

This section summarises the perceived speech quality of the various speech codecs.

Fixed rate codecs

• Full rate (13kbs-1)

- Regular Pulse Excited Long Term Prediction (RPE-LTP)

- Adopted 1989, with a maximum MOS of 3.7

• Enhanced full rate (13kbs-1)

- Advanced Code Linear Predictive (A-CELP)


• Half rate (5.6kbs-1)

- Vector Sum Excited Linear Predictive (V-SELP)


Figure 11 GSM Fixed Rate Codecs MOS against CIR



45

Advantages

• Enhanced full rate vocoders provide subscribers with improved perceived speech quality, close to that of a land-line link. Enhanced full rate can increase the perceived speech quality MOS from 3.7 to 4.0, (compared to 4.3 for land-line)

• Half rate vocoders provide increased network capacity (for example, an increase of up to 40%, when 60% of mobiles are capable of half rate operation)

• New base stations are not required

• Additional back-haul link capacity is not required

Disadvantages

• Subscribers may not consider half rate speech quality as acceptable, particularly in mobile to mobile calls

• Additional network capacity provided by the use of Half-Rate Codecs depends on the proportion of subscribers with half rate mobiles

• Enhanced full rate improved speech quality is only available to subscribers with new, enhanced full rate mobiles

• Enhanced full rate and half rate speech coding will be superseded by Adaptive Multi Rate speech coding from 2001 (this will require new mobiles and network infrastructure upgrades)




Adaptive Multi Rate codecs

Adaptive Multi Rate devices provide the benefits of high speech quality from the EFR Codec combined with the bandwidth efficiency of using HR channels.

The following graph shows measurements taken during ETSI characterisation tests of the AMR codec for AMR clean speech in AFS and AHS modes, compared against GSM EFR, FR and HR.

M eanO pinionScore

C IR (dB)1

2

3

4

5

N oErr ors

7 410131619

W ireline Q uality R egionITU -T G .711 PC M- 64 &A nalogue Local Loop

HR

AH S

AFSEFR

FR

Figure 12 Speech quality (MOS) against CIR for GSM speech codecs

Advantages

• Greater robustness to channel errors in FR mode under poor radio channel conditions

• Increased network capacity through codec operation in AMR Half Rate mode and allowing a lower CIR compared to existing GSM codecs

• Significant capacity gains (up to 90%) can be achieved through AMR mode

• Extra capacity is freed up for expected increase in data traffic

• Extended coverage deeper into buildings

• AMR operation can be tailored to meet different operator needs



47

AMR can be used in the following modes:

• Full Rate mode for maximum robustness to channel errors

• Half Rate mode to maximise network capacity

• Mixed Full/Half Rate mode. This allows operators to balance the trade-off between speech quality and network capacity in accordance with prevailing radio channel and traffic load conditions




This page is intentionally left blank


Mobile Handover

3


49

3. Mobile Handover

This chapter outlines the three main types of UMTS mobile handover, with particular emphasis on inter system handover between a UTRAN and a GSM network.

3.1. Handover types

Three types of mobile handover are supported:

• Soft / softer handover

A soft handover takes place when the mobile establishes a connection with a new cell, while it is still in communication with its current serving cell. This can only take place if both cells use the same RF channel (frequency).

A softer handover is a special type of soft handover that takes place if both the cells involved are controlled from the same BTS (or Node B) network element. The main difference is in the implementation within the UTRAN. Uplink maximal ratio combining instead of selection combining can be used for the softer handover, as the former is performed at the BTS, while the latter takes place at the Radio Network Controller (RNC) network element.

• UTRAN to UTRAN hard handover

A UTRAN to UTRAN hard handover may take place under the following conditions:

- Handover between cells operating on different RF channels (frequencies). For example when cells are allocated different numbers of RF channels, belong to different networks, or to different cell layers




- Handover between cells operating in different duplex modes (FDD or TDD). For example, when entering a building with a local high speed TDD network

- Handover between cells when operating in packet mode, when soft handover is not needed

- Handover between different radio transceivers at the same cell, operating on different RF channels

• UTRAN to GSM handover

- A UTRAN to GSM handover occurs when a mobile is handed over between a UTRAN and a GSM traffic channel, where an isolated area of UTRAN coverage exists within a GSM network. This type of handover is to be expected in the early stages of UTRAN deployment, when UTRAN coverage may not be contiguous over network operators’ areas. Additionally, to promote competition some countries are encouraging the issue of UTRAN licences to companies that do not operate an existing GSM network. Under these circumstances, existing GSM licenses have been amended to require support for roaming by UTRAN subscribers

To support inter frequency hard handover, (for example, between UTRANs or between a UTRAN and a GSM network), the mobile must conduct a cell search on a different frequency from that used for the serving cell, without interrupting the data flow associated with the call in progress. This can be achieved by either of the following methods:

• Using a dual receiver

• Slotted downlink transmission



51

3.2. Cell sets

The mobile maintains three cell sets:

• Active Set

This lists all UTRAN cells that are currently assigning a downlink Dedicated Physical Channel (DPCH) to the mobile. It includes any cell with which the mobile is in soft or softer handover.

• Handover Monitoring Set

This lists all the cells (UTRAN and GSM) that the mobile has been tasked to monitor. The list corresponds to the BCCH Allocation (BA) list sent on the Slow Associated Control Channel (SACCH) in the GSM network.

A separate set can be defined to monitor for cell selection, corresponding to the BCCH Allocation (BA) list sent on the Broadcast Control channel (BCCH) in the GSM network.

• Handover Target Set

This lists the cells received by the mobile at a level sufficient for the associated DPCHs to be decoded, but which are not currently in the Active Set. These cells may be on different frequencies from the current serving cell, and part of a UTRAN or GSM network.




3.3. Preparation for UTRAN to UTRAN handover

The Handover Monitoring Set is sent to the mobile via the BCH of the serving cell or via specific signalling on the Dedicated Channel (DCH). The Handover Monitoring Set contains the following data for each monitored cell:

• Downlink scrambling code

• Cell ID number

The network knows the mapping of cell scrambling codes to synchronisation codes (groups indicated by the secondary synchronisation channel). During the neighbour cell measurement process the mobile uses the primary and secondary synchronisation channels to synchronise to the cells, together with knowledge of the scrambling codes in use.

From both the serving and neighbour cells, the mobile measures the downlink:

• Received signal level

• Relative timing between serving and neighbour cells, based on the phase difference of the scrambling codes

• Bit Error Rate (BER) / Block Error Rate (BLER)



53

3.4. Preparation for UTRAN to GSM handover

The UTRAN super-frame structure is 720ms in length and comprises 72 frames of 10ms each. The GSM super-frame structure is 6.12s in length and comprises 26 frames of 120ms each.

As 720ms is a multiple of 120ms, it is possible to use compatible timing for inter system measurements. A dual-standard (UTRAN/GSM) mobile, when operating in UTRAN mode, can use the GSM Frequency Correction Channel (FCH) and GSM Synchronisation Channel (SCH) data to maintain the timing between a UTRAN carrier and a GSM carrier. The effect is as if there were two asynchronous GSM carriers.

A dual-standard mobile can make GSM network measurements during the idle periods in the downlink transmission that are created using downlink slotted mode. The use of slotted mode is controlled by the UTRAN, which informs the mobile which frame is slotted.

Slotted frames are achieved by compressing the data that would normally be sent in a 10ms frame, by either code puncturing or reducing the spreading factor, so that it can be sent in only 5ms. This creates a 5ms silent period (the Silence Duration) during which the mobile can make GSM signal level measurements.

To avoid compromising the quality of the radio link during transmission of the compressed slot data, the transmitted power is increased. When high data rate services are used, it may not be possible to create the idle slot (the silent period) by compressing a single time slot of data. In this case the data in a number of slots can be compressed to create the required idle slot.

As an alternative to slotted frames, mobiles can use an independent GSM receiver branch. However, although more flexible, this approach is likely to increase the cost of mobiles.

GSM system information must be exchanged between the two networks. This is to allow the UTRAN to broadcast the GSM BA list (GSM frequencies to monitor) to the mobiles in the UTRAN area.

Due to the inherently lower maximum data rate available from a GSM network, a more integrated form of network operation is required if the particular type of service is to be maintained during inter system handover.




Silence Duration parameters

The absolute length of the Silence Duration (SD) depends on the time taken for the mobile to switch from the UMTS frequency to a GSM frequency, decode and measure the GSM channel data, and switch back to the UTRAN frequency.

Note: Appendix A contains Silence Duration parameter definitions reproduced from the ETSI UTRAN Handover standard.

GSM cell timing unknown

If there is no knowledge of the relative timing of the UTRAN and GSM cells, Silence Duration Patterns are used to search and decode the GSM Synchronisation Channel (SCH). The process is repeated whenever the mobile receives a new SCH.

Depending on the mobile’s capabilities, the SCH search may be either sequential (track first GSM Frequency Correction Channel before decoding SCH), or parallel (parallel tracking of FCH and SCH). The parallel option decodes the SCH faster and thus needs fewer SD patterns.

For example, a parallel search with 2 SD patterns every 0.48s alternately using Tpattern1 (the delay between successive SD patterns) of 226.92ms (47 * 4.615ms) and Tpattern2 263.08ms (57 * 4.615ms) is as efficient as a sequential search with 4 SD patterns per 0.48s with Tpattern 120ms.

The number of successive patterns used to scan a specific GSM frequency (Npattern) before it is assumed to be unsuccessful, can vary the probability of detection against the number of slotted frames (that is, the impact on UTRAN link). The default settings for Npattern is 11 for serial searches and 6 for parallel searches, to ensure successful detection under worst case GSM cell timing conditions.

If the SDs are allocated by the UTRAN on a periodic basis, the mobile triggers the search procedure within the available SDs. So no specific signalling is needed between the mobile and the UTRAN.

Alternatively, the mobile may initiate a search by sending a Request New Cell Search message to the UTRAN, within which it indicates its serial/parallel search capability. The UTRAN calculates a suitable SD pattern and advises the mobile using the normal SD indicators. The network operator can delay implementation of this SD pattern according to the timing priority assigned for New BSIC (Base Station Identity Code) identification. When the mobile completes its search, it signals to the UTRAN the timing of the associated SCH (or SCH Not Found).

Examples of Silence Duration and associated SD patterns are given in Appendix A.



55

GSM cell timing known

The mobile may gain some knowledge of the relative timing between the UTRAN Active Cell set and the GSM neighbour cells, either from the UTRAN, or following acquisition of the GSM FCH. When the timing is known, the Silence Duration pattern may be assigned for a specific frame intended to capture a specific GSM SCH/FCH.

The control is implemented over 306 frames (13 GSM multi-frames). As the UTRAN 720ms super-frame shifts ¼ of a super-frame during the period, the 4 x 306-frame period is used to align the timing of a UTRAN cell and a GSM cell.

The mobile performs an SCH search whenever a new GSM neighbour cell is received, by sending a Request New Cell Search to the UTRAN. The UTRAN responds with:

• Frame number where the slotted mode takes place (frame number {[x] + [n*306]} where n is an integer)

• GSM carrier for which the particular slotted frame is intended (BSIC, CI, ARFCN)

When the search is complete, the mobile signals the UTRAN the timing of the SCH (or SCH Not Found), and the UTRAN stops the SD pattern.

GSM cell BSIC reconfirmation

Once the mobile has successfully received the SCH for a neighbour GSM cell, it must reconfirm the cell Base Station Identity Code (BSIC) to establish the identity of the cell.

If Silence Durations (slotted frames) are allocated on a periodic basis on the downlink, the mobile autonomously performs this process, and no specific signalling between the UTRAN and mobile is necessary.

If Silence Durations are not allocated on a periodic basis on the downlink, the mobile advises the UTRAN of the timing of the SCH and the length of the Silence Duration required to capture one SCH burst. The UTRAN operator sets the BSIC reconfirmation target time and this together with the timing data received from the mobile is used by the UTRAN to determine the Silence Duration pattern. The Silence Duration pattern is communicated to the mobile using the normal Silence Duration Indicators. If the UTRAN already has timing data regarding the neighbouring GSM cells, it can calculate the Silence Duration pattern with no mobile involvement.




Silence Duration parameters for GSM power measurement

In order to measure the received power level of GSM carriers, additional Silence Duration (that is, slotted frames) will be required by mobiles without a separate GSM receiver branch.

Monitoring multiple GSM neighbour cells

The previous sections introduced the concept of initial synchronisation, identification, and power measurement of a single GSM neighbour cell. In real deployment situations this must be extended to monitor multiple GSM neighbour cells in the following circumstances.

• Handover owing to lack of UTRAN coverage

When a dual-standard mobile moves outside the coverage of the UTRAN, but remains within coverage from a GSM network, it may handover from the UTRAN to the GSM network.

The number of downlink slotted frames per reporting period (0.48s) depends on:

- Type of slotted frame used (start, end, double start/end or mid-frame)

- Number of BCH measurements per slotted frame that the mobile can perform

- Number of measurements required per reporting period

• Handover owing to bearer services

A network operator with both a UTRAN and a GSM network may wish to use the UTRAN for high data rate services and the GSM network for low data rate services (such as voice). In this case, when a dual-standard mobile asks to set-up a call on the UTRAN, it may be assigned channels on either the UTRAN or the GSM network, according to the bearer service it negotiates with the network.

When the call set-up involves DCH allocation, the slotted mode is required to handle GSM handover preparation. The downlink slotted mode starts as soon as the network knows the requested bearer service. For example, when included in the initial bearer request sent from the mobile for a mobile originated call, or known by the network for a network originated call. The measurements are performed over a number of reporting periods.

When the call set-up does not involve DCH allocation, downlink slotted frames are not required. The mobile cannot listen to GSM channels as the downlink UTRAN messages can occur at any time.



57

Monitoring and reporting GSM neighbour cells

A dual-standard mobile which is monitoring GSM neighbour cells, performs the measurements specified in the GSM recommendations:

• Monitors up to 32 GSM carriers

• Synchronises with up to 6 GSM carriers

• Sends measurement reports back to the network on the 6 strongest GSM cells for which the BSIC is correctly decoded, (assuming the GSM Multi-band Reporting parameter is set for Normal Reporting)

• Performs the measurements down to the reference sensitivity and reference interference levels specified in GSM Recommendation 05.05

• Demodulates the GSM SCH on the BCCH carrier of each neighbour cell and decodes the BSIC as often as possible, and as a minimum once every 10s

The UTRAN to GSM handover may be invoked in either of the following circumstances.


If the mobile can make systematic GSM neighbour cell measurements, it does so when served by a UTRAN cell on the edge of the UTRAN coverage area. If the mobile cannot make systematic GSM neighbour cell measurements, the UTRAN decides whether to activate this as a background task, based on the uplink and downlink received signal level.


If a Dedicated Channel is allocated during call set-up, implementation of slotted mode on the downlink is necessary to allow the mobile to make GSM cell measurements. If a Dedicated Channel is not allocated during call set-up, no measurements are needed.

The handover procedure will stop in the following circumstances:

• The mobile is instructed to execute a handover to GSM

• The UTRAN includes a new cell in the Active Set (that is, start soft hand-off)

• Execution of an inter frequency handover within the UTRAN




3.5. Handover execution

This section describes the execution of intra-UTRAN soft handover and UTRAN-GSM handover.

UTRAN soft handover

The bearer service used by the mobile is known to the serving cell(s) of the Active Set. When a new cell is added to the Active Set the UTRAN passes the following data to the new cell:

• Maximum data rate and other service parameters associated with the duplex connection (for example, coding schemes, number of parallel code channels)

• Mobile ID and uplink scrambling code

• Timing information for the new cell, with respect to the time synchronisation of the mobile with the current serving cell(s), as measured by the mobile at its current location. The new cell uses this to calculate the timing of its common channel (for example, BCH) transmission

The mobile is given the following information via its current serving cell(s):

• The frame that the new cell uses to start transmission to the mobile, (assuming the Active Set update is accepted)

The channel code(s) used for the transmission by different cells need not be the same, and each cell uses a different scrambling code

UTRAN to GSM handover


When the mobile executes the handover from the UTRAN to the GSM network it stops communication with the UTRAN and establishes its signalling link on the target GSM cell (as defined in GSM Recommendations 05.08 and 04.08). If the handover is not successful the mobile may attempt to resume communication with its previous UTRAN serving cell(s).


If the call set-up involves Dedicated Channel allocation, after receiving the handover execution message from the UTRAN to the GSM network, the mobile stops communication with the UTRAN and establishes its signalling link on the target GSM cell (as defined in GSM Recommendations 05.08 and 04.08). If the handover is not successful the mobile may attempt to resume communication with its previous UTRAN serving cell(s).

- If the call set-up does not involve Dedicated Channel allocation, the mobile is assigned a GSM channel and stops communicating with the UTRAN serving cell(s). As it has not yet synchronised to the GSM cell, the mobile then listens to the GSM target cell BCH frequency to establish its time slot and frame



59

synchronisation by decoding the FCH and SCH. Once completed the mobile establishes a dedicated traffic channel on the required frequency of the target cell.

3.6. GSM to UTRAN handover

To simplify cell identification, the GSM system may also indicate the UTRAN base station scrambling codes in the GSM/UTRAN service area. As the UTRAN does not use a super-frame structure to indicate synchronisation, a dual-standard mobile operating in GSM mode may obtain frame synchronisation with the UTRAN once the UTRAN base station scrambling code timing has been acquired. The scrambling code has a 10ms period and is synchronised to the UTRAN Common Channel frame timing.


Subscriber Services

4


61

4. Subscriber Services

UMTS supports numerous subscriber services. This chapter outlines a few of the most common services, and identifies a number of features which may impact network RF engineering from a coverage and capacity viewpoint.

Prior to transmission of user data services, the following channel coding and service multiplexing functions are performed:

• Channel coding and interleaving for subscriber services

• Service multiplexing

• Rate matching

• Channel coding and interleaving for control channels




4.1. Coding and interleaving for subscriber services

Three classes of Forward Error Correction (FEC) are offered:

• Service specific coding

In this case no Layer 1 FEC coding is applied.

• Standard Services coding

Inner convolutional coding and inner interleaving is applied, yielding a Bit Error Rate (BER) better than 1 x 10-3

• High quality service coding

In addition to the coding and interleaving used for Standard Services, an outer level of Reed-Solomon coding and interleaving is applied, yielding a BER better than 1 x 10-6

4.2. Services multiplexing

When a single connection (call) requires multiple services, they may be time multiplexed onto a single bearer. This multiplexing may take place either before or after the inner or outer coding.

After multiplexing and coding, the multiple service data stream is mapped onto one Dedicated Physical Data Channel (DPDCH). If the total data rate exceeds the upper limit for a single channel code, the data stream is mapped onto several DPDCHs.

Alternatively, instead of multiplexing multiple services, each service may be separately coded and interleaved. Then they can be mapped onto separate DPDCHs using multiple codes. This approach allows independent control of the power and quality of each service but requires added complexity in the mobile to support multiple code transmission.



63

4.3. Rate matching

The rate matching process takes the data stream resulting from the channel coding and service multiplexing, which may be highly variable, and matches it to the closest corresponding fixed rate available from a DPDCH.

The process differs in the uplink and downlink.

Uplink

The data rate of the nearest DPDCH is achieved using unequal repetition (a sub-set of bits are repeated) or code puncturing. Code puncturing is used for bit rates of less than 20% above the nearest lower DPDCH data rate. Otherwise unequal repetition is used to increase the user data rate to that of the nearest higher DPDCH data rate. The repetition or puncturing pattern follows a pre-defined pattern at both the mobile and base station.

Downlink

Rate matching by unequal repetition or puncturing is only used for the highest data rate (after channel coding and multiplexing) of a variable rate connection, and for fixed-rate connections. For the lower of a variable rate connection, the same repetition or puncturing pattern used for the highest data rate is used, and the rate matching is achieved by discontinuous transmission (by transmitting for part only of each time slot). This technique simplifies the implementation of blind-rate detection at the mobile.




4.4. Control channel coding and interleaving

Dedicated Control Channel

The Dedicated Control Channel (DCCH) uses the same convolutional coding scheme as the traffic channels. Intra-frame block interleaving is performed after channel coding. The DCCH is mapped to the Dedicated Physical Data Channel (DPDCH) in the same way as the Dedicated Traffic Channels.

Downlink Common Control Channels

The Downlink Common Control Channels (BCH, FACH and PCH) also use the same convolutional coding scheme as the traffic channels. Intra-frame block interleaving is performed after channel coding, prior to mapping onto the primary and secondary Common Control Physical Channels (CCPCH).

On the secondary CCPCH, the FACH and PCH are time division multiplexed on a frame-by-frame basis within the super-frame structure. The frame set used by the FACH and PCH is broadcast to mobiles on the BCH.

4.5. Channel mapping examples

8kbs-1 bearer - speech

An 8-bit Cyclic Redundancy Check (CRC) and 8 tail bits are added to each 8kbs-1 speech frame prior to channel coding and mapping onto a DPDCH. Unequal repetition is used to match the 28.88kbs-1 data rate after channel coding to the closest DPDCH rate of 32kbs-1.

The mapping process is:

1. Data (80 bits) + CRC (8 bits) + Tail (8 bits) = 96 bits

2. Data of 96 bits (convolutional code rate 1/3 K=9) gives (96 * 3) = 288 bits

3. 288 bits (unequal repetition 9>10) gives (288 * 10/9) = 320 bits

4. 320 bits mapped onto the closest DPDCH rate of 32kbs-1



65

144kbs-1 bearer - data

Each 144kbs-1 data frame is Reed-Solomon coded, has tail bits added, and is convolutionally frame coded. Code puncturing is used to match the 542.4kbs-1 data rate after channel coding to the closest DPDCH rate of 512kbs-1 DPDCH.


1. Data (1440 bits) (Reed-Solomon code rate 180/225) gives 1800 bits

2. Data of 1800 bits + tail (8 bits) gives 1808 bits

3. 1808 bits (convolutional code rate 1/3, K=9) gives (1808 * 3) = 5424 bits

4. Data of 5424 bits (code punctured 339>320) gives (5424 * 320/339) = 5120 bits



Each 384kbs-1 data frame is Reed-Solomon coded and convolutionally frame coded. Unequal repetition is used to map the resulting 964.8kbs-1 data stream onto a 1024kbs-1 DPDCH.


1. Data (3840 bits) (Reed-Solomon code rate 192/240) gives 4800 bits

2. Data of 4800 bits + 3 tails (3*8 bits) gives 4824 bits

3. 4824 bits (convolutional code rate 1/2, K=9) gives (4824 * 2) = 9648 bits

4. Data of 9648 bits (unequal repetition 603>640) gives (9648 * 640/603) = 10240 bits






Each of the sixteen parallel blocks (each block is 300 bits) have a 12-bit sequence number and CRC added, prior to convolutional coding and mapping onto a 1024kbs-1 DPDCH. Rate matching is not necessary.


1. Data (300 bits) +SN & CRC (12 bits) + tail (8 bits) gives 16 blocks per frame

2. Data of 16 blocks of 320 bits each (convolutional code rate 1/2, K=9) gives (16*2*320) = 10240 bits



Lucent Equipment

5


67

5. Lucent Equipment

This chapter describes Lucent UMTS base station system equipment and associated attributes and interdependencies in relation to RF network capacity and design.

The principal network elements are:

• Node-B or Base Transceiver Station (BTS)

• Radio Network Controller (RNC)

The primary interfaces between the BTS and the RNC are illustrated in the following diagram:




Figure 13 Radio access network elements

5.1. Node-B (BTS)

The Node-B (or BTS) is the base station transceiver rack and is the furthest extremity of the UTRAN. It is located in a radio station and has two principal interfaces:

• The air (Uu) interface is connected to an antenna system, through which it communicates with the mobiles (known as User Equipment in UMTS). Its primary tasks are to provide:

- The air interface

- Combining/splitting functions for information streams borne by several physical channels

- Implementation of macro-diversity through soft (and softer) handover

• The network interface (Iub) to the Radio Network Controller (which performs similar functions to the GSM BSC)

Lucent produces four types of BTS:

• Distributed Milli-cell

• Macrocell

• Milli-cell



69

• Microcell

Distributed Milli-cell

The Distributed Milli-cell comprises an indoor or outdoor central unit which houses the network interface, BTS controller, and baseband processing hardware which can be linked to up to 6 remote RF heads. The remote link between the RF heads and the baseband unit can be optical fibre. The Distributed Milli-cell uses the indoor or outdoor Milli-cell rack assembly for the baseband, power, control and interface components, and the Microcell enclosure for the remote RF heads.

The Distributed Milli-cell and the Microcell are illustrated below:

Figure 14 Distributed Milli-cell BTS and Microcell BTS




Microcell (Ultra-small cell)

The Microcell (or Ultra-small cell) BTS is a single sector single carrier system, AC powered, and suitable for wall or pole outdoor mounting. If additional carriers are required, up to three Microcells may be linked together. The Microcell uses different circuit cards to the other BTS units and can only output 5W per carrier.

Macrocell

The Macrocell BTS is a three sector system, transmitting three carriers with transmit diversity, fitted with any combination of baseband units (32 or 64 voice channels). It uses the same indoor cabinet as the Milli-cell, but has more powerful power amplifiers and a second baseband card frame.

The Macrocell BTS is not available in Lucent Network Release 1.0.

Milli-cell

The Milli-cell BTS is a three sector system, with one or two 20W carriers and transmit diversity, housed in either an outdoor or indoor cabinet. It supports either a masthead low noise amplifier (MHA) or the standard rack mounted receiver multi-coupler amplifier.



71

1600 mm

1100 mm

266 mm

110 mm

75 mm

600 mm

75 mm

140 mm

775 mm

76 mm

1300 mm

Figure 15a Milli-cell BTS rack configuration




Lucent circuit pack names in bold

N+1

Misc. StationInterfaces

…

BTS-CNT

Radio Unit

BRU

BaseBandProcessor

BBU

BTS/RNCInterfaces

TX Amp

M CLA

Duplexer, LNAs

Clock UnitCDU

(EXT-INF)EXTINF

LNA controller

BTS

CCNT

SCNT

NIU

MaintenanceTools

M T

Figure 15b Milli-cell BTS internal architecture

The Milli-cell rack incorporates a number of features to optimise reliability:

• Hot-standby redundancy

- Precision oscillator card

- Clock distribution card

• Warm-standby redundancy

- Controller cards

• Load sharing redundancy

- Multiple back-haul interfaces (for example, 2Mbs-1 ‘E1’)

- Baseband processing cards

- Power supply units

- Transmitter power amplifier (if the transmitter diversity option is used)



73

• N+1 redundancy

- Radio transceiver cards

The Milli-cell base station has capacity for 11 Baseband Unit (BBU) cards. The BBU cards may be:

• BBU4 – available in Network Release 0.1, based on fully programmable hardware. Capacity for 4 voice circuits

• BBU32 – available in Network Release 1.0, based on the first generation of ASIC. Capacity for 32 voice circuits

• BBU64 – available in later releases, based on the second generation ASIC. Capacity for 64 voice circuits

• BBU128 – planned in the future. Capacity for 128 voice circuits

Note: When considering BTS capacity be aware that the common control channels use approximately half the capacity of a BBU32 card.




Characteristic Distributed milli-cell

Macrocell[6] Milli-cell Microcell

Sectors 6 sectors of 1 carrier, or 3 sectors

with 2 carriers

3 3 1

Carriers @ Power 3 @ 20W 1 or 2 @ 20W 1 @ 5W

Cabinet size (h x w x d) /mm

RF head 650x300x100

BBU 800x600x400

1800x600x400 indoor

1600x1300x650 outdoor

1800x600x400 indoor

1600x1300x650 outdoor

300x650x100 indoor & outdoor

Weight (max.) /kg ~270kg

Number of baseband units 11 22 11 N/A

Number of voice channels[1] 256 with BBU32[2]

608 with BBU64[2]

<120 voice circuits per carrier per sector

560 with BBU32 1264 with BBU64


256 with BBU32 608 with BBU64


16 with ASIC[5] no.1

48 with ASIC no. 2

Aggregate data capacity[1] 8.192Mbs-1 BBU32 9.728Mbs-1 BBU64

17.92Mbs-1 BBU32 20.224Mbs-1 BBU64

8.192Mbs-1 BBU32 9.728Mbs-1 BBU64

1.024Mbs-1 BBU32 3.072Mbs-1 BBU64

Typical power consumption ~6kW [3] ~3kW [4] ~250W

Antenna diversity Transmit (10W per carrier per antenna)

/ Receive

2 carriers with no transmit diversity

1 carrier with transmit diversity

Transmit (2.5W per antenna) Receive

Table 8 Main characteristics of Lucent BTS types

[1] Assumes the following: FACH/RACH uses 8 BBU4/BBU32 or 4 BBI 64 cards per sector per channel element Data rate 64kbs-1 per channel element One voice channel is a duplex channel pair comprising one 32ksymbol s-1 downlink physical channel and one 64ksymbol s-1 uplink physical channel The air interface capacity is ~120 voice circuits or 1Mbs-1 per sector per carrier Capacity calculated for maximum number of carriers

[2] BBU32 accommodates 32 voice channels. BBU64 accommodates 64 voice channels [3] Configuration of 3 sector, 2 carrier, 13 BBU [4] Configuration of 3 sector, 1 carrier, 13 BBU [5] Application Specific Integrated Circuit [6] Not available in Lucent Network Release 1.0



75

BTS traffic capacity

Where possible, the Lucent BTS range has been designed to use a number of common replaceable units.

The number and type of the sub-assemblies fitted to a specific BTS determine the capacity of the BTS per RF carrier. The number and type of baseband units (BBU) is a key factor in determining the capacity of the BTS.

The following table illustrates typical traffic capacity per RF carrier in terms of the number of the different circuit types supported and number and type of BBUs fitted.

Traffic Type Air Interface Max. No. of Simultaneous Channels (Calls) per Sector per Carrier

Air Interface Max. No. of Simultaneous Channels (Calls) per Tri-Sector Cell per

Carrier

Max. No. of Simultaneous

Channels (Calls) per BTS

(fully fitted with BBU)


Channels (Calls) per BBU

Using Baseband 4 Units

8kbs-1 voice channels

80 (~108 including all

mobiles in soft handover)

324 32 4

64kbs-1 data channels

9 (~12 including all mobiles in soft

handover)

36 16 2



handover)

21 8 1



handover)

9 Not Applicable Not Applicable





324 256 32



handover)

36 128 16



handover)

21 40 5



handover)

9 16 2




Traffic Type Air Interface Max. No. of Simultaneous Channels (Calls) per Sector per Carrier

Air Interface Max. No. of Simultaneous Channels (Calls) per Tri-Sector Cell per

Carrier


Channels (Calls) per BTS

(fully fitted with BBU)


Channels (Calls) per BBU





324 608 64




36 304 16



handover)

21 95 8



handover)

9 38 4

Table 9 Voice channel capacity of Air Interface and BBU cards compared

Note: Transmit diversity is not used.

The higher the data rate of the required circuit, the fewer that can be supported per RF channel (regardless of any BTS hardware limitations) and per BBU (and consequently per BTS). This is because a higher branch of the channel coding tree is used to support a higher data rate, and hence fewer branches are available.



77

The capacity of the air interface to support voice circuits can be compared with the capacity of the Milli-cell BTS to accommodate BBU32 and BBU64 cards, and hence provide hardware support for voice circuits.

Base station configuration

Voice circuit capacity of air interface

Voice circuit capacity using BBU32 Cards

Voice circuit capacity using BBU64 Cards

1 Sector 1 Carrier

80 336 688

2 Sectors 1 Carrier

160 320 672

2 Sectors 2 Carriers

320 288 640

3 Sectors 1 Carrier

240 304 656

3 Sectors 2 Carriers

480 256 608

Table 10 Comparison of air interface voice channel capacity and Milli-cell BTS

Notes:

1. Voice circuits are 13kbs-1 duplex

2. Half the capacity of one BBU32 is used for control channels

3. Estimated capacity of air interface is 80 voice circuits at 13kbs-1 per sector

4. Eleven BBU cards are installed

BTS further reference

For detailed information about BTS configurations, Lucent personnel can refer to the UBTS-SAS-HW System Architecture Specification for the Unicell BTS Hardware UMTS Node B Volume 400, Chapter 500, Section 241, Sub-section 1.

Please note that as this document contains commercially sensitive information, it is not generally available to our customers. If you need more details in this area, please contact your Lucent representative, who will arrange to provide the required extracts from the document.




5.2. BTS Antennas

In addition to the usual BTS site antenna system design criteria (such as RF characteristics and site and electrical constraints) the existence and type of any network already owned by the organisation is crucial. The organisation may own a single band (900MHz or 1800MHz) GSM network, a dual band GSM network, or may not own an existing network. Each situation is discussed in this section.

At the time of writing, dual band (GSM/UMTS) antennas have been developed by a number of manufacturers and are about to enter production. Tri band antennas, with separate isolated ports and elements for 900MHz, 1800MHz and 2GHz are at the prototype stage and expected to enter production early in 2001.

No existing network

The simplest situation is if a company obtains a licence for an area where it does not own an existing network, does not wish to share antennas with an existing network, and has no future intention of operating a network on a different band.

In this case it is only necessary to install antennas for the UTRAN band. For convenience, we refer to these as 2GHz antennas.

Existing single band network

An existing GSM network operator may obtain a UTRAN licence, and intend to operate both a UTRAN and a GSM network for a period of time. The GSM network may operate on the 900MHz or 1800MHz band.

For logistic and economic reasons the operator is likely to want to use existing GSM network BTS sites to provide UTRAN coverage. So it will be necessary to install an antenna system at the GSM sites, which can serve both the GSM and the UTRAN BTSs.

Separate GSM and UTRAN antennas

The simplest way to operate a dual band (GSM and UTRAN) base station is to install separate antennas for the two bands. In an existing network this needs only one additional antenna system to serve the new 2GHz band. However, this approach has some significant disadvantages, including:

• Greater loading on the antenna support structure

• Increased number of antenna feeders and associated duct/riser space required

• Increased rental costs

• Increased environmental impact with subsequent implications for planning consent

• Requirement to carry spares for each antenna type



79

As a result, operators are likely to install dual band antennas.

Dual band GSM/UTRAN antennas

In view of the magnitude of the frequency span, dual band antennas used for the 900MHz and 2GHz bands normally employ separate radiating elements for each band. These are enclosed in a common radome, and may comprise elements printed on a substrate (microstrip elements), or a number of three-dimensional (for example, wire) individual dipole elements.

Although both the 900MHz and 2GHz element arrays are contained within the same radome, and may be laid on the same substrate, different electrical down-tilts may be applied to each array. Additionally, the electrical down-tilt may be either fixed or electrically adjustable.

Typically, a fixed tilt is used for one array and its ground level radiation pattern is adjusted by mechanically tilting the whole antenna assembly. An electrical down-tilt is applied to the second array allowing its ground level radiation pattern to be separately adjusted.

Dual band antennas used for the GSM 1800MHz and UMTS 2GHz bands may use the same radiating elements. This is because the bandwidth required to cover both bands (+/-230MHz) is only ~+/-12% of the centre frequency (1.94GHz). When a single antenna is used for both bands it is clearly not possible to have different values of down-tilt (either electrical or mechanical) applied to the two different bands.

In this case, if a single dual band antenna is to be used for duplex operation on both bands, it is necessary to:

• Divide the received signal and provide a receive feed to both the 1800MHz and the 2GHz BTS

• Combine the transmit signal from the 1800MHz and 2GHz BTS, and provide a common feed to the antenna

Both the receive path receive distribution amplifier and power divider, and the transmit path combiner can be located in the equipment room.

As the 2GHz band is separated by ~40MHz from the 1800MHz band, relatively low loss broadband frequency selective band combiners can be used.

If a dual polarisation dual band antenna is used to provide receive polarisation diversity, and transmit diversity is not required, the 1800MHz BTS transmitter can feed one polarisation, and the 2GHz BTS transmitter can feed the other, thereby reducing transmitter combining loss.




Existing dual band network

Existing dual band GSM network operators who obtain a UTRAN licence, may want to continue to operate the dual band GSM network in addition to a UTRAN for the foreseeable future. The existing dual band network may operate with either single or dual band (900/1800MHz) antennas.

There are three possible options:

• Install additional 2GHz antennas

• At sites with separate 900 and 1800MHz antennas, replace the 1800MHz antenna with a dual band 1800MHz / 2GHz antenna

• At sites with dual band 900/1800MHz antennas, replace the dual band 900/1800MHz antenna with a tri-band 900/1800/2000MHz band antenna

Separate GSM and UTRAN antennas

This option is unlikely to be popular, for the same reasons previously outlined for single band networks upgrading to UTRAN coverage.

Dual band GSM/UTRAN antennas

A dual band GSM/UTRAN antenna can be used under two circumstances:

• At sites with a separate 1800MHz antenna, the 1800MHz band antenna can be replaced by a dual band 1800/2000MHz antenna

• At sites with a dual band 900/1800MHz antenna, the dual band 900/1800MHz antenna can be replaced by a single band 900MHz antenna and a dual band 1800/2000MHz antenna

Both options yield a 900/1800/2000MHz antenna system using two antennas.

The previous comments about dual band GSM/UTRAN antennas in an existing GSM single band network also apply to existing dual band networks.

Tri band GSM/UTRAN antennas

To date, tri band antenna developments have been based on the use of two arrays of radiating elements, one array for the 900MHz band and one broadband array for both the 1800MHz and 2GHz bands. Both arrays are contained in one radome.

This technique allows different electrical down-tilts for the 900MHz beam and the combined 1800/2000MHz beam. This is reasonable since the propagation characteristics at 1800MHz and 2GHz are similar, when compared with 900MHz propagation.



81

Tri band antennas are likely to be installed at sites which currently use dual band 900/1800MHz antennas. They allow the capability to provide service on 900MHz, 1800MHz, and 2GHz from a single antenna system.

As with the dual band GSM/UTRAN antenna, a similar 1800MHz and 2GHz signal power receive path division and transmit path combination is required. However, in the case of the tri band antenna, if a single 900/1800/2000MHz feeder (for each polarisation) is used, it is also necessary to:

• In the equipment room, combine the transmit and divide the receive 900, 1800, & 2000MHz power

Due to the spacing between the three bands, frequency selective broadband power combiners with relatively low insertion loss (<1dB) can be used.

• At the masthead, or within the tri band antenna:

Combine the received 900MHz power from the 900MHz antenna elements with the 1800 & 2000MHz received power from the broadband 1800/2000MHz antenna elements.

Divide the BTS transmit power at 900MHz from that at 1800 & 2000MHz, and present it to the 900MHz antenna elements, and the broadband 1800/2000MHz antenna elements respectively.

The power division/combining between the 900MHz and 1800/2000MHz paths can be implemented externally or within the antenna radome.

In addition to the standard choices of single or dual polarisation, azimuth and elevation beamwidth, and gain, some dual band GSM/UTRAN and tri band antennas provide an internal 900MHz and 1800/2000MHz band diplexer (cross-band coupler).

The main configurations available for dual band GSM/UTRAN antenna feeders are illustrated in the following diagram:




Figure 16 Example configurations for dual band GSM/UTRAN antennas

Note: Dual polarisation antennas have been illustrated to demonstrate the most complex configuration, but they are not mandatory for dual GSM/UTRAN band operation.

Tri band antenna with internal band diplexer

This combines the signals from both the 900MHz and 1800/2000MHz band elements onto a single mast feeder.

The advantages of a single feeder system are:

• Mast/tower loading is similar to that for GSM operation

• The same number of feeders is required as for GSM operation (no additional feeders to install when upgrading a GSM site for UTRAN operation)

• Reduced justification for rental increase at leased sites

• Reduced installation time

• Reduced maintenance effort

• No additional cable riser/duct/tray/ladder space required



83

The disadvantages of a single feeder system are:

• Because common feeders are used, the improved cell availability for users of dual standard GSM/UTRAN mobiles due to the redundancy arising from provision of GSM and UTRAN coverage is reduced

• Losses incurred in the internal antenna band diplexer

• Masthead pre-amplifiers cannot be added at their optimum position (which is immediately after the antenna)

• Additional losses incurred in the cross-band combiner installed near the base station transceiver rack

Tri band antenna without internal band diplexer

This type of antenna has two separate ports: one for the 900MHz feeder and one for the 1800/2000-MHz feeder.

It can be used with an external cross-band combiner to achieve a similar effect to a dual band GSM/UTRAN antenna with internal cross-band diplexer. However, this approach requires additional tails, and the components that are not housed within the antenna require weatherproof housing and mounting.

The relative advantages and disadvantages of using a dual band antenna without internal cross-band diplexer are the exact converse of those for an antenna with the internal cross-band diplexer.




Dual band and tri band GSM/UTRAN diplexers

Dual band and tri band diplexers are used to combine and divide two and three (possibly duplex) signal bands, so that more than one band can share a common antenna or antenna feeder.

Isolation between the ports for each band is necessary to avoid high power transmit signals from one band impairing the performance of the receiver(s) operating on the other bands.

Dual band diplexers

Dual band diplexers may take the following configurations:

• 900MHz and 2GHz

• 1800MHz and 2GHz

Dual band diplexers are three port devices used for the following purposes:

• To combine a Band 1 signal on port one with a Band 2 signal on port two, and present the composite signal on port 3

• To divide a composite Band 1 and Band 2 signal input on port 3, and present the Band 1 signal on port one, and the Band 2 signal on port two

• Provide inter port isolation between ports 1 and 2

Example

Assume that a base station site wishes to operate a GSM 900, GSM 1800, and UTRAN BTS using a single dual polar duplex antenna with a single port for 900, 1800, and 2000MHz signals, and an internal 900MHz and 1800-2000MHz diplexer.

A 1800MHz and 2GHz dual band diplexer can be used to combine the signals from the GSM 1800 and UTRAN BTSs. A second dual band diplexer can be used at the ‘ground end’ of the feeder to combine this signal with that from the GSM 900 BTS. This configuration is shown in the right of the following diagram.

If a dual band antenna without an internal diplexer is used with a single 900/1800/2000MHz feeder, a second dual band diplexer with 900MHz and 1800-2000Morts may be used at the ‘antenna end’ of the feeder. This second diplexer divides the 900MHz and 1800-2000MHz signals from the common feeder and presents them to the appropriate antenna ports. This configuration is shown in the centre illustration in the following diagram



85

Figure 17 Example use of separate dual band GSM1800/UTRAN & GSM900/GSM1800-2GHz diplexers (cross-band couplers)

Dual band GSM/UTRAN diplexers are available for both indoor and outdoor mounting.




Tri band diplexers

A tri band diplexer is a four port device used for any of the following purposes:

• To combine a Band 1 signal on port one with a Band 2 signal on port two and a Band 3 signal on port 3, and present the composite signal on port 4

• To divide a composite Band 1, Band 2, and Band 3 signal input on port 4, and present the Band 1 signal on port one, the Band 2 signal on port two, and the Band 3 signal on port 3

• Provide inter port isolation between ports 1, 2 and 3

The tri band diplexer can be used instead of two separate dual band diplexers (one for 1800MHz and 2000MHz and one for 900MHz and 1800-2000MHz). This simplifies design, installation and maintenance.

Typical performance figures for a tri band diplexer are:

• Pass band

- Port 1 880-960MHz, Port 2 1710-1850MHz, Port 3 1900-2170MHz

• Return loss >20dB

• Insertion loss

- Port 1 <0.2dB, Port 2 <0.5dB, Port 3 <0.5dB

• Inter-band isolation >60dB

• Continuous power through each port 200W (CW)

• Intermodulation 7th order product

- Port 1 <-160dBc, Port 2 <-160dBc, Port 3 <-175dB

• Intermodulation 3rd order product

- Port 1 <-155dBc, Port 2 <-155dBc, Port 3 Not Applicable

• Intermodulation 2nd order product

- Port 1 <-160dBc, Port 2 <-160dBc, Port 3 Not Applicable

• Size 230 x 120 x 55mm

• Connectors 7/16 female

• Operating temperature –40 to +850C



87

Broadband power divider

A range of broadband power dividers covering the 900, 1800, and 2000MHz bands in one device, is available from Lucent’s preferred antenna manufacturers. The range includes devices that divide the power between up to 4 ports, either equally according to a fixed ratio or according to a user variable ratio between the ports.

The components may be used to combine or divide power in an antenna feeder system. For example in an indoor multiple antenna distribution system or in a network of radiating cables.

Frequency selective isolation is not provided.

Broadband indoor antennas

The operating environment and close proximity of subscribers usually dictates that small size, covert styling, and broad beam-width are the principal design criteria for broadband indoor antennas. Dual polarisation and high gain are not usually provided.

A number of broadband 1800-2000MHz antennas have been developed that can be used for dual (GSM/UTRAN) band applications.

A few tri band antennas are available that cover the GSM 900MHz band as well as the 1800-2000MHz GSM and UTRAN bands.

Typical examples of these antennas include:

• Omni-directional ceiling mounted antennas with 2dBi gain are ~210mm diameter and 80mm high

• Directional low profile panels with a gain of 7dBi and an azimuth beam-width of 900 designed for wall mounting are ~205 x 155 x 42mm

However, these antennas usually provide no isolation between the two (or three) bands that are covered. A single common antenna feeder cable connector is used.




Antenna feeder

Apart from the higher loss experienced in the 2GHz band compared with the 900 and 1800MHz GSM bands, UTRAN networks impose no additional restrictions on the choice of antenna feeder cable compared to those applicable to GSM networks.

When upgrading an existing 1800MHz antenna system (or dual 900/1800MHz band) for simultaneous operation at 2GHz, the additional frequency dependent feeder loss is unlikely to be significant. It is normally possible to use the existing 1800MHz feeder for both services, provided that the additional loss associated with the dual (GSM/UTRAN) band diplexers (cross-band couplers) is acceptable.

Using an existing 900MHz band feeder for UTRAN services may introduce unacceptable loss except where the feeder length is relatively short. The combination of longitudinal loss in the feeder together with insertion losses in the two dual (GSM/UTRAN) band diplexers (one at each end of the feeder) may become unacceptable.

Factors to take into account when considering a common antenna feeder system for a UTRAN/GSM network are the same as for a dual band GSM 900/1800MHz network. The following table compares attenuation of common types of antenna feeder at 900MHz, 1800MHz, and 2000MHz:

Foam Dielectric Cable Nominal Diameter

Attenuation at 894MHz /dB Attenuation at 1.7GHz /dB Attenuation at 2GHz /dB

½” 0.722dB for 10m length

1.805dB for 25m length 3.61dB for 50m length

7.22dB for 100m length






11.3dB for 100m length 7/8” 0.403dB for 10m length










11/4” 0.298dB for 10m length









15/8” 0.252dB for 10m length









Table 11 Comparison of antenna feeder loss



89

Masthead amplifier

A masthead amplifier can be used at a UTRAN base station to improve the effective receiver system noise figure when a long length of feeder cable is used, in a similar manner as used at GSM 1800 base stations. As in a GSM receiver system, the reduction in the receiver system noise figure is translated into an improvement in the uplink power budget.

The worst case specification is usually associated with:

• Duplex operation. The masthead amplifier must incorporate two duplexers, to allow the transmit signal to bypass the receive amplifier

• Power feed using antenna cable. The masthead amplifier must incorporate a bias tee

Typical performance figures based on a duplex configuration, with DC power delivered via a bias tee, are likely to be:

• Transmit path insertion loss 0.25dB

• Receive path noise figure 1.8dB

• Receive path gain 12dB

• Size 300 x 170 x 96mm

• Optionally fitted inside antenna size 200 x 50 mm x length according to gain




Repeater

A repeater may be used to provide coverage for:

• Black-spots or ‘holes’ in areas of intended contiguous coverage

• Structures such as surface and underground buildings, and tunnels

• Long distances beyond the maximum operating range of individual base stations

The application and design criteria for using a UTRAN repeater are similar to those for GSM repeaters, with particular emphasis on the following matters.

Repeater adds noise to the donor cell receiver

To compensate for the added noise, a higher signal level must be received from all mobiles within the cell (donor cell and repeater cell) to maintain the minimum value of bit energy to noise ratio (Eb/[N0+NI]) necessary for satisfactory reception. However, as the mobiles have a finite output power, this has the effect of reducing the coverage range of the donor cell.

As there is a trade-off between traffic load (capacity) and coverage range (path loss power budget), when a repeater is used to extend the coverage range, the repeater cell range is traded-off against donor cell coverage. So when repeaters are used for continuous range extension, it is important to re-evaluate coverage of the donor cell. The greater the range extension, the greater the reduction in donor cell coverage area.

Using repeaters for scattered range extension (such as along a major road or to serve a minor outlying settlement) may provide economic coverage with a minimum number of BTS where coverage does not need to be contiguous.

Under these circumstances, it may be possible to serve the repeater from a BTS antenna sector with surplus capacity/available path loss budget.

Repeater coverage reduced when multiple carriers are received

If the repeater receives carriers from neighbouring cells in addition to the intended donor cell, the repeater’s power output is divided between the multiple carriers, resulting in a reduction in the power of each carrier.

Reception of multiple carriers also usually entails the transmission of unwanted noise to the neighbouring cells, thereby reducing their available capacity/available path loss.

This problem can be mitigated by the use of donor cell antennas that have narrow beam-width.



91

Repeater cell does not provide allow receive diversity

Whether the repeater is a band or channel selective type, it does not usually provide any facility for receive diversity (such as polarisation or space). This must be taken into account at the design stage when the path loss power budget is calculated.

Repeater specification details are not yet available.

Active/smart/adaptive antennas

The terms adaptive, smart and active cover many concepts. At the simplest level, a number of manufacturers are developing adaptive antennas based on conventional multi-dipole arrays, with remotely adjustable down-tilt. The Lucent BTS supports remotely adjustable antennas by providing 8 relay-driven discrete outputs.

Active or smart antennas based on planar arrays and multiple electronic beam steering techniques are being investigated. However, these solutions require close integration with the BTS, as signal processing and beam control must be controlled by the transceiver. Apart from the technical difficulties involved, the need for such antennas has not yet been demonstrated. During the early years of UTRAN deployment the base stations will be coverage (not capacity) limited and electronic beam steering antennas may not deliver a large improvement in gain. They are mainly promoted to improve capacity through interference cancellation and forming multiple narrow beams.

It is anticipated that electronic beam steering antennas may be deployed in 5-10 years if the cost becomes economic and frequency congestion becomes a problem.




General antenna comments

Many variations of base station antenna system design are possible. The most suitable configuration will depend on local site conditions and the existing antenna system type used by any existing network.

This document highlights just the principal antenna system arrangements and components that may be used, together with the design factors to be considered. Detailed analysis of the options and their respective merits will be required on a site by site basis.

The dual band 1800MHz and 2GHz, tri band 900MHz and 1800-2000MHz, and broadband 900-2000MHz antennas that are currently available provide no isolation between the 1800MHz and 2GHz signals - a common radiator array is used with no internal filtering. Therefore, when duplex antenna operation is used, both the 1800MHz and the 2GHz BTS duplexer antenna ports are presented with the transmitter power from the ‘other’ band. So in addition to providing protection for the BTS receiver from its transmitter (tx/rx isolation) it must provide sufficient out-of-band attenuation from the antenna port to the receiver port to protect its receiver from the other band transmitter.

If necessary, an existing 1800MHz BTS duplexer may be supplemented by a 1800/2000MHz diplexer (cross-band coupler) and/or circulator.

When an existing network operator deploys a dual GSM/UTRAN band or tri band antenna, a significant concern will be the need to minimise disruption to their existing service as the antennas are changed. This can be minimised by the use of temporary antennas and transitional mounting arrangements.

Deployment of dual band 1800/2000MHz or tri band 900/1800/2000MHz antenna systems should be considered at the outset when new GSM sites are constructed by existing GSM network operators or new GSM licensees. For a small increase in initial deployment cost, the potential disruption and costs associated with an upgrade of the antenna system to provide UTRAN coverage in future can be minimised if the operator later obtains a UTRAN licence or wishes to rent antenna facilities to a UTRAN operator.

A number of Lucent’s existing preferred antenna suppliers have developed a range of single band UTRAN and dual band GSM/UTRAN antennas.

Further reference – antennas and BTS accessories

Detailed information about antennas and other base station system ancillary equipment provided by Lucent’s preferred suppliers is available from our Product Management – BSS Ancillaries group. Please ask your Lucent representative for details.



93

5.3. Radio Network Controller

The Radio Network Controller (RNC) performs similar functions to the Base Station Controller (BSC) in a GSM network. It is usually positioned at the centre of a coverage region and has three principal interfaces:

• Core network (Iu) interface through which it communicates with the Mobile Switching Centre (MSC) for circuit switched traffic, and the Serving GPRS Support Node (SGSN) for packet data, both via Asynchronous Transfer Mode (ATM) interfaces at the MSC and SGSN respectively

• Neighbouring RNC (Iur) interface through which it communicates with surrounding RNCs when the serving RNC requires radio resources from cells controlled by a neighbouring RNC to maintain a connection on the air interface (for example, in soft handover)

• Base Transceiver Station (Iub) interface through which it communicates with the BTSs (Node Bs) that it controls

The RNC also provides interfaces to the Operations and Maintenance Centre (OMC) terminal, Local Maintenance Test Terminal (LMTT), and Short Message Service Cell Broadcast Centre (SMS-CB).

The primary tasks of the radio network controller are to control:

• Handover decisions that require signalling to the mobile

• Combining and splitting in support of the macro diversity available from using multiple BTSs to maintain a communication channel with the mobile

• Assignment of radio resources from one or more BTSs (either locally controlled by the same RNC or controlled from a neighbouring RNC)

The RNC rack layout and architecture block diagram is shown in the following diagram:




Figure 18 RNC configuration

The RNC is functionally split into two areas:

• Traffic Processing Unit (TPU)

Responsible for processing user data.

• Base Station Controller (BSC)

Responsible for implementing the signalling protocol.

This allows TPU and BSC processing to be handled by separate hardware platforms, with the following benefits:

• The processing capacity of each element can be selected individually according to the relative processing demands for signalling control and traffic processing

• The BSC RNC element can control base stations which operate a different air interface standard (such as GSM)



95

Figure 19 RNC Traffic Processing Unit and Base Station Controller

TPU tasks

• Traffic signalling

- Routing of signalling traffic between the BTS and BSC

- Routing of user data traffic

• Handover

- Frame selection and distribution

BSC tasks

• UTRAN Level 3 Signalling

- Air interface and core network interface

• Connection admission control

• Operations and Maintenance

- Configuration management

- Alarm reporting

- Performance collection and reporting




• Outer loop power control

RNC characteristics

The principal design attributes of the Lucent RNC are summarised below.

Interfaces

• Star or daisy chain connection of BTSs

- Up to 4 STM-1 circuits (Iub interface)

- Up to 160 sector carriers per RNC

• Core network interface (Iu)

- Iur interface multiplexed onto Iu

• Optical and electrical termination for STM-1 Iu and Iur interfaces

• Choice of two OMC interfaces:

- In-band with Iu interface

- Separate 100Base-T Ethernet

• Choice of two LMTT interfaces:

- 100Base-T Ethernet

- RIA RS 232

Capacity

• BTS link capacity for 160 E1 or 4 STM-1 or 1 STM-4

• Circuit capacity of 4,000Erl for voice channels with expansion to 10,000Erl (compared to 2,000Erl for the BCF)

- Limit is based on Intelligent Carrier Card (ICC) capacity in the Traffic Processing Unit

- TPU can support 10 active ICCs, each with a capacity of 400Erl

• Up to 6 soft handover branches per mobile connection

• Up to 144,000 Busy Hour Call Attempts (BHCA)



97

- Limit based on the processing capability of the Application Processors (AP) in the BSC

- BSC can support 4 APs, each with a capacity of 48,000 BHCA, plus some additional capacity for redundancy

The mix of subscriber services shown in the following table illustrates the capacity of the RNC:

Service User Data Rate /kbs-1

Busy Hour Call Attempts

Circuit Holding Time /s

Number of Subscribers per Service

Traffic Volume per Service /Erl

Voice 8 0.6 100 191,250 3,188

Switched Data 64 0.2 200 63,750 708

High Interactive Multimedia 144 0.03 1,200 8,500 85

Video Conference 384 0.03 3,600 425 13

Total Number of Subscribers Supported by RNC

263,925

Total Traffic Volume Supported by RNC /Erl

3,994

Table 12 RNC capacity

Example configurations for different sized RNCs are given below.

Minimum Configuration 1,200Erl

Medium Configuration 2,200Erl

Maximum Configuration 4,000Erl

Number of Subscribers 72,000 144,000 240,000

Busy Hour Call Attempts 43,200 86,400 144,000

Traffic Volume /Erl 1,200 2,400 4,000

Number of ICC Cards 3 + N Redundant 6 + N Redundant 10 + 2 redundant

Number of APs 2 for Redundancy 3 for Redundancy 4 for Redundancy

Table 13 Example RNC configurations

Notes: 1. Circuits are voice at 8kbs-1 2. 0.6 BHCA per subscriber 3. Call holding time 100s




Installation and commissioning

• Once frame and cabling is complete the installation of RNC should take <4hrs

• Once hardware installation is complete RNC downloads software diagnostics in <2hrs

Reliability

• 99.9998% availability at full capacity (1 minute per year non-availability)

• Mean Time Between Failures (MTBF) that reduce capacity of 200,000 hours

• MTBF of 10,000 hours for each Line Replaceable Unit (LRU)

• Single point of failure can be withstood with capacity reduction (for example, 15% of active stable calls)

• Routine maintenance visit frequency of one per annum

• Mean Time To Repair (MTTR) by exchanging LRUs is less than 5 minutes

• Hot-swap cards supported

Physical characteristics

• Operating temperature –5 0C to +45 0C (ETS 300 019-1-3)

• Operating voltage –40.5 to –57.0 VDC (ETS 300 132-2)

• EMC to EN 55022 Class A

• Power consumption to be decided (1.5kW for BCF)

• Size 2000 mm high x 600 mm wide x 600 mm deep

• Weight 250kg

• Floor loading <15kNm2



99

RNC further reference

For detailed information concerning the composition and configuration of the RNC, we recommend the following documents:

• UMTS RNC Release 1.0 Release Requirements Document

Volume 100, Chapter 100, Section 30, Sub-section 101

• UMTS BSC Release 0.1 System Architecture Specification


• UMTS RNC Release 0.1 System Architecture Specification


Please note that as these documents contain commercially sensitive information, they are not generally available to our customers. If you need more details in this area, please contact your Lucent representative, who will arrange to provide the required extracts from the documents.

5.4. Radio Resource Control software

The Lucent Radio Resource Control (RRC) software runs on the Radio Network Controller (RNC). The main functions of the RRC are to provide:

• Radio resource allocation

• Reverse outer loop power control

Radio resource allocation

Resource allocation within a UTRAN is based on a number of mobiles sharing a pool of physical channels. Access (CDMA) to the channels is controlled by Orthogonal Variable Spreading Factor (OVSF) codes.

Depending on the type of service requested by the mobile, and hence the channel data rate required, OSVF codes are selected from the appropriate level in the code tree (Figure 7 illustrates the code tree).

The Dedicated Channel (DCH) of each user is coded and multiplexed into one (or more) Dedicated Physical Data Channels (DPDCH). Additional rate matching is applied to balance the Eb/N0 (ratio of energy per chip to noise density) requirement of the different DCHs. The DPDCHs are weighted by a factor prior to multiplexing with the Dedicated Physical Control Channel (DPCCH), and spread to the chip rate of 3.94M chips s-1.

Variable spreading gain is used to realise different symbol rates, within the maximum DPDCH symbol rate for one DPDCH of 960kbs-1 .




Uplink

To decode the uplink transmission the BTS must receive each mobile at a level such that the Eb/N0 of its signal is above the minimum threshold.

Each mobile that transmits an uplink channel, produces an apparent rise in the noise floor. It is this effect that ultimately limits the capacity of the shared multi-user RF channel. The extent to which each mobile raises the noise floor is governed by its transmit power, activity factor (duty cycle) and data rate.

In addition to this mutual interference between mobiles within each serving cell, there is also a contribution to a rise in the effective noise floor caused by mutual interference from mobiles operating on the same RF channel in neighbouring cells. This inter-cell interference can be considered as part of the load of the serving cell.

The background noise floor is not constant, it comprises thermal noise and environmental noise, which varies slowly according to environment, time-of-day, and temperature. To obtain an accurate measure of the background noise, the level at the base station should be measured when no mobiles are in use.

The rise in the effective noise floor is used as a measure of the RF channel loading, and thereby to determine when the maximum traffic loading capacity has been reached. To allow a margin for overload prevention and coverage limitations, the entire theoretical RF channel capacity is not used. Usually, the capacity limit that causes a 3dB rise in the effective noise floor is used, which represents approximately 60% of the theoretical maximum capacity.

Mathematical models have been developed to represent these effects, and are used in traffic capacity design. This is discussed in the next chapter.

The following figures illustrate the impact of:

• Increasing traffic load on the consequential background noise rise

• Increasing data rate, required Eb/N0, and the resulting use of the RF carrier resource



101

Figure 20 Uplink RF resource capacity use

Downlink

A broadly similar process takes place for the downlink. Although in this case all the physical channels are transmitted from the same source, because more physical channels are transmitted, the effective noise floor will rise. This is due to a combination of:

• Path loss between mobile and base station

• Channel code orthogonality (both equipment limitation and propagation channel)

• Transmission power

• Inter-cell interference

• Thermal and environmental noise floor

Generally, for a service with a given bit rate and associated Eb/N0, the downlink will consume the same proportion of RF channel resources as it does on the uplink.

The transmitted power is measured at the BTS, which is a measure of the overall downlink RF channel load. As with the uplink, only a portion of the full theoretical capacity is used in order to avoid overload and coverage problems. Note that when a mobile’s connection is threatened by poor coverage (for example, low signal level), the mobile itself may autonomously handle the problem by means of lowering the transmission data rate once the maximum output power has been reached.




As the downlink OSVF code tree is shared by all physical channels, a channel code shortage can occur. For instance, when a high level code is allocated to a high data rate service, its derivative codes cannot be used for lower data rate services.

Figure 21 Downlink RF resource capacity use



103

Physical layer measurements

Radio resource management uses the following physical layer measurements:

Link Measurement Occurrence Purpose

Mobile to RNC Downlink Primary CCPCH Signal-Interference-Ratio Downlink Primary CCPCH Echip/N0

Event triggered Soft handover evaluation Downlink transmission power

BTS to RNC Downlink Primary CCPCH transmission power

Semi-constant Soft handover evaluation Downlink load estimation Downlink transmission power

BTS to RNC Uplink received signal strength

Periodic of interval 120 to 1000ms

Uplink load estimate

BTS to RNC Uplink background noise level

BTS initialisation (no users present)

Uplink load estimate

BTS to RNC Downlink transmitted signal level

Periodic of interval 120 to 1000ms

Downlink load estimate

BTS to RNC Uplink BER or BLER per transport block

Per 10ms frame Quality estimate for Reverse Open Loop Power Control

Table 14 Radio resource management physical layer measurements

Service dependent parameters

The radio resource consumption calculation is based on a single value of Eb/N0 for each service. The value takes account of the coding, interleaving, and rate matching involved.

Thus the Eb/N0 for each Radio Access Bearer (RAB) service is quoted for a specific Quality of Service (QoS), which may be defined in terms of BER, delay and so on, for a given radio environment.




The following table summarises typical Eb/N0 values for different services. The values are based on a vehicular environment with no base station receive diversity.

Application Reference service Uplink Eb/N0 /dB Downlink Eb/N0 /dB Voice 6kbs-1 voice

20ms interleaved 50% voice activity factor

5.4 9.2

Signalling 5.4 9.2

High quality video Low Constrained Delay (LCD) 384kbs-1 data

3.2 5.0

Low quality video Low Constrained Delay (LCD) 144kbs-1 data

2.7 5.4

Internet data Un-constrained Delay Data (UDD) 64kbs-1 data

3.6 6.0

Table 15 Example Eb/N0 values

Radio Resource Allocation functions

The main functions performed by the Radio Resource Allocation element of the Radio Resource Control software are:

Admission control

The purpose of admission control is to admit new mobiles, radio access bearers, and links (for example in handover) to the system, and to avoid overload and maintain stable operation by use of radio resource and interference measurements. Admission control is only invoked during initial mobile access, handover, and radio access bearer assignment. The access decision may be different in each of these cases, depending on the prevailing network conditions.

Congestion control

The purpose of congestion control is to detect and manage network conditions where overload is likely to occur, in a manner such that network stability is maintained, and the maximum number of subscribers are satisfied.

Congestion control is not supported in the Lucent prototype software release 0.1.

Dynamic radio access bearer control

This controls connection set-up and release in the UTRAN. The purpose of this function is to participate in processing the end-to-end connection set-up and release, and to manage and maintain the element of the end-to-end connection, which is located in the UTRAN.

In the first case, this function is activated by request from other functional entities at call set-up or release. In the second case, (when the end-to-end connection has already been established), this function may also be invoked to cater for in-call service modification or at handover execution.



105

The following Radio Resource Control procedures are Radio Access Bearer (RAB) service dependent.

• New Radio Access Bearer Establishment

Establishment based on QoS, assignment of Radio Link Control (RLC) parameters, multiplexing priority for the DTCH, scheduling priority for DCH, Transport Format Set (TFS) for DCH and update of Transport Format Combination Set (TFCS). It may also include assignment of a physical channel(s) and change of the used transport channel types / RRC state.

• Radio Access Bearer Release

This procedure may also release a DCH, which affects the TFCS. It may include release of physical channel(s) and a change of the used transport channel types / RRC state.

• Radio Access Bearer Reconfiguration

This procedure reconfigures parameters for a currently established radio access bearer or the signaling link to reflect a change in QoS. It may include changes in RLC parameters, multiplexing priority for DTCH/DCCH, DCH scheduling priority, TFS for DCH, TFCS, assignment or release of physical channel(s) and change of type of transport channel used.

Reconfiguration of the RAB is not supported by Lucent software release 0.1.

• Dynamic Channel Allocation

This function translates connection element set-up, reconfiguration or release requests into channel parameters, reserving or releasing the corresponding physical radio channels on the air interface.




Network element

Parameter RRC connection set-up

RRC connection clear-down

RAB establish ment

RAB release

Soft/softer handover radio link addition

Soft/softer handover radio link deletion

RNC to Mobile

TFCS * * *

RNC to Mobile

RLC Data * *

RNC to Mobile

DL Channel Code

* * * *

RNC to Mobile

DL Scrambling Code

* [1] * [1]

RNC to Mobile

UL Scrambling Code

*

RNC to Mobile

UL/DL Power Offset

* * *

RNC to Mobile

Frequency * [2]

RNC to BTS

TFCS * * * * [3]

RNC to BTS

DL Channel Code

* * * *

RNC to BTS

DL Scrambling Code

* [1] * [1]

RNC to BTS

UL Scrambling Code

* * [6]

RNC to BTS

UL/DL Power Offset

* * * * [3]

RNC to BTS

Frequency * [2] * [2][3]

RNC to BTS

DL Transmit Power

* * * [3]

RNC to BTS

UL Eb/N0 Target

* * [3]

Table 16 Parameters assigned by DCA for Lucent software release 0.1

[1] Used when multiple scrambling codes are used. Not supported in Lucent software release 0.1.

[2] Software release 0.1 supports only one RF channel frequency. [3] In soft handover the new BTS is informed of the current value (not necessary in softer handover).



107

Radio Access Bearer parameters

Transport Format Combination Set (TFCS)

This is in two parts: ‘dynamic’ and ‘semi-static’. The dynamic part of the TFCS (transport block size, transport block set size) determines the transport volume of the Radio Access Bearer (RAB). The semi-static part (transmission time, channel coding, rate matching) influences the transmission quality of the RAB.

Radio Link Control information

This contains two elements relevant to radio resource management, the type of Automatic Re-transmission Request (ARQ) (for example, selective repeat, hybrid), and the number of re-transmission attempts for the RAB.

Lucent software release 0.1 supports one fixed type of Radio Link Control (RLC) information which is used by all services that require ARQ. This is detailed in the UMTS RLC Layer Protocol Description (SDD-2) Volume 100, Chapter 50, Section 110, Sub-section 201.

In future releases the RLC information elements can be allocated depending on the subscriber service and environment.

Physical channel related parameters

Downlink channel codes

These Orthogonal Variable Spreading Factor (OVSF) codes are used to implement the CDMA to the common RF bearer, both for uplink and downlink.

Each node in the code tree defines a channel code corresponding to a Spreading Factor (SF) for the resultant spread spectrum transmission. However, not all codes within the code tree can be used simultaneously. A code can only be used for a physical channel if no other code on the path from the specific code to the root of the tree, or in the sub tree below the specific code, is used by another physical channel.

In the downlink the code tree is shared amongst all mobiles in the same cell. A matrix is generated of all codes and a field indicating whether or not the code is available. If a code is allocated to a physical channel, the code tree node itself and all nodes in the upward and downward direction in the code tree are indicated as not available. Several nodes are reserved for Common Channels and these are always indicated as unavailable.

When assigning codes from the code tree, the request is only accepted if there are enough codes available. If more codes are requested than are available, the request is denied even if the traffic load is below the admission threshold.

If multiple downlink scrambling codes are used (for example, multiple RF carriers) then multiple code trees are available for code allocation. However, this is not supported by Lucent software release 0.1.




Downlink transmitter power

The BTS is advised of the initial value for this for establishing a new radio link (such as RRC connection establishment, soft HO radio link addition). The initial value must be sufficiently high to ensure fast synchronisation of the new DPCH in the downlink (and uplink), but not so high that it unnecessarily contributes to the interference on the currently established downlink connections.

In case of soft handover, the new BTS is informed of the downlink transmission power currently used. During RAB establishment the downlink transmission power is readjusted when the Spreading Factor (SF) changes.

• As inner loop power control is already established there is no new calculated initial power. Instead a power offset from the currently used transmission power is calculated

• The BTS can calculate the value independently when the SF is known

• Inner loop power control can be used to track variations in the SF of 2

Downlink scrambling code

This must be allocated if multiple downlink scrambling codes are used. For example where more than one RF carrier is radiated. Lucent software release 0.1 does not support multiple scrambling codes.

Uplink channel code

The channel codes used for the uplink are predefined, and the Transport Format Combination Indicator (TFCI) determines the spreading factor.

Uplink scrambling code

When a mobile connection is established, the mobile is sent an uplink scrambling code that it uses for transmission over dedicated physical channels (including cases when only common channels are used). The uplink scrambling code remains valid until the RRC connection is released.

In a similar manner to downlink channel codes, uplink scrambling codes are managed using an allocation table in which scrambling codes are marked as to their availability. When a soft handover takes place the new BTS is advised of the scrambling code currently in use.



109

Uplink Eb/N0 target

To establish a new radio link (for example, RRC connection establishment, soft handover radio link addition) the uplink target Eb/N0 is sent to the BTS as an initial value for the inner loop power control algorithm. Once the connection is established, the outer loop power control algorithm takes control of maintaining the target Eb/N0.

The Eb/N0 target is based on:

• A minimum value of Eb/N0 determined by the detection reliability (error rate) of the Transport Format Combination Indicator (TFCI) and Transmission Power Control (TPC) bits

• An offset value for Eb/N0 that sets a threshold down to which the Reverse Outer Loop Power Control is allowed to reduce the Eb/N0, owing to good QoS, while maintaining the TFCI and TPC error rates.

A default offset value of 3dB is proposed for Lucent software release 0.1.

During soft handover the new BTS is advised of the modified Eb/N0 target.

Power offset between DPCCH & DPDCH (uplink/downlink)

The power control algorithms (closed loop and outer loop) maintain the Eb/N0 target on the DPCCH. However, it is the Eb/N0 of the DCH that determines the QoS of each RAB and hence the call quality.

This problem is addressed by applying a power offset between the DPCCH and DPDCH. The offset value is set so that an adequate Eb/N0 is maintained on the DPDCH without using excessive power. For example, a high data rate DPDCH has to be transmitted with a higher transmission power than the low data rate DPCCH.

Lucent software release 0.1 uses a fixed value for this power control offset (G) which is specific for the service combination used. The value of G for each service combination is detailed in UMTS Service Channel Multiplexing Algorithm Requirement Document (ARD 10) Volume 100, Chapter 50, Section 120, Sub-section 1001.

During soft handover the new BTS is advised of the power control offset value (G) that is currently in use for the uplink and downlink.

Operating RF channel (frequency)

Lucent software release 0.1 only supports operation on a single frequency.




Reverse outer loop power control

Closed loop (inner loop) and outer loop power control is used on the uplink and downlink. The exception is for common channels in the uplink, where open loop power control is applied to adjust the transmit power of the mobile.

The uplink power control adjusts the uplink transmitted power to ensure the same signal level is received from each mobile, and thereby (by use of the orthogonal channel codes) minimises the level of mutual interference.

The connection to a mobile may involve both a Serving Radio Network Subsystem (SRNS) and a BTS from a neighbouring or ‘Drift’ Radio Network Subsystem (DRNS). In this case the uplink outer loop power control (at the SRNC) sets the target quality level for the uplink inner loop power control function (at the BTS).

The downlink power control adjusts the downlink transmitted power on several radio links. It consists of inner and outer loop power control that resides in the mobile. The SRNC sends the target downlink power range based on the measurements reported from the mobile (so downlink outer loop power control is located in the mobile, but control parameters are set by the SRNC).

This concept is illustrated in the following figure (the figure applies to both the uplink and downlink).

Figure 22 Inner and outer loop power control



111

Summary

• Uplink. All power control functions reside in the SRNC for controlling the transmit power of the mobile

• Downlink. Control parameters are set by the Radio Resource Allocation software located in the SRNC

The inner loop adjusts the transmit power to meet a target Signal-to-Interference Ratio at the receiving station. The Eb/N0 value is not directly mapped to the QoS, as the target Eb/N0 is adjusted by the outer loop power control to compensate for environmental factors (such as shadowing and mobile terminal speed). Thus the outer loop power control is primarily responsible for controlling the required QoS.

Further reference – power control system

Detailed information about the power control system, including the algorithm flow-chart, is contained in the UMTS Service Channel Multiplexing Algorithm Requirement Document (ARD 10) Volume 100, Chapter 50, Section 120, Sub-section 1001. Supporting information is also available in the UMTS Inner Loop Power Control Algorithm Requirement Document (ARD 1) Volume 100, Chapter 50, Section 120, Sub-section 101.


Power control parameters

The principal reverse link (uplink) open loop power control parameters are listed below.

• Threshold for reporting QoS worse than its acceptable limit (BERmax)

- Set by the maximum BER for the specific service (BERrequired)

- Therefore (BERmax) = (BERrequired)

- Values for the maximum BER (BERmax) against types of services are given in the UMTS BTS UBTS System Requirement Document Volume 100, Chapter 100, Section 100, Sub-section 101

• Threshold for reporting QoS better than its acceptable (BERmin)

- Proposed to be such that lg (BERmin) = [ lg (BERmax) – 3 ]

• Step size to increase the target Eb/N0

- (Eb/N0)up is proposed to be 0.5 or 1dB

• Step size to decrease the target Eb/N0




- (Eb/N0)down is proposed to be 0.5 or 1dB

• Value for the lower limit of the target Eb/N0 set by the Radio Resource Allocation software

- This is min(Eb/N0)ctrl

• Value for the upper limit of the target Eb/N0 set by the Radio Resource Allocation software

- Proposed to be max(Eb/N0)ctrl = { [ (Eb/N0)min ] + [2 * ∆ (Eb/N0)ctrl ] }

Radio Resource Control software – further reference

For detailed information regarding the Radio Resource Control software, refer to the UMTS Service Channel Multiplexing Algorithm Requirement Document (ARD 10) Volume 100, Chapter 50, Section 120, Sub-section 1001.

Please note that as this document contains commercially sensitive information, it is not generally available to our customers. If you need more details in this area, please contact your Lucent representative, who will arrange to provide the required extracts from the document.

5.5. Handover

Handover is the process by which one or more radio links between a mobile and a UTRAN may be added or removed when an RRC connection exists and it is known in which cell the mobile is located.

In addition to the established inter-cell and intra-cell handover, a UTRAN exhibits the following additional handover types and related features.

• Hard handover

This procedure removes all existing radio links before new links are established, for instance when an inter-frequency handover occurs. Lucent software release 0.1 does not support this facility.

• Macro diversity

Macro diversity is effected when the mobile has radio links with two or more UTRAN access points simultaneously with the aim of improving the radio connection quality or providing seamless handover.

• Mobile evaluated handover

In this case handover is triggered by an evaluation made in the mobile. The mobile determines the necessity of handover based on the measured radio environment and criteria defined by the network. When the evaluation meets the hand-off criteria the necessary information is sent from the mobile to the network. The network then decides on



113

the necessity of the handover based on the reported evaluation result and other conditions such as the uplink radio environment and/or availability of network resources.

• Network evaluated handover

In this case handover is triggered by an evaluation made in the network. There are three cases.

- The mobile measures and reports the measurement to the network on request from the network either periodically or on demand. The network then evaluates the necessity of handover

- The network measures and evaluates the necessity of handover

- Both the mobile and the network make measurements

In all cases, the network decides the necessity of handover based on the measurements and other conditions such as availability of network resources. The network always executes the handover.

• Soft handover

In soft handover, radio links are added and dropped so that the new link is added before the old one is lost. During soft handover the mobile maintains at least two radio links with the UTRAN. The term soft handover is generally used for inter-cell handover. The term softer handover is generally used for intra-cell handover.

• Inter-system handover

This allows handover between two public cellular radio systems with different but compatible radio air interface standards. For example, between UTRAN and GSM or UTRAN and GSM-GPRS. Lucent software release 0.1 does not support inter-system handover.

• UTRAN Registration Area (URA) updating

This procedure updates the UTRAN registration area for a mobile when a RRC connection exists and the position of the mobile is known on the URA level in the UTRAN.

The Cell Update and URA Update procedures are used for common channel handover. They are not supported by Lucent software release 0.1.

Measurement reporting

Accurate handover decisions rely on signal measurements made by the mobile and BTS. These may include:

• Received signal strengths (serving and neighbour cells)




• Estimated bit error rates (current and surrounding cells)

• Estimated propagation environments (for example, high-speed, low-speed)

• Transmission range (for example, from timing information)

• Estimated speed (for example, Doppler shift)

• Synchronisation status

• Received interference level

• Total downlink transmission power per cell

For these measurements to be meaningful, they must be associated with the channels to which they relate. This association may include the use of identifiers for the network, the base station, the cell (base station sector) and/or the radio channel. This function is located both in the mobile and in the UTRAN.

Measurement cell sets

Handover preparation measurements are performed for different sets of cells. Each set comprises a list of cells that the mobile should monitor for a given period of time.

• Handover Monitoring Set

This lists the cells that the mobile should monitor when it is operating in active mode. The cells may be part of a UTRAN, or from another network using a GSM air interface.

• Active Set

This lists the UTRAN cells that are currently assigning a downlink DPCH to the mobile, and corresponds to the cells with which the mobile is in soft handover.

• Handover Candidate Set

This lists cells that are not currently in the active set but have been received by the mobile at sufficient strength to indicate that their associated DPCHs could be successfully demodulated. These cells may or may not be on the same radio channel (frequency) as the current serving cell(s), and may form part of the UTRAN or a neighbouring GSM network.

Assignment of cells to measurement cell sets

Cells are assigned to measurement cell sets as follows:

1. The RNC sends the Handover Monitoring Set to the mobile to monitor the neighbour cells and to report when the reporting criteria have been met, by use of the Measurement Control Messages. The measured cells may form part of a UTRAN or GSM network. Lucent software release 0.1 only supports soft handover to UTRAN cells operating on the same radio channel (frequency).



115

When the mobile reports that there is a cell to which it may handover, this handover candidate cell is added to the Active Set or the Candidate Set, depending on the availability of free active links. At the same time, the RNC instructs the mobile periodically to monitor the new cell (measurements are periodic for Active Set and Candidate Set cells).

2. Cells in the Active Set are moved to the Handover Monitoring Set when their received signal level falls below a threshold given by: (Received Level of Strongest Link) – (margin δdrop)

If the received signal level has not fallen below the above threshold, but other cell(s) in the Candidate Set are received at a higher level, the cell is moved from the Active Set to the Candidate Set.

3. As in Step 2, a cell in the Candidate Set will be transferred into the Handover Monitoring Set if its received signal level falls below a threshold given by:

(Received Level of Strongest Link) – (margin δdrop)

If the received signal level from a Candidate Set cell is greater than that of one of the Active Set links, then this link from the Candidate Set will replace the weak Active Set link.

Handover preparation measurements from UTRAN (FDD)

The measurements used for handover preparation form part of the radio resource control software detailed in the UMTS Radio Resource Control Algorithm Requirement Document (ARD 6).

• Measurements on the same radio frequency as the serving cell(s)

The mobile synchronises with the cells that are to be measured using the primary and secondary synchronisation channels and knowledge of the scrambling codes used by neighbouring cells. The mobile performs the following measurements:

- Signal level of the measured cell. Downlink primary CCPCH SIR. Downlink Primary CCPCH (Echip / No). (These measurements may change as part of the 3G harmonisation process)

- Relative timing between the cells. Measured from the phase difference between the scrambling code, which is directly related to the timing difference between the cells

• Measurements on different radio frequencies from the serving cell(s)

Lucent software release 0.1 does not support monitoring of cells which operate on different radio frequencies to the serving cell(s).




Measurement messages

The Radio Network Controller (RNC) dynamically updates the measurement information on the individual cells listed in the three cell sets. It is assumed that the measurements are made by the mobile.

Intra-frequency Measurements and Traffic Volume Measurements are defined in the UMTS Radio Resource Control (RRC) Protocol Description (SDD-3) as mandatory for the mobile. The measurements to be provided by the mobile are defined in the 3GPP UMTS 25.302 Services Provided by the Physical Layer document. The measurements required by Lucent software release 0.1 are given in Table 1 of the UMTS Radio Resource Control Algorithm Requirement Document (ARD 6).

The RNC defines the measurements that the mobile should make by sending a Measurement Control message to the mobile. This also specifies the reporting mode, which may be periodic, event-triggered, or immediate reporting. Periodic measurement reports are used for cells in the Active Set or Candidate Set. Event triggered measurement reports are used for cells in the Handover Monitoring Set.

The mobile sends a Measurement Report message to the UTRAN when the reporting criteria specified in the UMTS Radio Resource Control (RRC) Protocol Description (SDD-3) are met.

Lucent software release 0.1 uses the following event triggered reporting criterion:

[Received Level (new link)] > [Received Level (strongest active link)] – [margin δadd ] where: Received Level is either Downlink Primary CCPCH SIR or Downlink Primary CCPCH (Echip / No)

In future software releases it may be possible, by using predefined (default) reporting criteria, for the mobile to report measurements without being instructed to do so. 1

1 Note for readers familiar with the US IS-95 system: IS-95 mobile measurements are not reported periodically but only following certain events when the reporting criteria are met. As the handover algorithm has been standardised the mobile itself can decide when to add / remove radio links. It performs measurement averaging, maintains the necessary timers and has knowledge of the handover algorithm. However, UMTS mobiles do not have such knowledge of the handover algorithm.



117

Measurement performance

Measurement averaging is necessary to overcome short-term shadowing effects. However, this is not straightforward, as the reporting intervals may be of unequal and unpredictable length when event triggered measurement reports are used.

The current working assumption is that the measurements are reported every 1s.

In future software releases the following additional measurements may prove useful, particularly in multi-layer networks:

• Estimated bit error rate (serving and neighbour cells)

• Estimated propagation environments (high-speed, low-speed, etc.)

• Transmission range (from timing information)

• Speed (from Doppler shift)

• Synchronisation status

Soft (and softer) handover algorithm

This section outlines the handover algorithm used for soft (inter-cell) and softer (intra-cell) handover. This section is based on the Handover Algorithm Requirement Document (ARD 5), which contains detailed information and flow charts of the algorithms concerned.

Assumptions

• The Soft Handover (SHO) algorithm is activated on every reception of a measurement report

• The RNC determines whether to add new links and/or to remove existing ones. The mobile is informed via the Active Set Update Command replacing the entire current Active Set

• The Active Set may have up to 4 entries, so the maximum number of soft handover links is 4

If the Active Set is limited to 2 or 3 entries, performance may degrade owing to link preservation timers maintaining weak links for too long. This may occur if upcoming new links with higher power than the active links cannot be added to the Active Set because the preservation timers of weak links have not yet expired.

Also, if the Active Set is too small, very frequent handover (links frequently dropped and added) takes place producing an unnecessary and significant signalling overhead. Apart from the signalling load disadvantage, performance may be degraded owing to the relatively long time to set up a new link (~1s), during which time receiver performance will suffer.




There is a clear trade-off when maintaining many active links in the downlink. Network capacity is reduced owing to increased interference, but network link quality is improved and fewer calls are dropped.

• A restriction on the maximum number of Active Links according to the User Bit Rate may also be required for similar interference / capacity reasons. Consequently, the RRC algorithm has to reduce the bit rate before SHO (so SHO is restricted to those mobiles not exceeding the limit). However, this is a potential problem for the RRC software and has no impact on the SHO algorithm. The solution depends on whether it is necessary to support high bit rate users at the edge of a cell

• Both the RRC and outer loop power control software influence the handover decision

• In future software releases the SHO algorithm may allow the maximum size of the active set be altered according to the required QoS, with new links only added if they substantially improve the received signal quality

General algorithm description

A new radio link from the mobile to a neighbouring BTS is made when the average received level of the BTS becomes higher than the received level for the strongest existing radio link less a margin of (δadd ).

The link to a weak BTS is deleted when the average received signal level of the BTS drops below the average received signal level of the strongest existing radio less a margin of (δdrop ). A link is removed only if the average received level remains below the specified level for a certain time period called the Link Preservation Time.

The thresholds that control when to add or drop a link are set dynamically. The second, third and fourth link are only added when they substantially contribute to the transmission quality of the connection.

For example, it may be reasonable to specify a margin of (δadd ) = 6dB for the second link. However, with two active links with similar received power, for the addition of a third link the margin should be higher because a new link with 6dB lower power than two strong links will contribute less to improving transmission quality than when there was only one link.

The SHO algorithm comprises three elements:

• Measurement processing

Measurement Reports giving the Downlink Primary CCPCH SIR and Downlink Primary CCPCH (Echip / No) are received periodically from all cells in the Active Set

The SHO algorithm also requests periodic measurements for defined Handover Candidate Set cells

• Link deletion



119

The Radio Network Controller periodically triggers the link deletion check procedure. The network operator can set the interval (rate) from the value range 20-5100 ms (up to 255 frames). The default value is 500 ms.

During this procedure active links might be removed and possibly replaced by Candidate Set cells.

• Link Addition

The link addition procedure is triggered by a measurement report from a mobile station indicating that a neighbour cell has been received at a sufficient level to be added to the Active Set.

During this procedure either the requested link is established (if there are not already four entries in the Active Set) or the neighbour cell is added to the Candidate Set.

Each of these three elements is described in more detail in the following sections.

Measurement processing

Measurement processing is conducted in the following stages:

Stage 1

Periodic received signal level measurements from the Active Set and Candidate Set cells are received.

Stage 2

The received measurements are averaged using a negative exponential window with operator selected history weighting, having a default value of 0.9:

(new_value) = (old_value) * (history_weight) + (1 – [history_weight]) * (current_value)

The soft handover algorithm is based on the received signal level, quality is controlled by the power control and radio resource management software.

Stage 3

All Active Cell links and Candidate Set cells are ordered according to decreasing received signal level, in order to identify the strongest and weakest links/cells.




Link deletion

Link deletion is conducted in four main stages:

Stage 1 – all links

Link deletion thresholds are checked for all Active Set and Candidate List cells.

• For each link, the difference between the receipt level of the strongest active link and the received level of the link under consideration is calculated

• The link is determined to be too weak if the difference is more than a margin of (δdrop ), where (δdrop ) is user adjustable within the range 0 to 16dB (default value 6dB). For optimum performance, ultimately the margins will be different for the second, third and fourth link. In Lucent software release 0.1 the value of (δdrop ) will be fixed for all links

• When the level is determined to be too weak, the Link Preservation Timer is started, or incremented if already running. The timer has a range of 20 to 5100ms (255 frames), with a default value of 2s

• When the timer expires the Deletion Flag is set and the link is dropped

• When the level is determined not to be too weak, the Link Preservation Timer is reset. This is always done for the strongest link, and may be done for the other links that are above the deletion threshold

Stage 2 – candidate cell conversion

Active Links may be converted to Candidate Cells, providing they have not fallen below the threshold received signal level, if new higher signal level cells are received, and placed in the Active List, which thereby reaches its maximum (4).

• For each Candidate Cell a timer is started to ensure the link remains a Candidate for a certain minimum time. The timer has the range 20 to 5100ms (up to 255 frames), with a default value of 1s

• Once the Candidate timer expires, a new Active Set is created implicitly by marking the weakest of the current Active Links for Candidate Conversion. The only Active Links which are not marked for Candidate Conversion, are those already marked for deletion or that are among the 4 strongest Active and Candidate Cells

• The Candidate timer is reset

Stage 3 – delete links marked for deletion

Any links marked for deletion in Stages 1 and 2 are deleted. This also includes Candidate Cells that have to be deleted only from the measurement processing procedure. For deleted Candidate Cells the reporting criterion is reset from periodic reporting to the standard criterion for neighbour cells.



121

Links marked for Candidate Conversion are also deleted from the Active Set, but remain as Candidate Cells in the measurement processing procedure using the periodic reporting criterion.

Stage 4 – create active link

If Active Links have been deleted and there are any Candidate Links, Active Links to the strongest Candidate Cells are established until either the maximum number of Active Links is reached or all Candidate Cells have been handled.

The Candidate Timer is reset if at least one Active Link to a former Candidate Cell has been established.

Link addition

Link addition is conducted as follows:

• The mobile reports that a signal from a cell has been received at a level higher than that from the current strongest link less a margin (δadd ). The parameter (δadd ) is sent to the mobile for the measurement evaluation and reporting by use of the Measurement Control message

• The current working assumption is that at most 4 active links are supported, (4-way soft handover). If there are not 4 entries in the Active Set list (that is, there is a free link), the new link is added to the Active Set, and the reporting criterion is set to periodic measurement for the received signal level

• If the Active Set list is full (no free links), the link is added to the Candidate List, and the reporting criterion is set to periodic measurement for the received signal level

• Although an optimised algorithm will ultimately use different margins for the second, third and fourth link, Lucent software release 0.1 uses a fixed value of (δadd ) for all links. The default value is 4dB




Hard handover algorithm

Software release 0.1 does not support Hard handover (for example, inter-frequency handover).

UTRAN – GSM handover algorithm

Software release 0.1 does not support handover between UTRAN and GSM networks.

UTRAN – GSM GPRS handover algorithm

Software release 0.1 does not support handover between UTRAN and GSM GPRS networks.

Handover control software – further reference

For detailed information regarding the Handover Control software, refer to the following documents:

• 3GPP UMTS 25.231 Physical Layer Measurements

• 3GPP UMTS 25.302 Services Provided by the Physical Layer

Please note that the following documents contain commercially sensitive information, and are not generally available to our customers. If you need more details in this area, please contact your Lucent representative, who will arrange to provide the required extracts from the documents:

• UMTS Radio Resource Control Algorithm Requirement Document (ARD 6) Volume 100, Chapter 50, Section 120, Sub-section 601

• UMTS Radio Resource Control (RRC) Protocol Description (SDD-3) Volume 100, Chapter 50, Section 110, Sub-section 301

• UMTS Handover Algorithm Requirement Document (ARD 5) Volume 100, Chapter 50, Section 120, Sub-section 501



123

5.6. Lucent equipment capacity

This section lists the principal capacities for Lucent’s current implementation of UTRAN equipment.

Hardware (Lucent Network Release 1.0)

• The Macrocell BTS (3 carrier tri-sector) is not yet available

• A tri-sectored milli-cell BTS supports:

- 2 carriers @ 20W with 11 baseband units

- 256 voice circuits with BBU32 cards, and 608 with BBU64 cards

- Aggregate data capacity 8.192Mbs-1 with BBU32 cards, 9.728Mbs-1 with BBU64 cards

- The air interface capacity is ~120 voice circuits or 1Mbs-1 per sector per carrier

• A single sector microcell BTS supports:

- 1 carrier @ 5W

- 16 voice circuits with ASIC1, and 48 with ASIC2

- Aggregate data capacity 1.024Mbs-1 with BBU32 cards, 3.072Mbs-1 with BBU64 cards

- The air interface capacity is ~120 voice circuits or 1Mbs-1 per sector per carrier

• RNC capacity:

- Maximum number of links to BTSs: 160 E1 or 4 STM-1 or 1 STM-1

- Circuit switching load of 4,000Erl of voice channels, with expansion to 10,000Erl

- Up to 6 soft hand-off connections per mobile

Software (Lucent Release 0.1)

• Up to 4 soft handover links

• Hard handover not supported (no inter-frequency handover)

• Inter-system handover not supported (no UTRAN to GSM handover)


RF Network Coverage & Capacity Design 6


125

6. RF Network Coverage and Capacity Design

This chapter discusses network deployment planning from a theoretical RF perspective, using mathematical modelling and computerised coverage prediction software. It covers the following topics:

• Frequency planning

• Code assignment parameters

• Air interface link power budget

• Area coverage prediction

• Traffic load capacity

• Service type and quality

• Inter-system boundaries




6.1. Frequency planning

The frequency band allocated for terrestrial FDD UTRANs is a duplex pair of 60MHz blocks:

• 1920 to 1980MHz uplink mobile transmit (60MHz band, duplex separation 190MHz)

• 2110 to 2170MHz downlink base transmit (60MHz band, duplex spacing 190MHz)

Frequency planning criteria

The UTRAN radio channel spacing is ~5MHz, so 12 UTRAN channels can be accommodated in the allocated 60MHz band.

To date, few countries have announced their plans for issuing UTRAN licences and the number of channels likely to be assigned to network operators has not yet emerged. However, it is a reasonable assumption that countries will licence a similar number of UTRANs as for second generation networks. In European countries this is typically 3 or 4 national networks.

In the case of a country licensing 4 equally sized national networks, each UTRAN can only be assigned a maximum of 4 unique radio channels, on a sole use basis.

FDMA concepts

Mobile radio systems based on a FDMA cellular concept usually assign a small set of frequencies to each of their cells, from the large pool provided by their licence. The same small set of frequencies is then reassigned to another cell some distance away from the first, so that the mutual interference level is acceptable. This is known as ‘frequency reuse’ and is normally based on a regular pattern. For example:

• A group of 12 frequencies, used at tri-sectored base stations, with a 4 site x 3 cells per site basis, gives a reuse pattern of 4/12

• A group of 9 frequencies, used at tri-sectored base stations, with a 3 site x 3 cells per site basis, gives a reuse pattern of 3/9

Thus each radio frequency is not used at every cell.

CDMA concepts

When using a wideband DS-SS CDMA air interface such as that used by UTRANs, it is possible to use the same radio frequency in adjacent cells (a reuse pattern of 1/1), provided that the traffic loading of each site is limited to ~65% of its theoretical maximum capacity.

For example, when a high capacity site is required, the Lucent Milli-cell BTS can be used to provide a tri-sectored cell radiating two carriers (radio frequencies) in each sector.

The introduction of UTRANs is not directly comparable to the introduction of CDMA networks in America. In the case of CDMA in America, 1.25MHz CDMA channels had to co-exist in the same band as 30kHz FDMA channels, and protection for the existing services was necessary.



127

UTRANs do not share their band with other services, subject to any particular limitations in the territory concerned, and all channels can normally be used without the need to introduce protection for existing systems.

Example UTRAN band assignment – United Kingdom

This section outlines the UK band plan. As the UK was among the first countries to announce its licensing plans, it is not yet known how typical of band plans it will be. Industry commentators have expressed doubts about the viability of issuing five licences to cover a population of about 57 million over a land area of 245,000 sq. km. Similar countries are expected to issue fewer licences.

The UK Government has offered 20-year licences allocated as follows:

• Licences C, D, and E for 2x10 MHz paired spectrum plus 5 MHz unpaired spectrum

• Licence B for 2x15 MHz paired spectrum

• Licence A of 2x15 MHz paired spectrum plus 5 MHz of unpaired spectrum. This is reserved for a company that does not operate an existing ETACS or GSM network in the UK




The UK band plan for UTRANs is illustrated in the following diagram:

Figure 23 Example of UTRAN band plan

The current intention is that spectrum in the 2010 - 2025 MHz band will not be used.

The UK licences were issued to the following operators and at the following prices:

Licence A £4.38 billion TIW UMTS (UK) Licence B £5.96 billion Vodafone Licence C £4.03 billion BT (3G) Licence D £4.0 billion One2One Personal Communications Licence E £4.09 billion Orange 3G

The licences will remain in force until 31st December 2021 unless earlier revoked by the Government or surrendered by the licensee.



129

6.2. Code assignment parameters

As each UTRAN will only be assigned a small number of radio channels (2 or 3 in the UK) each frequency may be reused in every cell. Therefore the mobile will probably be in range of a number of co-channel cells.

Primary Synchronisation Channel Code

During its initial cell search a mobile searches for the cell to which it has the lowest path loss, it then uses the Primary Synchronisation Channel (PSCH) to acquire slot synchronisation. This is performed using a single filter matched to the Primary Synchronisation Code (cp).

The Primary Synchronisation Code (cp) is the same for all cells of the same UTRAN.

Secondary Synchronisation Channel Code

The mobile then uses the Secondary Synchronisation Channel (SSCH) to determine frame synchronisation and identify the code group of the base station it has received. This is achieved by correlating the received signal at the positions of the Secondary Synchronisation Code (cs) with all 16 possible Secondary Synchronisation Codes. The position of the Secondary Synchronisation Code is calculated by measuring the timing offset between the Primary and Secondary Synchronisation Codes, and the un-modulated PSCH can be used as a phase reference for demodulating the SSCH.

• The 16 different Secondary Synchronisation Codes designate 16 demodulation sequences

To achieve frame synchronisation:

• The 16 demodulation sequences are correlated with the 16 different cyclic shifts of the Secondary Synchronisation Channel modulation sequence (d1, d2, d3, …d16). This gives 256 (16 x 16) different correlation values

This means that up to 256 co-channel (same frequency) carriers can be uniquely identified and synchronised.

At this stage, by identifying the Secondary Synchronisation Code (1 of 16) and the Timing Shift (1 of 16), the Code Group (1 of 256) and frame synchronisation is determined.




Scrambling code

The final stage is for the mobile to determine the scrambling code used by the identified base station. The scrambling code is identified by symbol-by-symbol correlation over the Primary CCPCH with all the scrambling codes used within the Code Group identified above.

The start of the scrambling code is calculated from the frame boundary found in the preceding stage.

The correlation must be conducted on a symbol-by-symbol basis because of the unknown modulation of the Primary CCPCH.

Once the scrambling code has been found, the Primary CCPCH can be detected, super-frame synchronisation acquired, and all the cell and system information read from the BCH.

Sixteen scrambling codes are used in each Code Group.

Code assignment summary

For each BTS there are three codes to assign:

• Primary Synchronisation Code (cp). The same for all cells of the same UTRAN

• Secondary Synchronisation Code (cs). Choice of one code from 16; designates one of the 16 demodulation sequences

• Scrambling Code. Choice of one code from 16

These define 256 possible unique BTS identities for each Code Group.



131

6.3. Air interface link power budget

Background

In FDMA TDMA systems such as GSM, the part of the link power budget available for path loss is fixed, and is based on the following (apart from set system losses):

• Maximum Effective Isotropic Radiated Power (EIRP) per user channel

• Minimum Discernible Signal (MDS) power (receiver ‘sensitivity’)

• Minimum Carrier-to-Noise Ratio (CNR)

• Minimum Carrier-to-Interference Ratio (CIR)

These factors are constant for each link under consideration, regardless of the volume and type of traffic carried by the link. This is a result of using a fixed modulation and coding scheme for all traffic (control, voice and data), and a fixed inter-user interference level by means of a set FDMA reuse pattern and TDMA. In practice, the inter-user interference level may vary with co-channel traffic load at neighbouring sites where the same frequency is reused, although frequency hopping may be used to average this effect.

Effect of coding scheme on link power budget

In a CDMA system such as the UTRAN, the basic elements of the link power budget that determine the part of the link budget available for path loss are similar:

• EIRP per user channel

• MDS power

• Minimum signal bit energy-to-noise (Eb/N0) ratio (similar to the CNR in a FDMA TDMA system)

• Minimum signal bit energy-to-interference (Eb/I0) ratio (similar to the CIR in a FDMA TDMA system)

However, unlike FDMA TDMA voice and circuit switched data systems, the UTRAN CDMA air interface can provide different data rate channels according to the type of service the user requests (for example, voice, email, web, video). Each channel has a different data rate and associated coding scheme (please see Chapter 4 for details).

The extent to which the coding scheme protects against errors and allows for their correction also has a major effect on the required values of (Eb/N0) and (Eb/I0).

Another factor that influences the performance of the coding scheme is the severity of the multi-path environment, and the rate and length of nulls in the received signal caused by




movement through it. Thus the type of propagation environment, and the speed of the mobile has a significant effect on the required values of (Eb/N0) and (Eb/I0).

To simplify the air interface link power budget calculation, the above effects have been combined to form a default value for the (Eb/I0) that has to be met (or exceeded) for successful demodulation and data decoding.

The current default values for (Eb/I0) for each type of user channel for the Vehicular A 120kmh-1 and Pedestrian ‘A’ 3kmh-1 Propagation Channels are given in the following tables.

Important: The values in the tables are for illustration only. For current recommended values, Lucent personnel should refer to the default parameters given in the UMTS RF Dimensioning Tool – Input Parameter Guidelines document. This is updated frequently and is available at: http://en0033svr06.uk.lucent.com/rfsystems/source/RF_DimensioningTool_Inputs_v1.1.doc

Service 8k voice

13k voice

16k LCD data

32k LCD data

64k LCD data

144k LCD data

384k LCD data

64k UDD data

144k UDD data

384k UDD data

User Bit Rate /bs-1

8k 13k 16k 32k 64k 144k 384k 30.4k 60.88k 243.2k

Uplink (Eb/I0) /dB

5.3 6.1 3.8 3.1 3.8 3.1 3.1 3.8 3.0 2.4

Downlink (Eb/I0) /dB

7.9 8.5 6.0 5.2 6.0 5.2 5.5 5.3 5.2 4.3

Service Type

Cct Cct Cct Cct Cct Cct Cct Pkt Pkt Pkt

Table 17 Required (Eb/I0) for Vehicular ‘A’ 120kmh-1 propagation channel

[1] Cct indicates a circuit switched service [2] Pkt indicates a packet switched service

http://en0033svr06.uk.lucent.com/rfsystems/source/RF_DimensioningTool_Inputs_v1.1.doc



133

The default values for (Eb/I0) for each type of user channel for the Pedestrian ‘A’ 3kmh-1 Propagation Channel are given in the following table.

Service 8k voice

13k voice

16k LCD data

32k LCD data

64k LCD data

144k LCD data

384k LCD data

64k UDD data

144k UDD data

384k UDD data

User Bit Rate /bs-1

8k 13k 16k 32k 64k 144k 384k 30.4k 60.88k 243.2k

Uplink (Eb/I0) /dB

3.3 4.1 2.4 1.7 1.3 1.5 0.7 0.4

Downlink (Eb/I0) /dB

6.1 6.7 4.2 3.8 3.4 3.5 3.4 2.4

Service Type Cct Cct Cct Cct Cct Cct Cct Pkt Pkt Pkt

Table 18 Required (Eb/I0) for Pedestrian ‘A’ 3kmh-1 propagation channel

[1] Cct indicates a circuit switched service

[2] Pkt indicates a packet switched service

Link power budget elements

This section briefly describes each of the principal elements of the link power budget calculation:

• Information data rate

• Transmitter power

• Cell loading

• Antenna systems

• Noise and interference

• Receiver sensitivity

• Receiver minimum discernible signal

• Link power available for path loss




Information data rate

This is the data rate of the source (user or baseband data). For example, if the user requires a circuit that can support a continuous throughput of 64kbs-1, the information data rate is 64kbs-1.

The information data rate forms part of the link power budget calculation because the UTRAN uses Spread Spectrum (in this case Direct Sequence) CDMA to a common RF channel. The user data is spread over a wider RF bandwidth than the modulation scheme requires, thereby yielding processing gain when the signal is de-spread at the receiver.

In a Direct Sequence-Spread Spectrum system, the processing gain is determined by the ratio of the RF bandwidth (of the transmitted signal) to the baseband information rate. For the QPSK modulation used in a UTRAN, the noise bandwidth is equal to the code bandwidth and the processing gain can be considered as equal to the ratio of the chip rate to the information rate.

Because the chipping rate is fixed in Direct Sequence-Spread Spectrum systems, the processing gain is determined by the baseband information data rate. Since the UTRAN provides different types of channels with various associated user data rates, with the fixed chip rate, the receiver produces a different processing gain according to the user data rate.

Data rates of 32 (for voice), 64, 144, and 384kbs-1 are available, and the duplex circuits need not be symmetrical. CDMA access to the common RF carrier is achieved by means of Orthogonal Variable Spreading Factor (OVSF) Codes from the code tree described previously. It is therefore necessary to include the baseband data rate in the link budget power calculation, in order to incorporate the correct value for the processing gain. The default value depends on the type of service - 64kbs-1, 144kbs-1, and so on.

This is an 8kbs-1 voice circuit example:

Base transmitter downlink

Mobile transmitter uplink

Information data rate /kbs-1

8 8



135

Transmitter power

In a narrow band FDMA TDMA system, normally one transmitter serves only one user circuit (such as a voice call) at a time. So at any one instance (such as a time slot) the whole power output is available for providing a link to a single user (for carrying voice or control data).

In a wide band CDMA system, one transmitter serves many simultaneous user circuits. So at any one instance the power output available for providing a link to a single user is divided between both the number of users involved, and the number of active downlink common control data channels.

Another feature of CDMA systems that must be considered is that of soft handover. As a mobile may be served by more than one cell (transmitter), a proportion of the downlink power transmitted by each cell is used (coded for reception) by a mobile executing soft handover. As such, this proportion of power is not available for use by the other mobiles served by the transmitter (cell).

Consequently, the following values must be entered into the RF link power budget calculation:

• Total transmitter output power

Default value +43dBm for base station, and +21dBm for mobile.

• Proportion of power used for common control channels

Default value 8% for base station. That is, there is an 8% power ‘overhead’ to allow for common control channels. This results in the power available for use by mobiles in the cell being reduced by a factor of 1.08 (0.334dB).

Default value 0% for mobile, as this situation only applies to the base station transmitter.

• Proportion of power used for soft handover

Default value 65% for base station. That is, there is a 65% power ‘overhead’ to allow for soft handover. 65% of mobiles are not in soft handover, but some may be in multi-way handover.

This results in the power available for use by mobiles in the cell being reduced by a factor of 1.65 (2.5090dB).

Default value 0% for mobile, as this situation only applies to the base station transmitter.




Using the default values, the transmit power available for the link between the base station and mobiles is:



Transmitter output power /dBm 43 21

Proportion of power used by common control channels /dB

-0.3342 0

Proportion of power used by soft handover /dB

-2.1748 0

Resulting power available for all traffic channels /dBm

40.4910 21

Table 19 Transmit power

Cell loading

The power available for all traffic channels calculated above, is divided amongst all mobiles in the cell which use the same RF channel. So it is necessary to specify the maximum number of mobiles to be served in the cell with the same RF carrier, before the power available to each one can be calculated.

When determining cell loading, the “pole capacity” is identified first. This is the theoretical maximum number of users that can simultaneously use the same RF channel. It is a function of the traffic channel data rate, and hence the type of service used. Typical pole capacity values have been found by simulation. For example, 125.7 for downlink 8kbs-1 voice circuits.

Having established the theoretical maximum capacity, the proportion of the maximum theoretical cell loading used is required. A default value of 50% is used.

Once a traffic channel has been assigned, it may not be used continuously. For example, during voice calls both parties do not normally talk at the same time for any length of time. The channel may only carry traffic for less than 50% of the time, thereby reducing the need to transmit data and the associated use of RF power (and inter-user interference).

For voice circuits, a default value of 50% is used as the proportion of time the channel is used.

Initially, the power available to all mobiles can be divided by the number of mobiles within the cell, by knowledge of the pole capacity (for the specific type of service) and the maximum Designed Cell Loading Proportion. However, this can lead to a pessimistic result, as mobiles close to the base station require relatively little power, releasing more for mobiles located at a greater distance. Also, close mobiles receive signals with better orthogonality than distant mobiles, again reducing their need for power.

A correction factor is therefore used to represent the ratio of the power that is likely to be available to mobiles at the edge of the cell, compared to that which would be available by evenly dividing the available power (regardless of the location of the mobiles).



137

The value of this edge of cell power correction factor has been evaluated by simulation, assuming a uniform distribution of mobiles:

• Default value 1.65 for base station – the power available to mobiles at the edge of the cell is 1.65 times that which would be available if all the power was evenly divided between all mobiles within the cell

• Less power is needed by mobiles that are close to the base station

• Default value 0% for mobile, as this situation only applies to the base station transmitter

The above data is now used to calculate the power available to a mobile at the edge of the cell. Using the following default values for downlink 8kbs-1 voice circuits:

Pole capacity 125.7 Cell edge downlink power correction factor 1.65 Activity factor 0.5

(Power available to edge-of-cell mobile) = (Power available to all channels) - 10 * lg { Activity factor * Cell loading * [ (Pole Capacity) / (Power correction factor) ] } (Power available to edge-of-cell mobile) = (+40.491) - 10 * lg { 0.5 * 0.5 * [ (125.7) / (1.65) ] } = +27.69301dBm




40.4910 21

Pole capacity for service 125.7 0

Proportion of pole capacity loading

0.5 0.5

Channel activity proportion 0.5 0.5 (not used)

Power correction factor for edge of cell

1.65 0

Resulting power available for mobile at edge of the cell /dBm

27.6931 21

Table 20 Cell loading




Antenna systems

The resulting power available for a mobile at the edge of the cell calculated above, is then affected by both transmitter and receiver antenna systems, in the same manner as for any other radio transmission.

The main components can be grouped into:

• Transmit antenna feeder loss

At a base station with a 30m mast, using 7/8” diameter foam dielectric cable, a loss of 2dB would be expected, to which a value of 1dB should be added for additional losses in the tail and connectors.

Giving a total typical base station antenna feeder loss value of ~3dB.

At the mobile, the loss is negligible in the case of an integral antenna; although a figure of 1-2dB is typical in the case of vehicle mounted antennas.

• Transmit antenna gain

Antennas operating at 2GHz can be physically smaller, for the same electrical performance, than the 900MHz and 1.8GHz equivalents. It is therefore possible to use antennas of at least the same gain as those used on the lower bands (although their electrical characteristics will vary).

In the case of base stations, typical values are 8-11dBi for omni-directional antennas, and 12-20dBi for directional antennas. In the case of mobiles, typical values are ~0dBi for integral antennas, and 2-3dBi for small vehicle antennas.

• Receive antenna gain

Whether separate antennas are used for transmission and reception or a duplex antenna is used, receive antenna gain is similar to that of the transmit antenna.

• Receive antenna feeder loss

Whether separate feeders are used for transmission and reception, or a duplex feeder is used, receive antenna feeder loss is similar to that of the transmit feeder.

By using the typical antenna system attenuation values quoted above, the Effective Isotropic Radiated Power (EIRP) is calculated in exactly the same manner as for a FDMA TDMA system, together with the receiver antenna system gain.



139



Resulting power available for a mobile at the edge of the cell /dBm

27.6931 21

Transmit antenna feeder loss /dB

2 0

Transmit antenna gain /dBi 17 0

Transmit EIRP per traffic channel at cell edge /dBm

42.6931 21.0000

Receive antenna gain /dBi 0 17

Receive antenna feeder loss /dB

0 2

Receive antenna system gain /dB

0 15

Table 21 Antenna system gain

Noise and interference

Unlike FDMA TDMA systems, CDMA systems suffer from an effect whereby the effective cell coverage area is reduced as traffic load (number of users and data rate) served by the cell is increased.

As the number of users within a CDMA cell increases so does the energy radiated (composite power) on the common RF carrier. If all the users had completely orthogonal channel access codes, this would be indistinguishable from a rise in the thermal noise floor. However, in practice the interference effect of the radiated energy is intensified, owing to non-perfect orthogonality between the codes used by the different users. Thus inter-user (or mutual) interference results. Clearly the level of mutual interference will rise as more users access the common RF channel, and as they make more intensive use of the channel (for example, sending more data).

The mutual interference effect can is calculated by use of the following parameters:

• Orthogonality factor

Default downlink value 0.4, specified by ETSI for the Vehicle-A channel, this channel is based on that expected from a macrocell, distant from the mobile, with little line-of-sight path.




The ETSI Pedestrian channel has a value of 0.16, representing a microcell channel, where the mobile is much closer to the BTS, and more line-of-sight paths exist.

Default uplink value 1.

• Intra-cell interference factor (dB)

This is the level of co-channel interference caused by mobiles operating in the same cell, using imperfect orthogonal channel access codes. It is a function of the cell loading, and the orthogonality factor, and is calculated from the expression:

Intra-cell interference (dB) = 10*lg {1 / [1 – (orthogonality factor * cell loading fraction) ] }

For the downlink, intra-cell interference (dB) = 10*lg {1 / [1 – (0.4 * 0.5) ] } = 0.9691dB

For the uplink, intra-cell interference (dB) = 10*lg {1 / [1 – (1 * 0.5) ] } = 3.0103dB

• Inter-cell interference factor

This is a measure of the level of co-channel interference, caused by mobiles operating in neighbouring cells, using imperfect orthogonal channel access codes.

A default value of 0.85 is used, which is based on that used by American CDMA systems at 600 tri-sector sites.

This results in a corresponding increase in interference by a factor of 1.85, or 2.6717dB.

Natural noise is handled in the link power budget calculation in the same manner as for FDMA TDMA systems:

• Thermal noise floor

Standard default value –174dBmHz-1 .

• Effective receiver noise bandwidth

This is approximated to the information data rate bandwidth. In the case of 8kbs-1 voice circuits, this is 8kHz, or 39.0309dBHz.

• Receiver noise figure

For mobiles, the ETSI link budget normally uses a value of 8dB and receivers used in American CDMA mobile often have a noise figure of 10dB. The typical default value 9dB.

For base stations, the Lucent BTS design specification is for a noise figure of 2dB, which compares well with the existing American CDMA base station receivers which have a noise figure ~5dB.

This data is used to calculate the cumulative effect of the system noise and interference.



141



Orthogonality factor 0.4 1

Intra-cell interference factor /dB 0.9691 3.0103

Inter-cell interference factor /dB 2.6717 2.6717

Total interference margin /dB 3.6408 5.6820

Thermal noise floor /dBHz-1 -174 -174

Effective receiver noise bandwidth /dBHz

39.0309 39.0309

Receiver noise figure /dB 9 2

Effective receive noise floor /dBm

-125.9691 -132.9691

Table 22 Effective receiver noise floor

Receiver sensitivity

For satisfactory operation, the receiver must be provided with a power level that exceeds its effective noise floor by the minimum bit energy-to-noise ratio [Eb/(N0+I0)] required for satisfactory demodulation and decoding. This is the minimum discernible signal level for the receiver.

• Minimum [Eb/(N0+I0)] value

This is the threshold above the receiver noise floor defining the minimum discernible signal for the receiver. Owing to the use of different coding schemes and user data rates the value varies according to the type of service and operating environment.

A downlink value of 7.9dB and an uplink value of 5.3 is used as the default for an 8kbs-1 voice channel operating in a Vehicular-A 120kmh-1 Propagation Channel

• Power control error standard deviation

To ensure satisfactory demodulation and decoding, a signal margin is added to allow for imperfect power control.

A default value of 0.2590dB has been adopted based on the experience of American CDMA networks, where a power control error with standard deviation 1.5dB has been observed.

• Total interference margin




To ensure satisfactory demodulation and decoding, a signal interference margin is added to the minimum [Eb/(N0+I0)] value.

Default values of 3.6dB for the downlink and 5.682dB for the uplink were calculated in the previous Noise and interference section.


Mobile Transmitter uplink

Minimum (Eb/N0) /dB 7.9 5.3

Marin for power control error /dB

0.2590 0.2590

Margin for interference /dB 3.6 5.7

Effective minimum (Eb/(N0+I0)) /dB

11.8 11.2

Table 23 Receiver minimum [Eb/(N0+I0)] threshold

Receiver minimum discernible signal

At this point the minimum discernible signal for the receiver system can be calculated from the following components.

• Effective receiver noise floor (as evaluated above)

• Effective receiver minimum [Eb/(N0+I0)] demodulation and decoding threshold (as evaluated above)



Effective receiver noise floor /dBm

-125.9691 -132.9691

Effective minimum (Eb/(N0+I0)) /dB

11.8 11.2

Receiver minimum discernible signal /dBm

-114.1693 -121.7281

Table 24 Receiver minimum discernible signal



143

Link power available for path loss

• Receiver antenna system gain (as evaluated above)

• Resulting power available for a mobile at the edge of the cell (as evaluated above)

• Receiver effective minimum discernible signal (as evaluated above)

• Gain derived from soft handover:

- When a mobile is operating in connected mode for circuit switched services it may have radio links with more than one base station. This provides an improvement in performance of the resulting composite link, which can be translated into a soft handover gain that can be incorporated in the link power budget

- However, when the mobile is operating in packet data mode rather than circuit switched mode, depending on the system configuration either a dedicated or a common packet data channel may be used. The same type of packet data channel does not have to be used for the downlink and the uplink

- If a common packet data channel is used, the mobile communicates with only one base station, and there is no soft handover during the transmission of a data packet. This mode of operation is analogous to that of a GPRS mobile. Under these conditions only a single link to a base station exists, and consequently there is no soft handover gain applicable to the link power budget

- When applicable, the value of the soft handover gain for circuit switched services is a probability function of the ‘overlapping’ cell area at the edge of the cell and the standard deviation, as shown in the following table:

Cell area overlap probability /%

Standard deviation 8dB



85 - 4.4dB 5.3dB

90 3.6dB 4.6dB 5.6dB

95 3.9dB 4.9dB 6.0dB

97 4.1dB 5.2dB 6.3dB

Soft handover gain

99 4.5dB - 6.8dB

Table 25 Soft handover gain

- A default value of 5dB is used for 87.6% coverage at the edge of the cell




In this example, the resulting power budget available for setting against the path loss is as calculated below:



Power available to a mobile at the edge of the cell /dBm

42.6931 21.0000

Soft handover gain /dB 5 5

Receiver antenna system gain /dB

0.0 15.0

Receiver effective minimum discernible signal /dBm

-114.2 -121.7

Maximum acceptable path loss /dB

161.8624 162.7281

Table 26 Maximum acceptable path loss



145

Example link power budgets

This section summarises the typical link power budget described in the previous section.

Base transmitter downlink Mobile transmitter uplink

Information data rate /kbs-1 8 8

Transmitter output power /dBm 43 21

Proportion of power used by common control channels /dB

-0.3342 0

Proportion of power used by soft handover /dB

-2.1748 0


40.4910 21

Pole capacity for service 125.7 125.7

Proportion of pole capacity loading 0.5 0.5

Channel activity proportion 0.5 0.5 (not used)

Power correction factor for edge of cell

1.65 0

Resulting power available for a mobile at the edge of the cell /dBm

27.6931 21

Transmit antenna feeder loss /dB 2 0

Transmit antenna gain /dBi 17 0

Transmit EIRP per traffic channel at cell edge /dBm

42.6931 21.0000

Receive antenna gain /dBi 0 17

Receive antenna feeder loss /dB 0 2

Receive antenna system gain /dB 0 15

Orthogonality factor 0.4 1

Intra-cell interference factor /dB 0.9691 3.0103

Inter-cell interference factor /dB 2.6717 2.6717

Total interference margin /dB 3.6408 5.6820




Base transmitter downlink Mobile transmitter uplink

Thermal noise floor /dBHz-1 -174 -174

Effective receiver noise bandwidth /dBHz

39.0309 39.0309

Receiver noise figure /dB 9 2

Effective receive noise floor /dBm -125.9691 -132.9691

Minimum (Eb/N0) /dB 7.9 5.3

Margin for power control error /dB 0.2590 0.2590

Margin for interference /dB 3.6 5.7

Effective minimum (Eb/(N0+I0)) /dB 11.8 11.2

Receiver minimum discernible signal /dBm

-114.1693 -121.7281

Soft handover gain /dB 5 5

Receiver antenna system gain /dB 0.0 15.0

Receiver effective minimum discernible signal /dBm

-114.2 -121.7

Maximum acceptable path loss /dB 161.8624 162.7281

Table 27 Example 8kbs-1 voice (Vehicular-A) UTRAN link power budget

This example shows that for 8kbs-1 voice circuits, operating in a vehicular-A environment, the uplink and downlink are broadly balanced, with a slight benefit of ~0.9dB in the case of the downlink.



147

For comparison purposes, the table below reproduces the maximum acceptable path loss that results from the GSM link power budgets, as detailed in Section 6 of the GSM Dual Band Operation Engineering Guideline (order number 401-380-366):

Operating band Base transmit – downlink maximum acceptable path loss

Mobile transmit – uplink maximum acceptable path loss

900MHz 155.2 to 158.5dB 155.1 to 161.2 dB

1800MHz 151.9 to 156.2 dB 148.0 to 159.0 dB

Table 28 Example GSM maximum acceptable path loss

The maximum acceptable path loss for the UTRAN compares favourably with the range of values given above for GSM networks, and in the case of 8kbs-1 voice circuits operating in a vehicular-A environment, the UTRAN value is similar to that for the 900MHz GSM band.

To give an indication of the extent to which the maximum acceptable path loss is reduced as the service data rate is increased, the following table summarises the maximum acceptable path loss for various services. The same general assumptions are used regarding the link power budget as for the previous example of an 8kbs-1 voice circuit operating in a vehicular-A environment.

The itemised link power budget for each of the services is given in Section 7 of the UMTS Radio Network Coverage and Capacity Model description document.

Subscriber Service Base Transmit – Downlink Maximum Acceptable Path Loss /dB

Mobile Transmit – Uplink Maximum Acceptable Path Loss

/dB

Voice circuit @ 13kbs-1 166.8 164.5

Data circuit @ 14kbs-1 169.5 166.1



Data packet @ 40kbs-1 [1] 164.1 157.8

Data packet @ 128kbs-1 [1] 165.1 152.7

Table 29 Typical UTRAN maximum acceptable path loss for other subscriber services

[1] Link power budget for packet data services does not include any soft handover gain.

This table illustrates that the maximum acceptable path loss falls with increasing subscriber data rates, falling to ~10dB less than that available for 8kbs-1 voice circuits, when 128kbs-

1packet data is used.




At data rates below 128kbs-1, the maximum acceptable path loss for a UTRAN generally exceeds that for the GSM cases discussed earlier.

Margins for fading and building attenuation

The previous section calculated the maximum path loss that could be tolerated while maintaining a satisfactory link.

However, it should be noted that from this total path loss budget, elements to cover the shadow fading (slow, or log-normal fading) and for indoor coverage, building attenuation must be apportioned.

Shadow fading

A shadow fading margin is used to account for the natural variation in signal level in different areas of the cell, according to shadowing produced by terrain and clutter features (not as a result of distance from the base station). This is a probability function, based on the chance of receiving a signal that exceeds the mean required service threshold at any point in the cell, and the standard deviation of signal within the cell. The default values are shown in the following table:

Probability of signal exceeding mean required

level /%




85 3.2dB 4.9dB 6.7dB

90 5.4dB 7.6dB 9.9dB

95 8.6dB 11.6dB 14.6dB

97 10.7dB 14.1dB 17.6dB

99 14.6dB 18.9dB 23.3dB

Table 30 Shadow fade margin

Indoor coverage building attenuation

Where indoor coverage is provided, an additional margin must be applied to the required service signal level to allow for attenuation of the signal when passing into buildings. As the attenuation of radio signals caused by buildings is a function of the magnitude of the radio frequency, the attenuation for UTRANs operating in the 2GHz band is similar to that experienced by GSM networks operating in the 1.8 and 1.9GHz bands.

The value of this attenuation is highly variable, according to the type of building and location in the building. For further information regarding indoor coverage, please refer to the Indoor Coverage Systems for GSM Networks Engineering Guideline (order number 401-380-363).



149

Typical default values based on four building types, are given in the following table:

Building area type Indoor attenuation /dB

Dense urban Most buildings >4 storeys, some

much higher

20

Urban Buildings up to 4 storeys, but with

some open space in-between

18

Suburban Modern industrial estates, high &

low density dwellings

12

Rural Water features, fields & heath-

lands, agricultural land, woodland, and hilly & mountainous areas

8

Table 31 Default building attenuation margin

6.4. Estimating coverage and traffic capacity

In a GSM system the area that can be covered by a cell is independent of the traffic load borne, as it is determined by the transmitter power plus antenna gain (EIRP), receiver minimum discernible signal, and the path loss. None of these depend on the traffic load. When planning a new network, initially there will be little traffic and for economic reasons a small number of large cells are normally used. The maximum cell range can be determined from the maximum available path loss power budget, and as traffic load increases this will have no effect on the cell coverage area.

However, when considering a UTRAN that uses a CDMA air interface, the maximum area that can be covered by a cell is not independent of the traffic load. As described in the previous section examining the path loss power budget, CDMA systems suffer from mutual interference caused by:

• Intra cell interference

- Each mobile served within the cell transmits energy on the common RF channel. This appears as uncorrelated noise in de-spread channels used by other mobiles (a rise in noise floor)

- As more energy is transmitted (for example, more mobiles served, higher power transmissions used), the apparent noise floor rises

- Typically, a cell operating at 50% of its maximum traffic load (50% pole capacity), experiences a 3dB rise in its effective noise floor

• Inter cell interference




- The cause and effect of this is similar to that of intra-cell interference, as it represents the co-channel energy transmitted in neighbouring cells and received as uncorrelated noise

- A typical maximum value of 2.67dB is used based on American CDMA networks using a tri-sector site with 600 beam-width antennas

The result of these factors is a dependency between traffic load, both in the serving cell and neighbouring cells, and the maximum available path loss power budget, and hence the maximum cell operating range.

The effect is illustrated in the following graph, where the maximum available path loss power budget is calculated for different values of traffic load (mutual interference), and the resulting path loss translated into a maximum cell range.

Figure 24 Maximum cell radius against traffic load based on path loss power budget

Note: The calculations in Figure 24 are illustrative only. They represent a specific case and should not be applied generally. The values shown on the axes can be disregarded.

Fixed cell loading

The maximum recommended cell load proportion is 50%. This point is shown in Figure 24. The point at which this cuts the uplink and downlink curves of range against load determines the

0 50 100 150 200 250 300 350 4000

0.5

1

1.5

2

2.5

3

3.5

←Downlink

Uplink→

Erlangs (2% GOS)

Max

imum

cel

l rad

ius

(km

)

8kbps circuit switched service capacity versus maximum cell radius

50% of pole capacity

Cell radius used fornetwork dimensioning



151

maximum cell operating range. In the previous case, which considers an 8kbs-1 voice circuit, the downlink is (marginally) the limiting case.

This cell radius is then used to estimate the number of cells required to provide coverage over the desired service area.

In areas where the offered traffic in the service area is greater, either an additional RF channel (and associated TRX) is used at each cell, or, if this is not possible, the network becomes capacity limited.




Adaptive cell loading

When the subscriber density and associated traffic density is known, the ideal cell loading point can be determined by plotting a graph of cell radius against offered traffic on the same axis as the graph of cell radius against cell traffic load capacity.

An example is shown in the following graph:

Figure 25 Optimum cell loading point for given offered traffic density

Note: The calculations in Figure 25 are illustrative only. They represent a specific case and should not be applied generally. The values shown on the axes can be disregarded.

As mentioned previously, as cell range increases, traffic capacity decreases. So with a given subscriber density, the offered traffic will increase with cell radius. The optimum loading point is found where the two factors balance. It is located at the point where the cell radius contains the amount of offered traffic equal to the maximum cell capacity at that range.

0 50 100 150 200 250 300 350 4000

0.5

1

1.5

2

2.5

3

3.5

Erlangs

max

imum

cel

l rad

ius

(km

)

cell radius versus voice user capacity for 3 sector urban macrocell

Maximum allowable loadingfor uplink and downlink

2000 subscribers per km2

(50 erlangs per km2)





25mE persubscriber

Optimumloading points



153

Estimating base station numbers

Lucent provides a UMTS Dimensioning tool that automates the process of calculating maximum cell radius for a given volume of offered traffic, and thereby determining the minimum number of cells and base station sites needed to cover a given area.

The tool is written in the Mathworks ‘Matlab’ programming language, and runs under Matlab version 5.3 (and only version 5.3) which can be used on an Intel x86 compatible PC using Microsoft Windows 95 or NT4.

The UMTS Dimensioning tool requires the following user data:

• System parameters

- Blocking rate, default value 0.02 (2%)

- Chip rate, default value 3.84M chips s-1

- Coverage area, no default value, value entered in square kilometres (1 sq. km = 100ha)

- Operating frequency, default value 2000MHz

- Maximum cell loading proportion, default value 50%

- Dimension based on fixed or adaptive loading, default choice is adaptive loading

• Link budget and BTS

Each of 4 types of land use (Dense Urban, Urban, Suburban, and Rural) must be selected and the following values entered for each type:

- Base station antenna height, default values 30m in all cases except for Rural, for which 40m is used

- Mobile station antenna height, default value 1.5m

- Base station power output (does not include antenna system gain), default value +43dBm

- Base station feeder loss, default value 2dB

- Mobile station antenna gain, default value 0dBi

- Base station noise figure, default value 2dB

- Mobile station noise figure, default value 9dB




- Number of carriers. The number of RF channels or transceivers available at each cell, default value 1. It is assumed that the traffic load is distributed uniformly across the channels

- Number of sectors. The number of antenna sectors, each producing one cell, available at each BTS site

- Normalised sector overlap. The angle over which the practical sector antenna provides coverage, which overlaps that provided by an ideal sector antenna. Together with the number of sectors used, this determines the inter-sector interference

Default values: Omni-directional antennas 10 Bi-sector 10 Tri-sector 120

- Shadow fade margin (dB). Default values as described previously

- Soft hand-off gain (dB). Default values as described previously

- Indoor penetration loss (dB). Default values as described previously

• Path loss model

- Choice of COST 231, Hata, or user defined model. Default is the COST 231 model

- Correction factors, choice of Hata or user defined correction factors. Default is the user defined correction factors, which are entered as follows:

Dense urban 3dB Urban 0dB Suburban -6dB Rural -18dB

- Downlink orthogonality factor. This value indicates the extent to which the downlink channel codes are truly orthogonal, with a value of 1 indicating the ideal case yielding no intra-cell interference, and a default value of 0.4

• Traffic and services

- Service bit rate (bs-1). The information data rate as presented to the subscriber, for the range of subscriber services that are available

- Eb/I0 [uplink] (dB). The minimum value of the ratio of the bit energy to interference (plus noise) required for satisfactory demodulation, decoding and user data recovery, such that the subscriber is presented with the data at the BER specified for the service

- Eb/I0 [downlink] (dB), as above, but for the downlink



155

- Activity factor. The duty cycle of the subscriber data. For example, a default value of 0.5 is used for voice circuits as people do not normally talk and listen simultaneously in phone conversations. For packet services, although the duty cycle is small, since the value is only used to evaluate the aggregate data throughput, it is recommended that a default value of 1 is used

- Service type. Defines the service type as either circuit or packet switched. However, when packet switched data is used, there is currently no means to specify the maximum packet delay that is acceptable. Packet data is analysed by finding the number of trunks that can be supported; throughput is calculated by multiplying this value by the service bit rate

- Common packet channel (uplink). Identifies whether a common or dedicated channel is used for uplink packet data. Common channel operation is more efficient. However, a link is only maintained to a single base station so soft handover gain is not realised. Dedicated channel operation is less efficient, but the link can benefit from soft handover gain

Default value is to use a Dedicated Channel (that is, not use a common channel) for uplink packet data

- Common packet channel (downlink). Identifies whether a common or dedicated channel is used for downlink packet data

Default value is to use a common channel for downlink packet data

- Channel elements per Baseband Unit (BBU). Specifies the number of subscriber data circuits for each user data rate that can be provided by one BBU. Default values are:

Service data rate /kbs-1 8 13 30.4 60.8 243.2 16 32 64 144 384

No. circuit elements per BBU 32 32 16 8 2 32 16 16 5 2




Using the UMTS Dimensioning tool

The UMTS Dimensioning tool can be operated in two modes. Batch operation on a pre-defined data file containing values for a number of traffic load cases, or on a case-by-case basis using data entered in the traffic load window.

If batch analysis is used, the user enters the paths and names for the input traffic data file, and the output (number of base stations result) file.

Input file

The input traffic load data file must be in MS Excel ‘.csv’ format. An example is shown below:

label,morph type,area,8k voice erlang density,13k voice erlang density,U/L 64k packet uplink,D/L 64k packet uplink,U/L 144k packet uplink,D/L 144k packet uplink,U/L 384k packet uplink,D/L 384k packet uplink,U/L 16k circuit data,D/L 16k circuit data,U/L 32k circuit data,D/L 32k circuit data,U/L 64k circuit data,D/L 64k circuit data,U/L 144k circuit data,D/L 144k circuit data,U/L 384k circuit data,U/L 384k circuit data<CR><LF> central area,1,20,200,0,0,0,0,0,0,0,0,0,10,10,0,0,0,0,0,0<CR><LF>

suburbs,3,200,10,0,0,0,0,0,0,0,0,0,0.5,0.5,0,0,0,0,0,0<CR><LF>

rural,4,10000,0.1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0<CR><LF> 2015_16_case1_du_g,1,275,55.34,0,0.0952,0.2825,0,0.3943,0,0.204,0,0,0,0,0,0,1.7457,1.745,0,0<CR><LF>

2015_16_case1_u_g,2,1782,29.4187,0,0.050,0.1502,0,0.2096,0,0.1084,0,0,0,0,0,0,0.92,0.928,0,0<CR><LF>

etc.

The file is in ASCII format. Values are comma separated. Carriage return and line feed characters are indicated by <CR><LF>.

Each input data field is described below:

1. Label – ASCII text name that identifies the particular traffic load case.

2. Morph type – a number in the range 1 to 4, indicating the type of land use in the area for which the traffic load relates (1 = Dense Urban, 2 = Urban, 3 = Suburban, 4 = Rural).

3. Area – coverage area for which the traffic load relates, in square kilometres (1 square kilometre = 100 hectares).

4. 8k voice erlang density – density of 8kbs-1 voice traffic in erlangs per square kilometre.

5. 13k voice erlang density – density of 13kbs-1 voice traffic in erlangs per square kilometre.

6. U/L 64k packet – density of 64kbs-1 uplink packet traffic in Mbs-1 per square kilometre.

7. D/L 64k packet – density of 64kbs-1 downlink packet traffic in Mbs-1 per square kilometre.



157





12. U/L 16k circuit data – density of 16kbs-1 uplink circuit switched traffic in erlangs per square kilometre.

13. D/L 16k circuit data – density of 16kbs-1 downlink circuit switched traffic in erlangs per square kilometre.








21. U/L 384k circuit data – density of 384kbs-1 downlink circuit switched traffic in erlangs per square kilometre.




Output file

The output file is in ASCII format and values are comma separated. An example is shown below:

Area identifier ,Number of BTS required ,Number of BBU cards per BTS,Cell radius (km) ,Cell area (km2) ,Loading (U/L) ,Loading (D/L) ,Number of sectors at BTS ,Number of carriers at BTS ,<CR><LF>

central area , 44.000, 6.000, 0.484, 0.457, 33.485, 41.650, 3.000, 1.000<CR><LF>

suburbs , 53.000, 3.000, 1.403, 3.835, 14.051, 17.472, 3.000, 1.000<CR><LF> rural , 39.000, 2.000, 11.586, 261.562, 6.974, 8.353, 3.000, 1.000<CR><LF>

2015_16_case1_du_g,1157.000, 3.000, 0.349, 0.238, 10.906, 18.946, 3.000, 1.000<CR><LF>

2015_16_case1_u_g ,4399.000, 3.000, 0.456, 0.405, 9.895, 17.196, 3.000, 1.000<CR><LF>

Each output field is described below:

1. Area identifier – this is the label used to identify the traffic load case in the input data file.

2. Number of BTS required – the number of base stations, each radiating one or more cells (sectors) that are required to cover the area specified in the traffic load input file.

3. Number of BBU cards per BTS – the number of Baseband Unit cards that are required at each base station site.

4. Cell radius (km) – the radius in kilometres of each cell radiated by the base station

5. Cell area (km2) – the area in square kilometres of each cell radiated by the base station.

6. Loading (U/L) – the proportion of the maximum (pole) capacity of the cells (that is, the capacity of the RF channel(s)) that is used for the uplink traffic.

7. Loading (D/L) – the proportion of the maximum (pole) capacity of the cells (that is, the capacity of the RF channel(s)) that is used for the downlink traffic.

8. Number of sectors at BTS – the number of cells (sectors) that are radiated from each base station site.

9. Number of carriers at BTS – the number of RF channels (frequencies) that are radiated from each cell (sector) at the base station.



159

Land-use classification

The results produced by the UMTS Dimensioning tool are extremely sensitive to the land-use (morphology) classification used. Classification of land-use can vary from country to country. What may be considered a Dense Urban area in one country may not be considered so in another country.

Examples of land-use categories used by some public network operators together with Lucent’s approximation of associated population densities are shown below:

Type Approx. population density /people per sq. km

Description

Dense Urban > 2,0001 Area in which most of the buildings have more than 4 storeys and some can be classed ‘skyscrapers’. This classification is restricted to the centre of a few large cities.

Urban 1,000 to 2,000 Area in which buildings have up to 4 storeys, some buildings have more than 4 storeys, but there may be some open space in-between.

Suburban 200 to 1,000 Area with either high-density dwellings or low-density dwellings and modern industrial estates.

[1] A high population density is not always associated with a Dense Urban area, as zones with a large proportion of high office blocks, often do not have a large residential population.

Table 32 Land-use classifications

UMTS Dimensioning tool – further reference

User documentation is provided in the following documents:

• UMTS Network Dimensioning Software – User Guide

• UMTS RF Dimensioning Tool – Input Parameter Guidelines

The theory behind the software is described in detail in:

• UMTS Radio Network Coverage and Capacity Model description

• UMTS RF Network Dimensioning Tool Theory description





6.5. Airpro coverage and traffic distribution prediction software

The previous section introduced the concept of estimating the coverage range of cells based on the path loss power budget developed according to the type and volume of traffic that is offered. The UMTS Dimensioning tool automates this process of estimating the number of base stations needed to cover a required service area.

However, this technique only estimates the number of sites required. It does not take account of:

• Terrain features (for example, hills blocking the signal) that vary over the coverage area

• Ground clutter (for example, building height and type, vegetation, land use) that varies within the coverage area

• Antenna radiation patterns

• Site specific antenna height (the estimation software uses a single antenna height for all base station sites)

• Offered traffic load and type non-uniformly distributed over the coverage area

Also, the UMTS Dimensioning tool cannot produce maps of predicted received signal and interference power.

Lucent has therefore developed software called ‘Airpro’ to perform the coverage prediction and interference analysis. This is used for the next stage of RF coverage planning.

The Airpro and UMTS dimensioning tool are independent programs and can be used in conjunction with each other or separately.

Introduction

Airpro is the latest version of the RF planning software known as CE4, developed by Lucent over the last 15 years and currently used by over 400 customers. It has a user-friendly graphical interface and runs on an Intel x86 compatible PC under the Microsoft NT4 operating system.

Principal input data categories include:

• Terrain map. A matrix of spot land heights against location, in either rectilinear grid format (for example, OSGB Grid), or geographic co-ordinates (for example, latitude and longitude)

• Ground clutter map. A matrix of spot clutter (such as buildings and vegetation) category (such as type/height) against location, in either rectilinear grid format or geographic co-ordinates



161

• Offered traffic density maps for each type of service, depicting polygons of differing traffic volume (such as Erlangs per square kilometre) against location, in either rectilinear grid format or geographic co-ordinates

• Antenna radiation patterns such as graphs of antenna gain against azimuth and elevation angle

• Road maps and special features, usually supplied in vector format such as Drawing Exchange Format

The major outputs take the form of:

• Screen and plotter output of system configurations and analysis results

• Standard and customised statistical reports

• Antenna by antenna data analysis

A range of applications is currently supported in the CE4 RF planning tool and these will be carried forward into the Airpro release. These include RF propagation analysis, interference prediction, and utilities to help set up and manage system designs for diverse operating environments.

Principal features

This section outlines the main features and facilities supported by the Airpro software.

Propagation models

The following models have been implemented:

• Lucent model. Based on the generalised slope and intercept model

• Cost 231 Hata model. A variation of the Hata model used in the 500 to 2000 MHz frequency range

• ITU-R M.1225 general model. For wide band CDMA, mandated by the international organisations that oversee the wide band CDMA standards

Power allocation

Airpro supports the following power allocation:

• UMTS. Beacon channel / total overhead channel / traffic channel / reserved power

• CDMA 2000. Pilot channel / total overhead channel / traffic channel / reserved power




Phase 1 Release - January 2000

The key features included in Airpro Version 1.0 are:

• Wide band CDMA analyses:

- Pilot (beacon) signal strength analysis

- Forward strongest Ec/I0 analysis (with and without traffic data)

- Forward soft handoff analysis (without traffic data)

- Reverse strongest traffic Eb/N0 analysis (without traffic data)

- Forward Pilot (Beacon) overlap analysis

- Forward nth Pilot (Beacon) analysis

- Forward link coverage analysis (Pilot)

• Automatic PN code assignment

• Wide band CDMA related graphic user interface screens and data base development

• User defined traffic region (first part)

• Airpro will continue to support applicable features contained in its previous releases of CE4

Phase 2 Release - June 2000

Airpro Version 2.0 planned for release in June 2000 will include:

• Analyses for wide band CDMA:

- Forward strongest traffic signal strength analysis

- Balanced Ec/I0 coverage analysis

- Balanced Eb/N0 coverage analysis

- Forward link coverage analysis (traffic)

- Forward Neighbour List

- Forward Combined Pilot (beacon) Ec/I0

- PN overlap

- PN delay



163

- Reverse link coverage

• User defined traffic region (second part)

• File import and export

• Traffic analysis

• Measurement data integration

• Bin level details

• Help and set-up

The analysis results can be displayed on screen and sent to a plotter. Major examples include System (Market and Scenario level), Cell Configuration, Technology, Model, Path-loss, Antenna, Terrain, and Clutter Data.

Scenario related data can also be exported in Microsoft Excel file format (for example, MSC, Cell, Antenna, Technology, and Path-loss).

Services and traffic model

Airpro Phase 1.0 supports the definition of services with any bit rate, Eb/I0 requirement, activity factor, operating point (percentage of pole capacity), traffic type (circuit or packet), and blocking rate (circuit services only) for both forward and reverse links, and Ec/I0 for the forward link.

Airpro Phase 2.0 will allow a delay requirement for packet switched data to be defined.

Services can be defined as having asymmetric data rates. The reverse link (traffic Eb/I0 analysis) and forward link (traffic Eb/I0 and/or Ec/I0 analysis) coverage are analysed separately. Capacity analysis is likewise performed separately for the uplink and downlink.

Simulations have been carried out to determine empirical factors for power allocation used in the software. At present these are used to determine the power available to a mobile at the edge of a cell (given uniform distribution of mobiles) in the forward link. Different factors have been determined for each type of service bit rate, and interpolation has been shown to perform well in determining these factors for new services.

When a mix of traffic services is involved, coverage planning is based on the service that must be provided throughout the coverage area that represents the minimum cell radius.

The situation is more complicated for capacity analysis of multiple service types. Lucent has performed both simulation and theoretical analysis to determine the best way to approach this problem. The following graphs show the capacity relationship between voice, 64, 144, and 384kbs-1 circuit switched data. These results were obtained by 2 dimensional Markov chain modelling and have been confirmed by simulation.




0 20 40 60 80 100 120 1400

2

4

6

8

10

12Circuit switched 64kbps versus voice, 2% blocking

Dat

a E

rlang

s

Voice Erlangs

Figure 26 Graph of 64kbs-1 circuit switched data and voice capacity (fixed 2% blocking)



165

0 20 40 60 80 100 1200

0.5

1

1.5

2

2.5

3

3.5Circuit switched 144kbps versus voice, 2% blocking

Dat

a E

rlang

s

Voice Erlangs





0 10 20 30 40 50 60 70 80 90 1000

0.1

0.2

0.3

0.4

0.5

0.6

0.7Circuit switched 384kbps versus voice, 2% blocking

Dat

a E

rlang

s

Voice Erlangs


It is clear from these graphs that a linear relationship exists between voice service and data services up to and including 144kbps. The 384kbps circuit switched data service shows a significant deviation from linear.

The next three graphs show the relationship between circuit switched voice and packet data. Here it should be stressed that the packet data throughput is an aggregated rate (the UTRAN is always loaded to the maximum possible data throughput). The circuit switched voice always takes priority over the packet switched data. The packet size is 1024 bytes and queue length set to 100, 10 and 1 packets.



167

0 20 40 60 80 100 120 1400

2

4

6

8

10

12x 10

5

Voice erlangs

Thro

ughp

ut

Packet switched 384kbps throughput versus voiceerlangs

Figure 29 Graph of 384kbs-1 packet switched data and voice capacity (fixed 2% blocking)

Queue Length = 10 packets

Queue Length = 1 packet





0 20 40 60 80 100 120 1400

1

2

3

4

5

6

7

8

9

10x 105

Thro

ughp

ut (b

ps)

Voice erlangs

Packet switched 144kbps throughput versus voice erlangs







169

0 20 40 60 80 100 120 1400

1

2

3

4

5

6

7

8

9

10x 10

5

Voice erlangs

Thro

ughp

ut

Packet switched 64kbps throughput versus voice erlangs


These graphs indicate that a linear relationship exists between all packet data services and voice.

Airpro uses this information to determine the overall capacity within a cell that contains multiple service types by exploiting the linear relationship that exists between services.

Only the linear relationships are modelled in Phase 1 of the Airpro software. Phase 2 will deal with the exceptional cases (high bit rate circuit switched services) by means of empirical functions.

Phase 2 will also allow a maximum delay to be specified as a means of quality of service (comparable to the blocking rate in circuit switched services). The following figures show the relationship of delay versus voice traffic (Erlangs) for 384kbps and 64kbps packet data.







0 20 40 60 80 100 120 1400

2

4

6

8

10

12

14

Voice erlangs

Del

ay (s

) Queue length = 100

Queue length = 10

Queue length = 1

Figure 32 Graph of 384kbs-1 packet data delay against voice traffic load (fixed 2% blocking)



171

0 20 40 60 80 100 120 1400

1

2

3

4

5

6

7

Voice erlangs

Del

ay (s

)

Queue length = 100

Queue length = 10

Queue length = 1

Figure 33 Graph of 64kbs-1 packet data delay against voice traffic load (fixed 2% blocking)

At present handover is modelled as a fixed gain for all services and a power budget overhead for forward link analysis. Simulations and analysis will be conducted to determine whether these assumptions need to be modified for packet services.

Similarly, power control is modelled as a fixed standard deviation for all services. It is possible that packet services will experience worse power control errors. Again this will be studied and if necessary included in the Phase 2 release.

PC platform requirements

The recommended minimum PC configuration for the Airpro software is:

• Intel Pentium II compatible CPU clock speed 266MHz

• 2GB FDD

• Microsoft Windows NT4 operating system




Source data file formats

Airpo supports the following file formats:

• Geographic data

- Elevation data from DEM/CD-ROM

- Mapinfo (.MIF), (.MPF), and (.SET) files

- Autodesk (.DXF) files

- Lucent (.ELV) elevation files

- Lucent (.ATR) clutter files

• RF Data

This data category includes the antenna height, azimuth, elevation, type, radiation pattern, transmitter output power, frequency, mobile antenna height, propagation model and so on.

The antenna radiation pattern is usually supplied by the manufacturer. Airpro imports this data in the (.PDF) file format.

• Measurement data

This is real signal strength data recorded from the survey measurements and used to calibrate the prediction model according to local environmental conditions.

Airpro imports measurement data in either Lucent (.CSV) or (.MDI) formats.



173

Output data file formats

Airpro Version 1.0 supports Lucent Analysis Result binary output files. Airpro Version 2.0 will support the following output data file formats:

• Lucent Analysis Result binary files

• Exported Airpro scenario and configuration files, in Microsoft Excel format

• ASCII files

In addition, interfaces to MS Access for the Airpro database will be available in Version 2.0.

Airpro default values

Default values for the principal parameters used by Airpro are listed in the UMTS RF Dimensioning Tool – Input Parameter Guidelines document. This is available on the RF Engineering web site:

http://en0033svr06.uk.lucent.com/rfsystems/UMTSdimenTool.htm

This document is updated frequently to reflect any changes to recommended default values.

http://en0033svr06.uk.lucent.com/rfsystems/UMTSdimenTool.htm




6.6. BTS and antenna settings optimisation software

Base station sites are identified, acquired and constructed in locations that match as closely as possible the ‘ideal’ positions provided by the Airpro coverage prediction software. Owing to local technical and commercial constraints, the base stations may not be in the ‘ideal’ locations.

Introduction

Lucent has developed software called ‘Ocelot’ which uses predictive optimisation algorithms to determine the best initial base station parameter settings during the RF design stage, in order to satisfy the required network coverage and capacity. Using Ocelot can reduce the need for post installation readjustments.

Ocelot operates on the following input data:

• Geography of the coverage area such as terrain, clutter, and offered traffic density

• Traffic density in coverage area

• Base station information such as location, number of sectors, target capacity, and maximum transmitter power

• Antenna details such as radiation pattern, height, and bore-sight

The analysis results in optimised values for:

• Transmitter power

• Antenna azimuth and elevation bore-sight

• Neighbour cell list

Ocelot can simultaneously adjust any combination of the above parameters for a single cell, region, or an entire network.



175

Applications

Ocelot can be used during the RF design of a new network or after part of the network has been implemented. It is typically used for:

• Green-field (new base station) designs

• Market upgrades (adding cells or carriers)

• Overlay to existing designs

It can also be used to re-optimise networks that are in service, to provide overload relief due to local traffic hot spots or other problems.

Cellular radio standards

In addition to the UTRAN version, Ocelot also supports versions for GSM and American CDMA One / IS95 / JSTD 008. Future releases will also support CDMA 2000.

Optimising strategy

Ocelot simultaneously optimises network coverage and capacity.

• Network coverage is the fraction of traffic-weighted area where service can be received with a certain quality. This captures the influences of thermal noise as well as interference from other cells and users

When optimising for coverage, the algorithm generally minimises mutual interference.

• Network capacity is the amount of traffic the network can handle at an overall network-blocking rate of 2%

When optimising for capacity, the algorithm generally adjusts the size of the cells so that the spatial distribution of network capacity and traffic match.

• A trade-off between capacity and coverage must be achieved, since optimising for network coverage and network capacity generally gives different results

Ocelot optimises for all trade-offs between network capacity and coverage. The result is displayed in a trade-off plot, which assigns network capacity and network coverage to the two plot axes. Each potential network configuration is represented by one point in the plot. All network configurations that Ocelot finds form a trade-off curve, representing all configurations with overall best performance. An example trade-off plot is shown below:




Trade-off Curve:Optimum Configurations

Initialconfiguration

Trade-off Space

Maximum CoverageConfiguration

Maximum CapacityConfiguration

“Knee” ofTrade-off Curve

0.80 0.85 0.900.7

0.8

0.9

1.0

Optimized Initial

Capacity

Coverage

Figure 34 Proportion of maximum capacity against proportion of maximum coverage

Figure 35 Ocelot data flow diagram

RF model

Ocelot models radio-propagation effects and radio-link performance to estimate the overall network performance. The currently implemented radio-propagation models are compatible



177

with those of most commercially available RF network design software and can include terrain and clutter effects. The radio-link-performance evaluation uses algorithms that capture power control and network-wide mutual-interference effects. These algorithms extend beyond the capabilities of those used by the RF network planning software that is currently available, thereby yielding improved performance predictions.

Ocelot incorporates an efficient meshing algorithm that utilises road-map data and therefore captures the actual traffic distribution in a realistic manner. The software can import network traffic data, from which it can optimise the performance of existing networks for capacity.

Results

Ocelot presents analysis results via the graphical user interface.

The following Ocelot displays illustrate the effect of adjusting the antenna down-tilt on coverage in the central coverage area. Red areas indicate poor coverage owing to interference or low signal level.

Coverage Improvement With Ocelot - Market XInitial Tilt Settings - Design Ocelot Tilt Settings

Uptilt -3.50 8.00 DowntiltUncovered Areas.Covered Areas.

Figure 36 Network coverage before (left) and after (right) Ocelot optimisation

Comparison trials

Ocelot has been tested in several international networks, for both overall optimisation and individual field trials. Ocelot optimisation was compared against costly in-field optimisation strategies. The trial results have shown that configurations proposed by Ocelot can lead to comparable or even better performance than those gained from conventional methods.




Further information regarding the Ocelot software, together with a list of Lucent contacts and plots illustrating the performance improvement obtained during field trials are contained in the Ocelot – Lucent’s Cellular Optimisation Tool document.

6.7. Inter-system boundary

Lucent Network Release 0.1 does not support inter-system handover, which will be implemented in a subsequent release.

6.8. Further reference

• GSM Dual Band Operation – Engineering Guideline (order number 401-380-366)

Please note that the following documents contain commercially sensitive information and are not generally available to our customers. If you need more details in this area, please contact your Lucent representative, who will arrange to provide the required extracts from the documents:

• UMTS Radio Network Coverage and Capacity Model (Description Document) from the UMTS Performance Analysis and Simulation Group

• UMTS Network Dimensioning Software – User Guide

• UMTS RF Dimensioning Tool – Input Parameter Guidelines

• UMTS Radio Network Coverage and Capacity Model description

• UMTS RF Network Dimensioning Tool Theory description

• Ocelot – Lucent’s Cellular Optimisation Tool document


RF Network Practical Implementation 7


179

7. RF Network Practical Implementation

This chapter considers the practical effects of transition from a GSM network to a UTRAN on the principal network elements and implementation techniques commonly found in GSM networks.

7.1. Use of existing sites

For various economic and environmental reasons, it is likely that existing GSM operators will initially want to re-use their existing GSM sites to provide UTRAN service.

This raises some potential problems to be considered by the network designer:

• Electromagnetic Compatibility (EMC) between UTRAN and GSM equipment. No insurmountable problems are anticipated. This issue is discussed later in this chapter

• The extent to which antenna systems can be shared. A wide range of antennas and supporting components will allow the same scope for shared antenna systems as exists for GSM networks. No insurmountable problems are anticipated. This issue is discussed later in this chapter

• Availability of floor space for additional UTRAN equipment racks in existing GSM equipment rooms or for more street cabinets

• Availability of site services for additional UTRAN equipment:

- Increased power load (mains, UPS and long-term alternative supplies where appropriate)

- Increased heat dissipation requiring greater cooling capability




- Increased floor loading

- If shared antenna systems are not used, more space is required in risers/ducts for cables, loading on the antenna support structure is increased as more cables and antennas are added

- Back-haul network capacity (such as Megastreams or microwave radio links) may have to be upgraded to cope with the UTRAN traffic, which may grow rapidly with the capability to support high data rate subscriber services

- Increased RF radiation from the base station sites

• Coverage that can be achieved from the sites. This is governed by the available link power budget, and is influenced by the antenna system, system parameters, desired subscriber services and quality of service, and the propagation conditions

- An example link power budget for 8kbs-1 voice transmission on 900 and 1800MHz GSM networks and a UTRAN was compared in Section 6.3. When operating in a Vehicular-A channel, the UTRAN link budget was similar to that of the GSM 900 network

- When higher subscriber data rates are considered, the UTRAN link power budget falls below that of the GSM networks by approximately 10dB

7.2. EMC at existing GSM sites

Although the upgrade of existing radio stations to provide UTRAN coverage is extremely attractive from a commercial viewpoint, from an engineering perspective a number of EMC problems can arise regarding the air interface.

Note: This document concentrates on EMC with GSM equipment. However, at shared radio stations, compatibility problems with other types of equipment should also be considered.

Transmitter isolation

Most multiple channel BTS transmitters use a single transmitter output stage for each RF channel that is radiated, from which the power is combined in either a narrow band frequency selective, or broad band ferro-magnetic hybrid, power combiner. The transmit path passes through one (or more) circulators which, together with any frequency selective filtering (for example, in the duplexer) provide sufficient isolation to prevent power from out-of-band transmitters causing intermodulation products within the transmitter output stages.

Recently, with the development of multiple carrier power amplifiers (such as Cartesian Loop and pre-distortion feed-forward ‘linearised Class-C’) there may be less protection in the transmitter output stages, and external circulators may be necessary. However, at present this configuration is unlikely except in the very latest GSM transmitters, and more commonly where multi-carrier masthead power amplifiers are used.



181

Receiver performance degradation

Receiver performance can suffer due to four principal mechanisms:

• Desensitisation (blocking)

• Spurious responses

• Interference from intermodulation products

• Increase in noise floor from wideband interference

Desensitisation

Desensitisation (receiver blocking) usually occurs when a signal enters the first RF amplifier stages of the receiver system at such a level that it drives the amplifier into compression, thereby limiting its ability to amplify weaker signals, and hence desensitising the receiver.

The effect can be mitigated by good out-of-band attenuation of the RF band pre-selector filter to prevent any large out-of-band signals entering the receiver, and high linearity of the first RF amplifier stages which allow amplification over a wide dynamic range.

When considering the co-existence of 900 and 1800MHz GSM and UTRAN BTSs, as the individual systems operate in different bands that are separated by guard bands, receiver desensitisation can best be avoided by use of a good RF band pre-selector filter. At sites where duplex antennas are used, the duplexer can partly perform this filtering. The minimum blocking antenna isolation (Lant_iso) required between BTSs is such that the power received from the out-of-band transmitter is less than that producing a 1dB reduction in the sensitivity of the ‘wanted band’ receiver.

1. Ptx_tot – Zdup - Lant_iso <= Prx_desens

where: Ptx_tot = total power transmitted by potential interferer Zdup = attenuation of receiver system duplexer and RF pre-selector filter in

unwanted band Prx_desens = input receiver power in unwanted band causing a 1dB sensitivity drop

2. The total power received from the ‘unwanted band’ BTS should be 5dB below its receiver 1dB compression point (P1dB)

Pblock = Ptx_tot – Zdup – Lant_iso

Spurious responses

The BTS receiver may exhibit responses at frequencies other than in the desired operating band. Such responses may appear at harmonics and sub-harmonics of the operating frequency and/or the Intermediate Frequency (IF).




If any of the spurious responses occur in the ‘other system’ transmit band, or transmit and/or receive Local Oscillator (LO) frequency, additional filtering may be necessary to prevent interference.

Interference from intermodulation products

Intermodulation products arise from the mixing (multiplication) of two (or more) signals in a non-linear device, resulting in a series of sum and difference products. The mixing may take place in active devices (for example, RF pre-amplifier in a receiver system), or passive devices (for example, antenna system or support structure). Depending on the frequencies involved, mixing products from the two systems transmitters may be generated in the receive band of either system.

The largest of these mixing products is the third order product, the level of which can be calculated from:

IMP3 = (3 * Pint) – (2 * TOI)

Where: IMP3 = Power of third order intermodulation product Pint = Power from ‘other band’ interferer at ‘own band’ receiver input

= Ptx_tot - Lant_iso - Zdup TOI = Third Order Intercept point specified for the receiver input

Ptx = Power transmitted from interfering signal

Transmitter spurii raises receiver noise floor

Transmitters can generate both discrete signals (for example, harmonics) and wideband that fall outside the transmit band. Although the radiation of discrete out-of-band signals should be tightly controlled at the transmitter, synthesised transmitters (particularly DDS based) are prone to generation of wideband low level noise.

The received interference power can be expressed as:

Prx_int = Ptx + CIRtx – Ltx_rej - Lant_iso + 10*lg (W interfered / W interfering)

Where: CIRtx = Interfering transmitter carrier-to-interference ratio (dBc) Ltx_rej = Attenuation of ‘other band’ transmitter filter in ‘wanted band’

receive band Lant_iso = Inter-antenna isolation W interfered = Bandwidth of ‘wanted band’ receiver

W interfering = Measurement bandwidth of interfering signal (‘other-band’ transmission)

Prx_int = Power of ‘other band’ interference signal presented at ‘wanted band’ receiver input



183

Inter-system isolation criteria

When assessing the potential sources of receiver performance degradation, the following antenna isolation guidelines may be used. At the ‘wanted band’ receiver:

• The power received from the ‘unwanted band’ BTS should be 10dB below the ‘wanted band’ receiver noise floor

• The power of the third order intermodulation products that are generated should be 10dB below the ‘wanted band’ receiver noise floor

• The total power received from the ‘unwanted band’ BTS should be 5dB below its receiver 1dB compression point (P1dB)

When these guidelines result in different values for the antenna isolation, the largest value should be used.

Estimating inter-system isolation

In order to estimate the isolation that is needed between the UTRAN and GSM systems, the following data is necessary:

• Selectivity of the BTS receiver

• Spectral purity of BTS transmitter

• BTS transmit power

• Linearity of the antenna system and support structure

For illustration purposes, typical Lucent BTS transmitter and receiver performance data of the type needed to calculate required inter-system isolation is provided in the following tables. Because of the number of possible combinations of equipment, we recommend that you determine the exact configuration of the BTS equipment, and obtain the corresponding performance data for it.

If the existing GSM equipment is not of Lucent’s manufacture, corresponding performance data must be obtained from the manufacturer concerned.




BTS receiver selectivity

Receiver Type ReceiverNoise Floor /dBm

Third Order Intercept Point /dBm

1dB Desensit-isation Blocking Input Power /dBm

1dB CompPoint /dBm

RF Pre-selector Filter Rejection at 900 Tx. Band /dB



UTRAN -105 -6 -17 70+40=110 70+40=110

BTS2000/900 -114 +12 -16 60+30=90 60+30=90

BTS2000/1800 -115 +10 -16 60+30=90 60+30=90

BTS2000/2C/900 -117 +11 -18 60+30=90 60+30=90

BTS2000/2C/1800 -117 +11 -18 60+30=90 60+30=90

Table 33 Typical Lucent BTS receiver performance data

BTS transmitter spurii

Transmitter Type Output Power /dBm

Spurii Power in 900 Receive Band



Spurii Power in 900 Transmit Band



UTRAN +43 <-98dBm <-98dBm <-57dBm <-47dBm

BTS2000/900 +45 <-98dBm <-30dBm <-30dBm <-47dBm <-30dBm

BTS2000/1800 +45 <-98dBm <-30dBm <-57dBm <-30dBm

BTS2000/2C/900 +36 <-98dBm <-30dBm <-47dBm <-30dBm

BTS2000/2C/1800 +36 <-98dBm <-30dBm <-57dBm <-30dBm

Table 34 Typical Lucent BTS transmitter performance data

Antenna coupling

The site specific data gathered in the previous section (supplemented by data from other manufacturers as necessary) is used with the methods described in the Receiver performance degradation section to provide the inter-system isolation required to satisfy the isolation criteria specified in the earlier Inter-system isolation criteria section.

The inter-system isolation must then be translated into a value isolation between the antenna systems used for the UTRAN and GSM networks.

The coupling between the two separate antennas is a function of both their operating bandwidth and their physical separation.

The operating bandwidth depends on the type of antenna. Owing to the proximity of the 1800 and 2000MHz frequency bands, it may be assumed that antennas designed for either band will



185

exhibit resonance across both bands. However, some degree of frequency selective isolation may be achieved in the case of 900 and 2000MHz antennas.

Again, the inter-antenna coupling will vary according to the type of antenna and any parasitic coupling with environmental clutter. For antennas that are sufficiently separated to be operating in the ‘far-field’ region, and in line-of-sight of each other, the standard free-space propagation formula can be used to estimate the coupling loss. When directional antennas are involved, a correction has to be applied to allow for the radiation pattern.

Lfs = 32.4 + 20*lg (f ) + 20lg (d) + Gtx + Grx

Where: Lfs is the free space loss (dB) f is the operating frequency in MHz

d is the separation in km Gtx is the transmit antenna gain (dBi) Grx is the receive antenna gain (dBi)

For estimating purposes, the following graphs show typical inter-antenna coupling for two dipoles:

Figure 37 Coupling between vertically and horizontally spaced antennas

Separate UTRAN and GSM antenna systems

In this case, an existing GSM site can be upgraded for UTRAN operation by installing an additional antenna system comprising antenna, feeder, and optional masthead amplifier.

Co-existence with GSM 900MHz BTS




• At the GSM 900 transmitter:

- If more than one transmitter is coupled to a single antenna, the transmit path will normally contain one (or more) circulators, in addition to either a frequency selective or ferro-magnetic hybrid combiner (and a duplexer if single transmit/receive antenna working is required)

- The presence of circulators will attenuate any received power from the ‘out-of-band’ transmitters, reducing the potential for the generation of intermodulation products in the transmitter output stages. The duplexer, where used, will provide another ~70dB isolation

- Recent developments in multi-carrier power amplifiers (both for equipment rack and masthead applications) will reduce the use of frequency selective and hybrid combiners, and the protection from out-of-band reverse power they can provide

• At the GSM 900 receiver:

- If a single antenna is used for both transmission and reception the GSM 900 receive path will contain a duplexer prior to the receiver distribution amplifier (in the case of a multiple transceiver site) and the receiver RF band pre-selector filter

- The stop band performance of the duplexer should typically be >70dB, which taken together with the RF pre-selector filter stop band attenuation of ~50dB should avoid intermodulation products and blocking. If greater isolation is necessary an additional 900MHz receive band filter can be added

- Particular care should be taken at sites where masthead amplifiers are used, in order to avoid blocking and intermodulation products in their receive path. Although the duplexer may have a stop band attenuation of 80 to 90dB in the 900MHz band, the device may not maintain this performance in the 2GHz band. Again, this can be overcome by the use of additional receive path filtering

• At the UTRAN transmitter:

- Although the UTRAN transmitter uses a multi-carrier amplifier, is protected from out-of-band reverse power by use of circulators



187

• At the UTRAN receiver:

- The UTRAN receiver desensitisation and intermodulation should not occur as the receiver is designed to operate in the vicinity of high power 900MHz GSM transmitters, and the duplexer and RF pre-selector filter are designed to offer adequate protection

- No in-band signals should be received from the 900MHz GSM transmitter, as the second harmonic of the highest 900MHz band channel falls outside the UTRAN receiver filter pass-band (<1920MHz)

- If the UTRAN receiver is affected by in-band spurii generated by the 900MHz GSM transmitter, the problem can be solved by fitting a GSM 900MHz transmit band-pass filter to the GSM transmitter

Co-existence with GSM 1800MHz BTS

• At the GSM 1800 transmitter:

- As for GSM 900 transmitter

• At the GSM 1800 receiver:

- As for GSM 900 receiver

- Intermodulation products from the UTRAN transmitter could be generated within the 1800MHz receive band. In the UK such products are likely to be of high order, as the maximum separation between UTRAN channels is 10MHz

• At the UTRAN transmitter:

- As for UTRAN transmitter co-located with GSM 900MHz BTS

• At the UTRAN receiver:

- As for UTRAN receiver co-located with GSM 900MHz BTS

- Depending on the 1800MHz channel spacing used at the base station, low order intermodulation products from the GSM 1800MHz transmitter may fall in the UTRAN receive band. Where these are a problem, suppression can take the form of an additional filtering of the 1800MHz band transmitter. Generation of intermodulation products in the passive antenna system is possible but unlikely




Co-existence with GSM dual band BTS

In the case of a dual band GSM base station site used to provide a UTRAN service, the effects of EMC with both the 900 and 1800MHz bands should be evaluated.

Figure 38 Antenna isolation for UTRAN BTS - separate UTRAN antenna at existing GSM BSS using dual band GSM antenna

Isolation between UTRAN and dual band GSM antennas

The line-of-sight path loss between the two antennas can be calculated from the standard far field free space propagation loss formula given in the Antenna coupling section earlier in this chapter.

Assuming that the antennas are spaced by 2.5m horizontally, are co-polar, both exhibit a gain of 10dBi at their operating frequency and 5dBi at other frequencies, and that there is no parasitic coupling through the antenna support structure, the following isolation values apply:

• At 900MHz

- Path loss = 40dB, less antenna gains of +10dBi and +5dBi

- Gives isolation = 40 – 10 – 5 = 25dB

• At 1800MHz





189

• At 2GHz



UTRAN receiver

The possible sources of interference are assessed in turn:

900MHz transmitter 2GHz spurii

1. Level of spurii generated by transmitter <-30dBm in the 2GHz receive band.

2. Attenuated in transmitter hybrid combiner by 30dB, resulting level <-60dBm.

3. Attenuated by two dual band (900/1800) diplexers by 20dB each, resulting level <-100dBm.

4. Attenuated by dual band (900/1800) antenna feeder by 2dB, resulting level <-102dBm.

5. Attenuated by antenna isolation of 25dB at 900MHz, resulting level <-127dBm.

6. Even if the spurii are broadband (across the whole 5MHz UTRAN channel) this level is below the noise floor of the 2GHz receiver (-105dBm). If the spurii are narrow band (more likely harmonics of 200kHz GSM channel), then the effective noise power in the 5MHz UTRAN receiver bandwidth would be reduced further by a factor of 10*lg (W interfered / W interfering).

900MHz transmitter 900MHz blocking signal

1. Transmitter output <+46dBm in the 900MHz transmit band.

2. Attenuated by two dual band (900/1800) diplexers by 0.5dB each, resulting level <+45dBm.

3. Attenuated by dual band (900/1800) antenna feeder by 2dB, resulting level <+43dBm.

4. Attenuated by antenna isolation of 25dB at 900MHz, resulting level <+18dBm.

5. Attenuated by 2GHz duplexer by 70dB, resulting level <-52dBm.

6. Attenuated by 2GHz receiver RF pre-selector filter by 40dB, resulting level <-92dBm.

7. The 2GHz receiver bandwidth is 5MHz, compared with the 200kHz bandwidth of the GSM transmitter, therefore a bandwidth correction factor of 10*lg (W interfered / W interfering) = 14dB can be applied, (assuming that there is only one GSM transmitter) giving a resulting signal level of –106dBm. This level is below the noise floor of the 2GHz receiver (-105dBm).

Both processes above are repeated for the 1800MHz GSM transmitter.




1800MHz GSM receiver

The analysis performed to assess the interference to the UTRAN receiver should then be repeated for the 1800MHz GSM receiver.

900MHz GSM receiver

The analysis performed to assess the interference to the UTRAN receiver should then be repeated for the 900MHz GSM receiver.

Single UTRAN and GSM antenna system

Depending on site conditions using a single UTRAN and GSM antenna system may prove an efficient alternative to adding a separate UTRAN antenna system at sites where a GSM antenna system already exists.

If this option is taken, in addition to the considerations outlined in the previous section, the following factors apply:

• Inter-band isolation from physical separation of antenna is lost

- This may be in the region of ~40dB for antennas separated horizontally by a distance of 3m

- When 900 and 2000MHz operation is required, a dual band (900/2000) antenna that has separate radiating elements with different connectors for the two bands may be used. These antennas provide isolation between the bands of ~30dB, and a third order intermodulation product power of <-150dBc for two carriers of 25W each

- When 1800 and 2000MHz operation is required, a dual band (1800/2000) antenna may be used. Such antennas are usually broadband devices covering the range 1710-2170MHz, and therefore provide no isolation between the 1800 & 2000MHz bands

- When these antennas are connected to 1800 & 2000MHz BTSs, additional isolation is required. This can be provided by a dual band (1800/2000) diplexer. Such devices can provide inter-band isolation of ~60dB, with the penalty of adding ~0.4dB insertion loss

• Use of a single antenna feeder for the GSM and UTRAN bands requires additional isolation

- When a single antenna is chosen for both bands, the single antenna feeder option is often used

- Again, this isolation can be provided by a dual band diplexer (also known as a cross-band coupler). Typical isolation is ~60dB, typical insertion loss is ~0.4dB. One dual band diplexer is used at the BTS end of the feeder to combine the signals from the GSM and UMTS BTSs. Another may also be used at the masthead if the dual band antenna has separate connectors for the two bands. In this case the insertion loss is increased to ~0.8dB for the pair of diplexers



191

• Masthead amplifiers for separate bands

- In the case of single antenna operation with a masthead amplifier (low noise pre-amplifier), the signal from the two bands has to be divided before it is presented to the masthead amplifier. This may be achieved in the single band antenna (which may have a separate connector for each band), but if not, it can be performed by a masthead dual band diplexer (cross-band coupler)

- Unfortunately, the insertion loss associated with the dual band diplexer impairs the noise figure of the masthead amplifier. This degrades the receiver system noise figure, and hence increases the level of the minimum discernible signal

- At sites where existing masthead amplifiers are used, their out-of-band performance when subject to power in the UTRAN transmit band must be checked. If the performance is inadequate, additional filtering should be used

UTRAN BTS spurious emissions

In view of the likely need for UTRAN and GSM BTSs to be installed at the same base station sites, the limits for UTRAN spurious emissions have been tightened in the GSM frequency bands.

Generally, the UTRAN BTS must comply with the Category B spurious conducted emission requirements given in the ITU-R Recommendation SM 329-7 summarised below:

Band Limit /dBm Measurement bandwidth

9kHz to 150kHz -36 1kHz

150kHz to 30MHz -36 10kHz

30MHz to 1GHz -36 100kHz

1GHz to 12.75GHz -30 1MHz

Table 35 Protection for GSM bands from UTRAN transmitter spurii




Additionally, the UTRAN 3GPP TS 25.104 includes the option to define the following tighter limits for spurious conducted emissions, in order to protect the GSM bands:

Band /Mhz Limit /dBm Measurement Bandwidth System Affected

876 to 915 -98 100kHz GSM900 Base Receive

921 to 960 -47 100kHz GSM900 Mobile Receive

1710 to 1785 -98 100kHz GSM1800 Base Receive

1805 to 1880 -57 100kHz GSM1800 Mobile Receive

1893.5 to 1919.6 -41 300kHz J-PHS (TDD System)

1920 to 1980 -94 100kHz UTRAN Base Receive

2100 to 2105 -30 +3.4(f-2100MHz) 1MHz Services in adjacent band

2175 to 2180 -30 +3.4(2180MHz –f) 1MHz Services in adjacent band

Table 36 Additional protection for GSM bands from UTRAN transmitter spurii

This table shows that:

• At a GSM base station, the level of UTRAN transmitter spurii should be <-98dBm in the GSM base receive band

- When antenna spatial isolation (for example, 46.4dB for a 2.5m spacing at 900MHz) is added, the sprurii level should be <-144.4dBm, which is below the receiver noise floor of –114dBm

- Where common antenna system elements are used, a relatively modest cross-band isolation (~40dB) will be sufficient to attenuate the spurii below the receiver noise floor

• GSM mobiles should experience UTRAN transmitter spurii at a level <-48dBm in the GSM receive band.

- When the mobile is operated at a reasonable distance from the BTS antenna (such as 30m) the antenna isolation (~62dB at 900MHz with line-of-sight) will attenuate the UTRAN transmitter spurii to –109dBm, which is approximately the noise floor of the receiver

- In practice, most mobiles are operated at greater distances and/or without line-of-sight to the base station antenna, and hence benefit from greater attenuation of the UTRAN transmitter spurii



193

Worked example of co-siting GSM and UTRAN

This example is based on the following system information:

Existing base station equipment

GSM 900 BTS using common transmit and receive antenna

New UTRAN base station equipment

- UTRAN BTS to use separate duplex antenna located at a horizontal separation of 2.5m from existing GSM antenna

- Separate 900 and 2000MHz feeders

Receiver characteristics

Receiver type Receiver Noise Floor

/dBm

Third Order Intercept

Point /dBm

1dB Desensit-

isation Blocking

Input Power /dBm

1dB CompPoint

/dBm

RF Pre-selector

Filter Rejection at

900 Tx. Band /dB

RF Pre-selector

Filter Rejection at

2000 Tx. Band /dB

UTRAN BTS -105 -6 +10 -18 110 N/A

GSM 900 BTS -114 -5 +8 -16 N/A 100

Transmitter characteristics

Transmitter type

Output Power /dBm

Duplexer 900MHz Stop-

band Attenuation /dB

Duplexer 2000MHz Stop-

band Attenuation /dB

Spurii Power in 900 Receive Band /dBm

Spurii Power in 2000 Receive Band /dBm

UTRAN BTS +43 80 N/A -98 N/A

GSM 900 BTS +45 N/A 70 N/A -30




Antenna isolation

The GSM and UTRAN antenna are separated horizontally by a distance of 2.5m.

Assuming that the antennas provide no directional or frequency selective isolation, that there is no parasitic coupling from environmental features, and there is a viable radio line-of-sight between them, the free-space loss between them will be:

• At 900MHz Lfs_900 = 32.4 + 20*lg (f ) + 20lg (d) + Gtx + Grx = 32.4 + 20*lg(900) + 20*lg(0.0025) +0 + 0 = 32.4+59.1-52.0 = 39.5dB

• At 2GHz Lfs_2000 = 32.4 + 20*lg (f ) + 20lg (d) + Gtx + Grx = 32.4 + 20*lg(2000) + 20*lg(0.0025) +0 + 0 = 32.4+66.0-52.0 =

46.4dB

GSM 900MHz transmitter

The GSM 900 transmitter will receive the following power from the UTRAN transmitter:

Prx_from_900tx = Ptx_by_2000tx – Lfs_2000 – Ldup_900 – Lcir_900 + 10*lg (W interfered / W interfering) = +43 – 46 – 70 –30 –14 = -117dBm

Where Prx_from_900tx is the 2GHz power received by the 900MHz transmitter Ptx_by_2000tx is the power transmitted by the 2GHz transmitter Lfs_2000 is the antenna isolation at 2GHz Lcir_900 is the 900MHz transmitter circulator directivity (30dB) Ldup_900 is the out-of-band stop-band attenuation of the 900MHz band duplexer W interfered is the receiver bandwidth

W interfering is the measurement bandwidth of the interfering signal

A power of –117dBm at 2GHz received by the 900MHz transmitter output port is unlikely to generate any intermodulation products of sufficient amplitude to appear above the noise floor of either the 900MHz or 2GHz band receivers.

UTRAN transmitter

The UTRAN transmitter will receive the following power from the GSM 900MHz band transmitter:

Prx_from_2000tx = Ptx_by_900tx – Lfs_900 – Ldup_2000 – Lcir_2000 = +45 – 39.5 – 80 –30 = -104.5dBm

Where Prx_from_2000tx is the 900MHz power received by the 2GHz transmitter Ptx_by_900tx is the power transmitted by the 900MHz transmitter Lfs_900 is the antenna isolation at 900MHz Lcir_2000 is the 2GHz transmitter circulator directivity (30dB) Ldup_2000 is the out-of-band stop-band attenuation of the 2GHz band duplexer



195

A power of –104.5dBm at 900MHz received by the 2GHz transmitter output port is unlikely to generate any intermodulation products of sufficient amplitude to appear above the noise floor of either the 900MHz or 2GHz band receivers.

GSM 900MHz receiver

This section considers the main mechanisms by which the 2GHz transmitter may affect the performance of the GSM 900MHz receiver, in relation to the inter-system isolation criteria described earlier.

Desensitisation (blocking)

At the ‘wanted band’ receiver, the total power received from the ‘unwanted band’ BTS should be 5dB below its receiver 1dB compression point (P1dB)

• GSM 900MHz receiver

The minimum blocking antenna isolation (Lant_iso) required between the BTSs is such that the power received from the out-of-band transmitter is less than that producing a 1dB reduction in the sensitivity of the wanted band receiver.

i) Ptx_tot – Zdup - Lant_iso + 10*lg (W interfered / W interfering) <= Prx_desens = +43 – 100 – 46.4 - 34 = -137.4dB

where: Ptx_tot = total power transmitted by potential interferer Zdup = attenuation of receiver system duplexer and RF pre-selector filter in unwanted band Prx_desens = input receiver power in unwanted band causing a 1dB sensitivity drop.

A blocking interference level of –137.4dBm is > than the 900MHz receiver 1dB desensitisation level of +8dBm.

Thus the performance of the 900MHz receiver should not be compromised.

ii) The total power received from the ‘unwanted band’ BTS should be 5dB below its receiver 1dB compression point (P1dB)

Ptx_tot – Zdup - Lant_iso + 10*lg (W interfered / W interfering) < (P1dB) - 5 +43 – 100 – 46.4 - 34 < (-16) - 5 -137.4dB < -21 , (condition met)

A blocking interference level of –137.4dBm is < than the 900MHz receiver 1dB compression point of –16dBm less 5dB.

Thus the performance of the 900MHz receiver should not be compromised.

• UMTS 2GHz receiver




The minimum blocking antenna isolation (Lant_iso) required between the BTSs is such that the power received from the out-of-band transmitter is less than that producing a 1dB reduction in the sensitivity of the ‘wanted band’ receiver.

i) Ptx_tot – Zdup - Lant_iso <= Prx_desens = +45 – 110 – 39.5 = -104.5dB

A blocking interference level of –104.5dBm is > than the 2GHz receiver 1dB desensitisation level of +10dBm.

Thus the performance of the UTRAN 2GHz receiver should not be compromised.

ii) The total power that is received from the ‘unwanted band’ BTS should be 5dB below its receiver 1dB compression point (P1dB)

Ptx_tot – Zdup - Lant_iso < (P1dB) - 5 +45 – 110 – 39.5 < (-18) - 5

-104.5dB < -23dBm, (condition met)

A blocking interference level of –104.5dBm is < than the UTRAN 2GHz receiver 1dB compression point of –18dBm less 5dB.

Thus the performance of the UTRAN 2GHz receiver should not be compromised.

Spurious responses

Owing to the use of duplexers at both the 900MHz and 2GHz receivers, with stop-band rejection of 100 and 110dB respectively, all receiver spurious responses are below the noise floor of each receiver.

Interference from intermodulation products

At the ‘wanted band’ receiver, the power of the third order intermodulation products that are generated should be 10dB below the ‘wanted band’ receiver noise floor.

• At GSM 900MHz receiver

Pint = Power from ‘other band’ interferer at ‘own band’ receiver input = Ptx_tot - Lant_iso - Zdup + 10*lg (W interfered / W interfering) = +43 – 46.4 – 100 – 14 = -

117.4dBm IMP3 = Power of third order product generated at ‘own-band’ receiver = (3 * Pint) – (2 * TOI) = (3 * -117.4) – (2 * -5) = -352.2 – (-10) = -342.2dBm Where: IMP3 = Power of third order intermodulation product

TOI = Third Order Intercept point specified for the receiver input Ptx = Power transmitted from interfering signal

W interfered is the receiver bandwidth W interfering is the measurement bandwidth of the interfering signal



197

A third order intermodulation interference level of –342.2dBm is >10dB below the 900MHz receiver noise floor of –114dBm. Thus performance of the 900MHz receiver should not be compromised.

• At UTRAN 2GHz receiver

Pint = Power from ‘other band’ interferer at ‘own band’ receiver input = Ptx_tot - Lant_iso - Zdup = +45 – 39.5 – 110 = -104.5dBm

IMP3 = Power of third order product generated at ‘own-band’ receiver = (3 * Pint) – (2 * TOI) = (3 * -104.5) – (2 * -6) = -313.5 – (-12) = -301.5dBm

A third order intermodulation interference level of –301.5dBm is >10dB below the 2GHz receiver noise floor of –105dBm. Thus performance of the 2GHz UTRAN receiver should not be compromised.

Transmitter spurii raises receiver noise floor

At the ‘wanted band’ receiver, the power that is received from the ‘unwanted band’ BTS should be 10dB below the ‘wanted band’ receiver noise floor.

• At GSM 900MHz receiver

Prx_int = (Ptx + CIRtx) – Ltx_rej - Lant_iso + 10*lg (W interfered / W interfering) = (-35) – 100 – 46.4 – 14 = -195.4dBm

A received interference level of –195.4dBm is >10dB below the 900MHz receiver noise floor of –114dBm. Thus performance of the 900MHz receiver should not be compromised.

• At UTRAN 2GHz receiver

Prx_int = (Ptx + CIRtx) – Ltx_rej - Lant_iso = (-30) – 110 – 39.5 = -179.5dBm

A received interference level of –179.5dBm is >10dB below the 900MHz receiver noise floor of –105dBm. Thus performance of the 900MHz receiver should not be compromised.




7.3. Use of repeaters

Repeaters are commonly used in GSM networks to extend coverage from an existing base station in a contiguous manner (such as along a road), to an isolated area, or within a structure such as a building or tunnel. This section discusses the differences that may be encountered when using repeaters in UTRANs.

The following diagram illustrates contiguous and scattered repeater coverage respectively:

Figure 39 Use of repeaters to extend contiguous or scattered coverage

Introduction

As in any cellular radio network, the maximum repeater range is determined by the repeater path-loss power budget. However, as the UTRAN employs a CDMA air interface, this power budget is not static but a function of traffic load.

The path-loss power budget for both the donor cell and the repeater cell vary according to traffic load. The path-loss power budget for the donor cell will be further reduced by the noise and traffic signal power it receives from the repeater.

So in a UTRAN, it is necessary to consider the composite (donor to repeater and repeater to mobile) path-loss power budget when evaluating the coverage extension derived. Although a unique maximum solution exists for the composite path-loss power budget, many pairs of loss values may satisfy the solution.

Donor cell coverage shrinks because mobiles must transmit more power in order to overcome the additional repeater noise. In the case of scattered range extension repeaters this may not cause difficulty, as the donor sector may not be required to cover a large area. For contiguous



199

range extension repeaters, it is possible to lose more network coverage by shrinkage of the donor cell than is gained in repeater range extension.

Design

The general method for designing repeater UTRAN sites is:

• Identify the candidate repeater and donor base station and estimate the line-of-sight free-space path-loss between them (in the same manner as for a GSM repeater)

• From this path-loss estimate, the allowed loss between the repeater and the mobile can uniquely be determined. The RF coverage associated with the allowed loss is calculated in a manner similar to that for GSM, and requires the path loss algorithms used in Airpro (or similar)

• Select antenna gain and repeater amplifier gain

- In general, higher gain provides greater range, but has a more serious impact on capacity owing to the generation of unwanted noise. This is described in more detail later

- The repeater donor antenna is of narrow azimuth beam width, with line-of-sight to the donor base station. The base station and repeater subscriber antennas are chosen to shape the required coverage area (for example, azimuth beam-width 200 to 900)

• Noise entering the donor receiver

- Each repeater generates noise that reaches the donor receiver

- The magnitude of the noise is a function of the separation between the donor and repeater, the repeater amplifier gain, gain of the antennas, and the number of repeaters used in each donor sector. Thus excess gain should be avoided, and no more than two repeaters per donor sector are recommended

• Capacity

- Repeaters cannot be used to add capacity to a cell (or network) and in fact total cell capacity is often reduced. The effect of repeaters can be thought of as spreading the existing BTS cell capacity over a larger or different area




Uniform range extension

In principle, repeaters could be used to uniformly extend the coverage of the donor cell, thereby reducing the number of BTSs required during the initial network roll-out.

However, this approach is not recommended as in many situations the repeater range extension will be more than offset by shrinkage in the donor cell. Also, as up to 6 repeater sites are required (2 per donor cell sector), the site acquisition and construction costs outweigh any saving in BTS equipment cost.

Cascaded range extension

Again, in principle, repeaters can be cascaded to provide greater range extension. This technique involves a mobile communicating with a repeater, which is itself linked to a second repeater which is finally linked to the donor BTS cell. This approach is often considered for coverage of roads into small towns.

Cascading of no more than two repeaters is recommended. However, there are situations where a BTS base station can be designed to have a larger range (particularly directional range along a road) than the addition of a repeater to a donor BTS cell.

Repeater gain and composite noise factor

The repeater to donor BTS cell link power budget is a key design element in determining the repeater range.

Figure 40 Calculation of composite noise factor for repeater and donor cell



201

Composite noise factor at repeater input is:

F’c = Fr + (L-1)/Ga + L(Fbs-1)/Ga (see Figure 40 above for definitions) F’c = Fr + (4π/λ)2 Fbs/Ga GdGbs – 1/Ga at the repeater input

Thus F’c is the equivalent noise seen by mobiles operating in the repeater cell.

This composite noise factor can be referenced to the BTS donor cell antenna output:

Fc = Fbs + GdGbs(4π/λ)2 (Ga Fr-1) at the donor cell BTS antenna output.

This is the composite noise factor seen by mobiles operating in the donor cell.

By comparing the functions F’c and Fc it can be seen that high repeater gain:

• Reduces the noise seen by mobiles operating in the repeater cell, and hence increases the operating range of the repeater, but:

• Increases the noise seen by mobiles operating in the donor cell, thereby reducing its coverage

Donor cell shrinkage

The minimum discernible signal density at the repeater (MDS’) is given by:

MDS’ (dBm/Hz) = I’ (dBm/Hz) –g + d Where g = processing gain d = minimum Eb/N0 I’ = internal and external interference power at repeater input Assume g = 27dB and d = 8dB MDS’ (dBm/Hz) = I’ (dBm/Hz) –27 + 8 = I’ –19 at the repeater input

The interference power density I’ (dBm/Hz) = Nth + F’c + Mi Where Nth = thermal noise power density = -174dBm/Hz

F’c (dB) = composite noise figure at repeater input MI (dB) = combined user adjacent carrier and IMD interference (at 50%

pole capacity assume to be 3.4dB)

MDS’ (dBm/Hz) = -189.6 + F’c at repeater input, or, MDS (dBm/Hz) = -189.6 + Fc at donor BTS antenna output with repeater

With no repeater, MDS (dBm/Hz) referenced to the donor BTS antenna output is found by replacing Fc with Fbs :

MDS (dBm/Hz) = -189.6 + Fbs at donor BTS antenna output with no repeater




Donor cell shrinkage s(dB) is defined as the increase in the minimum discernible signal required to overcome repeater noise compared with the case without a repeater, while supporting the same minimum Eb/N0 level. Hence:

s(dB) = ∆ MDS = Fc - Fbs

The following table gives some typical values that illustrate the effect of donor cell shrinkage:

Repeater Uplink Gain /dB

Separation of Repeater from Donor Cell /km

Donor Cell Shrinkage /dB

60 1 3 5 7

0.5 <0.5 <0.5 <0.5

70 1 3 5 7

4 0.5 <0.5 <0.5

80 1 3 5 7

11.5 4 2 1

85 1 3 5 7

16.5 7.5 4.5 3

90 1 3 5 7

21.5 12 8.5 6

Table 37 Donor cell shrinkage

Median repeater link budget calculation

The unadjusted repeater link budget LB’rep is the median path-loss between the EIRP of the mobile and the minimum discernible signal MDS’ at the repeater, offset by the gain of the repeater subscriber antenna Gr.

Thus LB’rep (dB) = EIRP (dBm) – MDS’ (dBm/Hz) + 10lg(BW) + Gr

Where BW is the RF channel bandwidth, and the term 10lg(BW) changes the units from relative (dBm/Hz) to specific power level (dBm).

As the link budget is directly related to the MDS’, the median path-loss is directly related to the maximum useable repeater amplifier gain.



203

Figure 41 Median path-loss model

Median repeater link budget adjustment

The median repeater path-loss must be adjusted to take account of soft handover gain, diversity gain, fade and indoor penetration margins.

In the case of continuous range extension, if BTS base stations are used, then where cells overlap, the mobile benefits from soft handover gain, thereby improving the link power budget. If a repeater is used instead, in the area of overlapping coverage from the donor BTS cell and the repeater cell there is no soft hand-off gain, and the mobile link power budget may suffer.

In this case the range of the donor BTS cell is reduced by a combination of loss of soft handover gain and shrinkage (s [dB]) due to repeater noise.

Repeater donor cell antennas

A critical aspect of repeater installation is the isolation between the donor and subscriber antennas. Typically, isolation must be more than 10 to 15dB in excess of the repeater amplifier gain in order to avoid instability.

Thus if a repeater amplifier gain of 90dB is used, antenna isolation as high as 105dB may be required. This can be difficult to achieve with reasonable antenna separations.

An antenna parameter that influences the physical separation required is the electric field suppression at +90 degrees to the bore-sight. This is because most repeater applications place the antennas vertically, with the subscriber antenna at the top of the mast and the donor antenna below.




If isolation greater than 15dB can be achieved, the repeater will benefit from an additional margin against instability owing to random reflections from changing environmental conditions.

Donor repeater antennas that feature side-lobe suppression of 50dB in the electric field plane and 40dB in the magnetic field plane, together with a front to back ratio greater than 50dB can be used. These also minimise the parasitic radiation from any metallic mounting structure.

Costs associated with the use of a high specification donor antenna can be offset against the use of a lower mast (as antenna separation can be reduced).

Antenna gain and beam-width should also be optimised when donor antennas are selected. Higher gain reduces the required antenna separation, as the overall repeater gain may be reduced in direct proportion to the antenna gain. Narrower beam-widths reduce the interference that may make selection of the repeater difficult, as it improves the repeater’s ability to discriminate between the target cell and neighbours. Typically an antenna gain greater than 24dBi and a beam-width less than 120 should be sought.

Summary

Repeaters are an economic means of extending coverage from an existing base station BTS into structures such as buildings and tunnels, and to isolated areas beyond the contiguous coverage area of the base station. However, the traffic load capacity of the donor BTS base station is never increased, and often the effect on coverage can best be viewed as ‘moving’ coverage from the donor BTS cell to the repeater cell. A net coverage gain is not guaranteed.

The principal factors to consider are:

• Repeaters add noise to the donor cell. This limits both repeater and donor cell coverage. Under many conditions, more donor BTS cell coverage is lost than is gained in repeater cell coverage. Both donor cell and repeater cell coverage must be re-calculated when a repeater is used. These effects severely limit the applications where repeaters are suitable for providing contiguous range extension

• Using repeaters for scattered (non-contiguous) range extension (structures and isolated areas) appears to offer the most successful application. In particular if the repeaters can use donor BTS sectors where long range is not required

• There is no single path link budget that is applicable to repeater use, each application is unique, and has to be individually calculated



205

7.4. Use of microcells

GSM networks may consist of:

• Macrocells with antennas mounted above the general rooftop height serving mobiles at a range of 0.75 to 35km

• Microcells with antennas mounted below the general rooftop height (propagation largely dictated by ground clutter), serving mobiles at a range of 200 to 750m

In GSM networks, microcells may be used to extend coverage in black spots such as buildings and tunnels, and to enhance network capacity in local hot spots such as indoor offices and shopping centres. A network comprising macrocells and microcells may be configured a a multi-layer network, in which slow moving mobiles are directed to use the microcell layer (the lower layer) and fast moving mobiles to use the macrocells (the upper layer). This maximises the efficiency of frequency re-use, while maintaining signalling traffic and switching load caused by inter-cell handover at acceptable levels.

As in GSM, a UTRAN may also use a combination of macrocells and microcells. The microcells can provide:

• Coverage in ‘holes’ and black spots, such as shadow areas caused by terrain features

• Capacity in local hot spots such as a shopping area or industrial estate. The microcell may be added within the coverage area of an existing macrocell to serve the local traffic, thereby releasing capacity at the macrocell to service the general cell area

• Indoor coverage and capacity. Macrocells can provide indoor coverage by the use of greater outdoor power density. However, this reduces the traffic capacity available in the macrocell. As an alternative, deploying indoor microcells provides improved coverage within the building and improves performance of the outdoor macrocell, by allowing more of the available link power budget to be used for capacity. In this case the screening effect of the building structure acts beneficially, providing a degree of isolation between the two co-channel transmitters

• Coverage along rural roads and in isolated communities. Microcells can be an economic solution to providing longitudinal coverage areas, where little capacity is required, the individual microcells can be daisy-chained. This approach can also be used for coverage in isolated communities, where contiguous coverage would be uneconomic




Embedded and non-embedded microcells

In principle, microcells may be deployed in either an embedded (i.e. analogous to hierarchical cells in a GSM network) or non-embedded (for example, indoor coverage extension in a single layer GSM network) manner.

Embedded microcells

To date, embedded microcells that act in a hierarchical manner have not been deployed in CDMA networks. Embedded microcells can create excessive areas of overlapping coverage, resulting in ‘pilot pollution’. This occurs when more viable pilot signals exist than rake receiver channels, and the pilots that cannot be used by the receiver act to raise the noise floor.

Non-embedded microcells

Non-embedded microcells can take two forms.

• Cell-splitting configuration

The microcell may be installed at the boundary of two or three existing macrocells in order to provide additional capacity at the edges of the macrocells. In the presence of the microcells, the coverage of the macrocells shrinks (see Figure 42).

• Longitudinal configuration

This arrangement may be used to provide coverage along roads, in areas where contiguous off-road coverage is not required.

Figure 42 Embedded and non-embedded microcell deployment



207

Microcell problems

Two of the most common problems that may be experienced when microcells are deployed are:

• Near-far effect

- When mobiles move from the macrocell into the microcell, they transmit a high power level at the edge of the macrocell. This can result in interference at the microcell just before handover is completed.

• Optimum handover power

- At the edge of the microcell the mobiles transmit relatively little power. As they move into the macrocell this power may be insufficient for reception at the macrocell and calls may be lost.

Co-channel macrocells and microcells

Satisfactory operation of a number of microcells within a macrocell network is very difficult to engineer owing to the pilot pollution effect mentioned earlier. The rake receiver used in a CDMA network can handle up to four soft handover links. If a fifth strong pilot signal becomes available it cannot be constructively combined in the rake receiver, and results in a rise in the noise floor, representing interference.

Networks are therefore designed to restrict to four , the number of strong pilot signals that can be received at any point within the network. Each cell (sector) radiates a pilot, so a tri-sectored base station with embedded co-channel microcells can become difficult to design without undue levels of interference.

Increased interference resulting from the use of co-channel microcells will reduce the traffic capacity of the macro-layer. Consequently, any theoretical increase in network capacity provided by the co-channel microcell is not simply added to that of the existing macrocell in a linear manner.

As well as the effect of mutual interference, the proportion of handover activity at any instant affects network capacity, both on the air interface and in terms of the network hardware and back-haul infrastructure.

In a network with layers of microcells and macrocells, the proportion of calls in soft handover will be greater than in a single macrocell layer network. Experience with American CDMA networks indicates that the proportion of resources used for soft handover from the embedded microcells is generally 40-50% and can rise to 60-70%.

As far as the network is concerned, the additional call processing required to support the increase in soft handover activity is not normally a problem, although care has to be taken when sizing the RNC and MSC to ensure that the network elements can handle the anticipated load.




Macrocells and microcells on different channels

The main factors to evaluate when considering macrocells and microcells operating on different radio channels are:

• Ability of mobile to report Eb/Io of different carriers

- The UTRAN standard allows for this, but in practice the level of support in early models may vary. Matching the footprint of the carriers on each frequency can mitigate this.

• Adjacent channel interference

- Although this has not proved a large effect in American CDMA networks, it has been reported that call quality can be degraded when one carrier becomes excessively strong compared to another. The wider bandwidth used for UTRAN transmissions is expected to further minimise this problem.

The frequency band plan designed for use in the first countries to licence UTRANs assigns adjacent channels to the same licensee.

Matching the footprint of the carriers on each frequency can mitigate this effect.

• Traffic distribution

- Sophisticated hashing algorithms can be implemented to achieve a balance of traffic load between the RF channels.

• Multi-carrier handover and trigger

Cells that only transmit pilot carriers may be required at border areas to improve handover trigger performance. This involves additional hardware cost.

• RNC and MSC loading

- Additional call processing resources in the RNC and MSC will be needed to handle the second carrier and inter-carrier handover activity.

Dual layer UTRAN

In narrow band FDMA / TDMA systems such as GSM, multi-layer networks provide both additional capacity through more efficient frequency re-use and improved link quality by segregating certain types of calls (for instance by forcing fast mobiles to use macrocells and thereby minimise the rate of handover).

In wide band CDMA systems such as that used for UTRANs, the rationale for the use of dual layer macrocell and microcell networks is not well developed. However, there may be situations where different layers could provide different types of service. For instance, macrocells could provide wide area voice and low data rate services, and microcells could provide high data rate services for a sub-set of the coverage area. However, such an approach using embedded



209

microcells would be complex to engineer, and the potential benefits would need to be overwhelming for such an approach to be recommended.

At present UTRANs have no mechanism to distinguish between microcells and macrocells, whether co-channel or not, when access control or call processing is considered.

7.5. Masthead amplifiers

Because of the increased feeder cable attenuation at 2GHz compared with that at 900MHz, and the losses associated with the use of tri-band diplexers, a masthead amplifier can yield a significant reduction in the effective receiver system noise figure.

This effect is illustrated by the following example, in which it is assumed that:

• An existing GSM 900 BTS site is to be upgraded to offer UTRAN service in addition to maintaining its GSM 900 service

• Mast loading constraints (commercial, environmental, or physical) mean that additional UTRAN antennas and feeders cannot be used:

- An existing GSM 900 antenna is replaced with a dual band GSM 900 and UTRAN antenna

- An existing 25m length of 7/8” foam dielectric cable feeder is used to serve the dual (GSM/UTRAN) band antenna




This configuration is shown in the following diagram:

Figure 43 Application of masthead amplifier

Without masthead amplifier

900MHz transmit path

The losses associated with the 900MHz transmit path are:

• Equipment room tails and connectors 0.3dB

• Equipment room dual band diplexer 0.2dB

• Antenna feeder 1.0dB

• Masthead dual band diplexer 0.2dB

• Antenna tail and connectors 0.3dB

• Total Loss 2.0dB

2GHz transmit path

The losses associated with the 2GHz transmit path are:





211





900MHz receive path

The noise figure of the existing 900MHz BTS receiver will be increased from its current value by:

• Equipment room dual band diplexer insertion loss 0.2dB

• Masthead dual band diplexer insertion loss 0.2dB

• Total increase in noise figure ~0.4dB

2GHz receive path

The noise figure of the UTRAN BTS receiver will be increased from the quoted 2dB by:

• Equipment room cables and connector loss 0.5dB


• Antenna feeder loss 1.65dB


• Masthead tail and connector loss 0.5dB


• Total effective receiver system noise figure ~5.65dB




With masthead amplifier

900MHz transmit path

The losses associated with the 900MHz transmit path are:







2GHz transmit path

The losses associated with the 2GHz transmit path are:





• Masthead amplifier transmit bypass duplexer loss 0.25dB



900MHz receive path

The noise figure of the existing 900MHz BTS receiver will be increased from its current value by:






213

2GHz receive path

The noise figure of the UTRAN BTS receiver will be increased from the quoted 2dB by:

Fin = F1 + [(F2 – 1) / G1)] + [(F3 – 1) / G1 G2)] + [(Fn – 1) / G1 G2 …Gn)] Where Fin = Receiver system Noise Factor F1…n = Noise Factor of Sage 1…n G1…n = Gain of Stage 1…n F1 = Noise Factor of Stage 1 = Noise Factor of antenna tail and connectors = alog (0.5/10) = 1.12 F2 = Noise Factor of Stage 2 = Noise Factor of masthead amplifier = alog (1.8/10) = 1.51 F3 = Noise Factor of Stage 3 = Noise Factor of feeder system to BTS input

= alog (3.15/10) = 2.07 F4 = Noise Factor of Stage 4 = Noise Factor of BTS input = alog (2/10) = 1.58

G1 = Gain of Stage 1 = Gain of antenna tail and connectors = 1 / [alog (0.5/10)] = 0.89

G2 = Gain of Stage 2 = Gain of masthead amplifier[1] = alog (3.15/10) = 2.07 G3 = Gain of Stage 3 = Gain of feeder system to BTS input = 1 / [alog (3.15/10)] = 0.48 G3 = Gain of Stage 4 = Gain of BTS receiver RF amplifier = 1 / [alog (15/10)] = 31.62

Fin = F1 + [(F2 – 1) / G1)] + [(F3 – 1) / G1 G2)] + [(Fn – 1) / G1 G2 …Gn)] Fin = 1.12 + [(1.51 – 1) / 0.89)] + [(2.07 – 1) / 0.89 x 2.07)] + [(1.58 – 1) / 0.89 x 2.07 x 31.62)] = 1,12 + 0.57 + 0.58 + 0.01 = 2.28 NFin = 10 log 2.28 = 3.58dB

Total receiver system noise figure ~3.6dB

[1] In practice, the effective gain of the masthead amplifier will be reduced such that the receive path between the masthead amplifier output and the BTS receiver input has unity gain, in order to avoid compression of the BTS receiver dynamic range.

This example is for illustration purposes only and for clarity it contains several simplifications.




Summary

The effect of using the masthead amplifier in the UTRAN receive path is summarised in the following table.

900MHz Transmit Power Reduction

2GHz Transmit Power Reduction

900MHz Receiver System Noise

2GHz Receiver System Noise

With masthead amplifier

-2dB -3.65dB 7.4dB 5.65dB

Without masthead amplifier

-2dB -3.9dB 7.4dB 3.58dB

Table 38 Effect of masthead amplifier

In this example, the addition of the masthead amplifier can improve the UTRAN receiver system noise figure by ~2dB.

7.6. Practical antenna considerations

When assessing the base station antenna performance required, in addition to the general criteria such as gain, radiation pattern, and physical aspects, the following additional factors may need particular attention when considering UTRAN and dual band GSM/UTRAN antennas:

• Generation of intermodulation products

• Front-to-back ratio

• Variable electrical down-tilt

Intermodulation products

Initially, most UTRANs will be deployed by existing GSM networks. Most of the GSM networks will be operating at 900MHz, some at 1800MHz and a few at both 900MHz and 1800MHz.

In the early stages of implementing UTRANs, the current GSM network operators will wish simply to add UTRAN base stations at their GSM sites. In many areas, owing to weight and space limits, and economic/ environmental constraints, it is anticipated that the existing GSM antennas will be replaced with dual (or tri) band GSM/UTRAN antennas.

Use of multi-band antennas at base station sites that may be shared with other system operators, will be susceptible to the generation of intermodulation and cross-modulation products, owing to the greater number of channels radiated, and cables, connectors and couplers used.

Intermodulation occurs when two or more signals pass through a non-linear component and produce mixing products. Although most pronounced in active system components, it can occur at bi-metallic junctions which are subject to corrosion (such as found in cables, connectors, antennas, multi-coupling components, and non-welded mast joints).



215

This effect can be minimised by:

• Reducing the number of cable joints and connectors

• Selection of high quality connectors (with good intermod specifications)

• Choosing antennas specified to generate low level intermodulation products

- Many GSM dual band 900 and 1800MHz antennas are specified to generate third order intermodulation products <-150dBc for two signals at +43dBm. This should be the minimum acceptable performance for multi-band GSM/UTRAN antennas.

A Lucent preferred supplier is currently using–162dBc as a working target, and has been given a specification of –159dBc by UTRAN network operators. Although these GSM/UTRAN antennas have yet to enter full production, measurements on prototypes indicate an intermodulation product level of ~-175dBc.

• Avoiding corrosion by means of regular maintenance and protective coatings

• Use of welded (rather than bolted) joints in antenna support structures

• Good base station engineering practices. For example, as a minimum adherence to the following standard: UK Department of Trade & Industry (DTI) Code of Practice for Radio Site Engineering (MPT 1331)

Front-to-back ratio

The front-to-back ratio is a measure of the directivity of the antenna - the extent to which it leaks radiation in the opposite direction from its bore-sight.

Current dual band GSM 900 and 1800MHz antennas have a front-to-back ratio specified at >30dB. The effect of energy not transmitted along the bore-sight in a GSM network is an increase in interference to neighbouring co-channel cells, and reduced overall network capacity.

During the initial deployment of UTRANs it is likely that many BTS sites, although using tri-sectored antennas, will be radiating the same RF channel on all base station sectors, and using code discrimination. Although inter cell interference is minimised by use of different codes, any co-channel energy that strays from the ‘wanted’ to an ‘unwanted’ cell, will manifest itself by raising the noise floor in the unwanted cell, and thereby reduce cell capacity.

Consequently, the quality of the beam pattern and magnitude of the front-to-back ratio for UTRAN antennas is equally important, or more so, than for GSM antennas.

Variable electrical down-tilt

GSM networks commonly use antennas that are down-tilted in the elevation plane, in order to constrain cell area and maximise signal level within the cell. This down-tilt can be achieved by




mechanically tilting the whole antenna assembly (mechanical down-tilt) or electrically altering the phase of the signals applied to the individual elements that comprise the antenna, thereby producing a composite beam that is tilted (electrical down-tilt).

The mechanical tilt can be continuously varied at the time of installation, and subsequently during network optimisation. However, if masts have to be climbed to adjust the tilt brackets, weather conditions (such as icing and electrical storms) and staffing requirements/training may be make re-adjustment costly.

The electrical tilt is usually fixed at the factory at the time of manufacture, and cannot be subsequently changed. However, a mechanical up or down-tilt may be added to an antenna that has been manufactured with an electrical down-tilt. In addition to producing a better ground level radiation pattern than an equivalent mechanically down-tilted antenna, electrical beam tilt can be used without the tilt being visually obvious. When large tilt angles are used, some local residents have complained that the antennas are ‘looking into’ their houses.

Recently a few GSM antennas have been introduced that have the facility to vary the electrical down-tilt, for example between 2 and 80.

This may be achieved by two means:

• Mechanically operated phase shifter

- The mechanical phase shifter is a dielectric or ferromagnetic device. This is generally adjusted by a user-adjustable knob on the antenna. Manual access to the antenna is required

• Electrically operated phase shifter

- The electrically operated phase shifter is controlled by an electrical signal, either applied directly at the antenna, or which can be extended to the equipment room

- The electrical signal generally controls a stepper motor servomechanism in the antenna that in turn operates a mechanical phase shifter. True, fully electric phase shifters are not normally used in the low-cost antennas used for GSM applications

Several manufacturers of UTRAN antennas have remotely controlled electrically operated variable down-tilt antennas. This moves control of the variable electrical down-tilt from the base station equipment room to the network operations and maintenance centre. This has several benefits:

• Avoids the need to visit each base station site in order to change the down-tilt

- It has been reported that site visits, climbing masts, and adjusting antennas can generate increased complaints from local residents

• Ability to remotely maintain a database of the actual down-tilt in use at each site. This prevents staff making local changes at base stations and not recording the results



217

Remote adjustment of electrical down-tilt may be used to alter antenna radiation pattern, and hence the cell coverage area, on a frequent basis. It has been proposed that this may be useful to:

• Extend cells that cover office areas during working hours, to a larger area when the traffic density in the office area falls outside working hours

• Extend cover from the urban area to more of the outlying sub-urban areas when more subscribers are likely to be at home in the evenings and weekends

The Lucent BTS supports remotely adjustable antennas by providing 8 relay-driven discrete outputs.

7.7. Transmit diversity

GSM networks use the following diversity techniques:

• Time diversity

- Uplink and downlink. Implemented by a combination of data redundancy and bit interleaving.

• Space or polarisation uplink receive diversity

- Implemented by use of two receive paths at the base station. The two receive paths are served by different antenna elements, which use physical separation or orthogonal polarisation to minimise the correlation between the signals they receive.

The diversity gain (effective path-loss power budget improvement) achieved depends on whether polarisation or space diversity antennas are used, the type of diversity combining used, and the operating environment (for example, rural or urban). Under favourable circumstances a diversity gain up to 4dB can be achieved.

Downlink receive diversity is not used owing to practical difficulties in implementing two antennas with non-correlated receive paths in the mobile.




UTRAN implementation

In addition to the above methods, UTRANs may also incorporate space or polarisation downlink transmit diversity.

This is feasible in a UTRAN because when the two transmissions are received at the mobile’s rake receiver, information can be extracted from both transmission paths.

Two forms of UTRAN base station transmit diversity may be used:

• Space Time Block Coding Based Transmit Antenna Diversity (STTD)

- The open loop downlink transmit diversity employs space time block coding based diversity (STTD). STTD encoding is optional in the UTRAN; support for STTD is mandatory in the mobile.

Channel coding, rate matching, and interleaving is done as in the non-diversity mode.

• Time Switched Transmit Diversity (TSTD)

- Transmit diversity in the form of Time Switched Transmit Diversity (TSTD) can be applied to the Synchronisation Channel (SCH). TSTD is optional for the SCH; support for TSTD is mandatory in the mobile.

The application of base station diversity modes to physical channels is summarised below:

Channel Open Loop Mode TSTD

Open Loop Mode STTD

Closed Loop Mode

P-CCPCH Available

SCH Available

S-CCPCH Available

DPCH Available Available

PICH Available

PDSCH (Associated with DPCH)

Available Available

AICH Available

Table 39 Transmit diversity on downlink physical channels

Note: Simultaneous use of STTD and Closed Loop Modes on DPCH and PDSCH is not allowed.



219

Effect of UTRAN implementation

Simulations indicate that implementing base station transmit diversity can reduce the bit energy to noise ratio [Eb/N0] for each type of subscriber service by ~2dB.

7.8. Coverage areas for different services

As described previously, the path-loss link power budget varies according to the type of subscriber service required. The minimum received power levels and associated maximum permitted path losses detailed in Chapter 6 (Air interface link power budget section) can be used as the basis for planning the minimum acceptable received power level for each type of subscriber service.

For example, assuming the system characteristics given in the Air interface link power budget section:

Subscriber Service Uplink Minimum Receiver Input Power Level /dBm

Downlink Minimum Receiver Input Power Level /dBm

8 kbs-1 circuit switched voice -121.7 -114.2

13 kbs-1 circuit switched voice -118.9 -111.9

64 kbs-1 circuit switched data -114.2 -107.0

128 kbs-1 circuit switched data -112.1 -105.4

128 kbs-1 packet switched data -112.4 -105.4

384 kbs-1 packet switched data -107.8 -100.9

Table 40 Received power level variation with subscriber service type

Note: The uplink and downlink input power levels in the table are the minimum receiver operating input power and do not include:

• Any additional power margin to allow for building attenuation or slow (log-normal) fading

• Any reduction in power to allow for soft handover gain in the forward link, and net gain in the receiver antenna system

This minimum acceptable received power level can be measured using the survey equipment described in the next section, in order to verify that the designed network coverage has been achieved, and to gather data for optimisation purposes.




The fact that the higher data rate services require a greater received power level can be used when planning coverage and base station sites, in order to minimise the number of required sites. This can be done in the following ways:

• Minimise the data rate

- Many applications such as internet access and file down loading from office based servers can use asymmetric data rates, without noticeably degrading their ease of use for the user.

The downlink data rate may be higher than the uplink data rate.

• Network performance will normally exceed the worst case situation, which is often evaluated during the RF network planning process

- This involves ensuring that the required service type and availability is achieved at the edge of the cell (and indoors) during the busy hour. Owing to their geographic distribution, the majority of subscribers will receive a better quality of service and higher data rate. For the majority of time, all users will receive a better quality of service and higher data rate.

It may be possible to reduce the number of base station sites by accepting a lower quality of service for a minority of time, and for a minority of subscribers.

7.9. Survey test equipment

A number of survey measurement equipment suppliers are currently developing equipment for UTRANs. One supplier expects the first release to be ready later this year. This will be based on:

• PC running Microsoft Windows 95 (or NT4)

• Scanning receivers

- Up to 4 receivers may be connected to the same PC for recording data from different networks, which may use different air interface standards and operate in different frequency bands

• Test phones

- Subject to availability, test phones will be introduced in the second software release (planned for autumn 2000)

Using an integrated solution comprising both a scanning receiver and a test phone provides a number of benefits:

• The test phone can identify the symptoms of the problem (such as call dropped or blocked, or poor FER)



221

• The scanner can help to identify the cause (such as low Ec, high Io, pilot pollution, timing problem, missing neighbour cell)

The test equipment is based on that currently used for American CDMA networks and records the following parameters:

• Test phone

- FER

- RSSI

- Dropped calls

- Blocked calls

- Messages

• Scanner

- [Ec/I0]

- Ec value

- I0 value

- Delay

- Aggregate [Ec/I0]

- Delay spread

- Channel power

- Spectrum for interference scan

• Data recording software

- Data is recorded to Microsoft Access database

- Measurement location data is recorded from Navstar GPS receiver

- Data export filter for Mapinfo

- Graphical user interface

- Simultaneous recording from up to 4 test phones and up to 4 scanning receivers, using different air interface standards and operating frequencies




Further details regarding survey equipment and techniques that may be used for UTRAN coverage verification and optimisation are given in the UMTS RF Optimisation Guidelines published by the UMTS RF Systems Engineering Group. Please note that as this document contains commercially sensitive information, it is not generally available to our customers. If you need more details in this area, please contact your Lucent representative, who will arrange to provide the required extracts from the document.

7.10. Site selection and design optimisation

This section describes the RF network engineering aspects of site selection and design. It does not cover site engineering or deployment aspects.

RF network engineering requirements

To successfully plan an optimum RF overlay, RF network engineers require information from a site selection survey. The required information includes:

• Clutter classification

• Site co-ordinates

• Available antenna heights

• Feeder cable lengths

• Foreground obstructions (including antenna support structures)

• Assessment of other site users (as potential interferers)

• Interference sweep

• Ground height and gradient

Clutter classification

During a site selection survey, the clutter found between the antenna position and subscriber locations is recorded and photographed.

Clutter is classified in a number of groups, such as Urban, Suburban, Rural, Trees and Water. Each clutter classification has a particular bearing on the overall RF coverage prediction, as RF energy is attenuated in different ways depending on the type of clutter that it encounters.

Photographic evidence of clutter types is recorded from the antenna position by taking a series of photographs, starting from north, to create a 360-degree panoramic view. This allows the RF engineers to identify potential problems due to clutter when predicting the RF coverage from the site.



223

Site co-ordinates

The location of each proposed transmission site must be determined as accurately as possible for inclusion in the RF coverage prediction tool.

We recommend that positions are determined from accurate maps by scaling from the grid provided. If suitable maps are not available for the survey area, a Global Positioning System (GPS) receiver should be used instead.

GPS receivers

A GPS receiver can give a position anywhere on the earth’s surface, providing it has line-of-sight to four satellites for a 3-D position, and to three satellites for a 2-D position. It can display the location on a national grid system (such as Ordnance Survey Great Britain) or on a world grid system (such as WGS 84) in latitude and longitude.

Important: Errors may be encountered when using a GPS receiver and these must be accounted for when locating a site. Factors such as selective availability and multipath can introduce errors in excess of 100 metres. It is therefore essential to use methods such as position averaging or differential measurements to give the most accurate results possible.

Selective availability is a deliberate degradation to the accuracy of the Clear/Acquisition (C/A) code signal used by commercial receivers. This degradation is expected to be reduced progressively in different geographic areas over the next five years, thereby improving the accuracy of C/A code receivers.

Available antenna heights

When surveying an existing mast or tower it is necessary to provide all the available antenna height options, bearing in mind where existing antennas are installed. This allows RF engineers to evaluate various combinations when determining the optimum antenna height.

Where possible, antenna heights are measured directly. If this is not possible, a clinometer is used to measure an angle from the ground, at a known range from the base of the antenna support structure, to the antenna position from which the height can be calculated using basic trigonometry.

If the intention is to use a high-rise building such as an office block or block of flats, then the height of the building plus the height of the support structure on the roof is recorded.

Feeder cable lengths

The length of cable required to connect the antenna to the transmission equipment is measured and used in calculations of signal power loss. It is the responsibility of the RF engineer and not the site selection engineer to determine the feeder cable type. This is because calculations for attenuation, and hence its effect on cell range, are necessary.




In addition to feeder length, data regarding any local site restrictions on cable selection should be recorded. For example: minimum bend radius, size and weight limits imposed by the support structure, ducts and risers, and any site specific safety conditions.

Foreground obstructions

The site selection survey should note any foreground obstructions and document their position by means of a bearing and distance. It is also important to measure the height of the obstruction and take photographs.

In addition to foreground obstructions, any potential effect on the antenna radiation pattern by the antenna support structure itself should be considered.

Assessment of other site users

The RF engineer must have information about any other site users, as they may become interferers.

Information and photographs of any existing antennas will allow the RF engineer to check, where possible, that the frequencies and power outputs already in use will not cause interference with the proposed installation.

Interference sweep

In addition to the potential interference from the existing site users it is necessary to check for outside interference from remote sources.

A spectrum analyser equipped with a Low Noise Amplifier (LNA) can be used to monitor a frequency band for a given period of time to see if any interference appears. By monitoring the frequency band for a period of time there is a greater chance of picking up interference from an intermittent source.

Ground height and gradient

The RF coverage prediction tool works with digital terrain data. It is important that the site selection engineer verifies the actual ground height against that shown by the terrain model. Significant errors can occur in hilly or mountainous areas where the actual ground height may differ significantly from the model, causing an error in the antenna height used in the final coverage prediction.


Deployment - Worked Example

8


225

8. Deployment – Worked Example

This chapter illustrates the main steps and RF design considerations involved in a transition to UTRAN operation in a small part of an existing GSM network.

8.1. Scenario

In order to describe one of the more complicated scenarios, this example is based on an existing dual band GSM network, where a UTRAN overlay is required in urban areas.

This allows the example to illustrate:

• Use of the existing 900 and 1800MHz base station sites

• The likely achievable performance with multi-band antennas (based on preliminary performance data)

• The resulting difference in coverage that can be expected on the 900, 1800, and 2000MHz bands

Owing to limited space in this document, coverage of only one city is examined. However, a similar approach may be adopted to deal with similar scenarios elsewhere.




Coverage area

• City of Braga, northern Portugal

• Population of 150,000

• Urban area of the city is assumed to cover approximately 10sq km

Existing base station sites

The example network uses Lucent 900 and 1800MHz equipment.

To preserve customer confidentiality, details of the existing base station such as location, antenna type, height, and feeder are not included here. Subject to the approval of the customer concerned, more detailed information may be available on request. If you would like more details please contact your Lucent representative.

Our coverage analysis in this example is based on data from 10 existing base station sites that cover the city and its surrounding rural area.

The general city setting and approximate location of 8 of the base station sites is shown on the following map extract (the other two base station locations lie to the north of the map area):



227

Figure 44 Map of city and surrounding area [© 1989 Instituto Portugues De Cartografia E Cadasto]

Note: The area shown on the map broadly corresponds to the area marked ‘City and Surrounding Area’ in the predicted coverage maps later in this section.




The city centre is covered by four base station sites, (1, 3, 8, and 10) as shown in the following map extract:

Figure 45 Map of city central area [© 1989 Instituto Portugues De Cartografia E Cadasto]

Note: The area shown in this map broadly corresponds to the area marked ‘City’ in the predicted coverage maps later in this section.

The following sections step through the RF network design process.



229

8.2. Background

Although this worked example has been based wherever possible on a real network and real BTS configuration data, where data is not available, typical values have been used.

The traffic model used is based on that anticipated by a UK UTRAN licence applicant in the early years after deployment. The network is expected to cover only the major cities, and relatively few UTRAN compatible mobiles are anticipated in the subscriber base.

The base station antenna performance follows that of a prototype multi-band GSM900 / GSM1800 / UTRAN cross-polarised sector antenna. A similar approach has been adopted for the performance of the tri-band diplexer, where provisional data sheet values have been used. The manufacturers of the antenna and diplexer have both indicated that the ‘prototype’ and ‘provisional’ performance is likely to be representative of the commercial product.

Different propagation models were used for the GSM 900, GSM 1800, and UTRAN coverage maps, and consequently any comparisons should be treated as indicative of the trend, rather than directly comparable.

The received power maps for UTRAN coverage are based on the 50% pole capacity downlink power budgets detailed in Chapter 6 (Link power budget elements section). In practice, for the majority of subscriber services the duplex link will be uplink power limited by ~2dB.




8.3. Estimating coverage and capacity

The first stage in the RF network design process is to estimate the air interface and BTS capacity required to support the anticipated traffic load. The Lucent Network Dimensioning Software was used to do this.

The Network Dimensioning Software requires three models on which to operate:

• Coverage model

• Traffic model

• System model

Each model is described below. For brevity, only coverage in the city itself is considered, but the technique illustrated can be extended to larger areas of the network and various land-use categories.

Coverage model

The first model required is for the required network coverage and for the quality and type of service to be provided.

Factors that are taken into account include:

• Land area and type over which coverage is required

• Grade of service (blocking rate) to be offered

• Extent of indoor coverage needed and the type of building construction method used (for example, traditional stone or lightweight steel framed), and hence the building attenuation value to be used

• Probability within the nominal coverage area, over which coverage is actually required, and hence the shadow fade margin and soft handover gain to be used



231

For the purpose of this example the following model is used.

• Land area

- Coverage of Braga city 10 sq. km

- Land-use category Urban

• Grade of service is 2%

• Building attenuation

- Dense urban area 20dB

- Urban area 20dB

- Suburban area 15dB

- Rural area 10dB

• Proportion of the nominal coverage area over which coverage is actually required is 95%




Traffic model

The traffic load that the network expects its subscribers to offer is apportioned according to the switching method used (circuit or packet) and the data rate required. As the required network coverage area is known, the anticipated traffic density per traffic type can be calculated.

The anticipated traffic density for this example is:

Zone = Urban

Area /sq km 10

8kb/s voice traffic density /Erl/sq km 20

13kb/s voice traffic density /Erl/sq km 0

Uplink 64kb/s packet data traffic density /Mb/s/sq km 0.045

Downlink 64kb/s packet data traffic density /Mb/s/sq km 0.225

Uplink 144kb/s packet data traffic density /Mb/s/sq km 0


Uplink 384kb/s packet data traffic density /Mb/s/sq km 0


Uplink 16kb/s circuit data traffic density /Erl/sq km 0

Zone = Urban

Downlink 16kb/s circuit data traffic density /Erl/sq km 0





Uplink 144kb/s circuit data traffic density /Erl/sq km 0.145

Downlink 144kb/s circuit data traffic density /Erl/sq km 0.145





233

System model

The system model is for the network and site specific system parameters that are added to the general UTRAN system model.

For the purpose of this example, the following system parameters are used.

General parameters

Parameter Value Chip rate 3.84 Mb/s Frequency of operation 2000 MHz Maximum cell loading 50%

Link budget and BTS

Parameter Value Base station antenna height DU - 30m (not used)

U – 37m SU – 25m (not used) RU – 30m (not used)

Mobile station antenna height 1.5m Base station power +43 dBm Mobile station power +21 dBm Base station antenna gain 16dBi Base station feeder loss 3 dB downlink 0 dB uplink

(MHA used) Mobile station antenna gain 0 dBi Base station noise figure 3 dB Mobile station noise figure 9 dB Number of sectors Tri-Sectored - 3 Normalised sector overlap Tri-Sectored (90°) - 12° Shadow fade margin 11.6dB Soft handover gain 4.9dB




Path loss model

Parameter Value Model COST 231-Hata Correction factors Dense Urban +3 dB (not

used) Urban 0 dB Suburban –6 dB (not used) Rural –18 dB (not used)

Downlink orthogonality factor

0.4



235

Traffic and services

For this example, the signal bit energy to noise energy ratios (Eb/N0) for the Vehicular-A channel is used:

ID 1 2 3 4 5 6 7 8 9 10 11 12

Service Description

8k voice 13k voice

16k lcd data

32k lcd data

64k lcd data

128 lcd data

144k lcd data

384k lcd data

16k udd data

32k udd data

64k udd data

128 udd data

Service bit rate (bps)

8000 13000 16000 32000 64000 128000 14400 384000 16000 30400

Eb/Io uplink (dB)

5.3 6.1 3.8 3.1 3.8 3.1 3.1 4.3 3.8

Eb/Io downlink (dB)

7.9 8.5 6 5.2 6 5.2 5.5 5.9 5.3

ID 13 14

Service Description

144k udd data

384k udd data

Service bit rate (bps)

60800 243200

Eb/Io uplink (dB)

3 2.4

Eb/Io downlink (dB)

5.2 4.3




The activity factor (or duty cycle) for each channel type is given below:

ID 1 2 3 4 5 6 7 8 9 10 11 12 13 14

Activity factor 0.5 0.5 1 1 1 1 1 1 1 1 1 1 1 1

Service Type Circuit Circuit Circuit Circuit Circuit Circuit Circuit Circuit Packet Packet Packet Packet Packet Packet



237

Analysis results

The Lucent Network Dimensioning Software was used to process the coverage, traffic and system models, yielding the following results.

BTS sites

The estimated number of tri-sector base station sites (BTS), together with the cell radius and coverage area, and traffic capacity is shown below:

Figure 46 Dense Urban Area – number of BTS sites and traffic capacity

The results in Figure 46 show the following information:

• Number of BTS required

- This displays the number of base station sites required (each tri-sectored in this case, thus supporting three cells). It also displays maximum cell radius and corresponding cell area relating to one of the tri-sector cells

• The cell loading (uplink) is limited to a ceiling of 50% of the pole (maximum) capacity by the user entering this as a system parameter

- The actual cell loading for the given traffic volume is shown in the results




- We recommend that the uplink cell loading is limited to between 50 and 75% of pole capacity. The downlink is not constrained in the same manner, and can be higher if the link budget permits

• The traffic supported by the whole tri-sectored base station site, comprising the three cells each operating on three RF carriers in this case, is also shown in Figure 46

- In this example the user has chosen adaptive cell loading, therefore the total supported traffic is equal to that dictated by the anticipated traffic density and the coverage area of each cell



239

Uplink power budget

The corresponding uplink power budget for the voice traffic is shown below:

Figure 47 Dense Urban Area – uplink power budget for voice traffic




Operating point

The operating point on the cell range against capacity graph is shown below:

Figure 48 Dense Urban Area – operating point on cell range against capacity graph

Conclusion

The dimensioning software results indicate that the offered traffic, spread uniformly over 10 sq. km, can be served by four tri-sectored base station sites. Each cell will have a radius of ~1.3km, and an associated coverage area of ~3.2sq. km.

As the existing network has four sites covering the city, it can be assumed that the desired UTRAN coverage can be achieved simply by installing UTRAN BTSs at the existing BTS sites.

However, the dimensioning software takes no account of the following factors:

• Local terrain features

• Ground clutter

• Base station location

In particular, although the dimensioning software estimates that only 4 BTS sites are needed, this assumes optimum site location, which will not be the case in reality. The next stage in the



241

RF network design process is to use the Airpro coverage prediction software to investigate the effect of these factors.

8.4. Coverage prediction with Airpro (CE 5)

The existing network provides coverage of the city and surrounding area from 8 BTS sites, and the city area itself from 4 BTS sites. These sites are used to provide both a 900MHz and 1800MHz service.

For economic reasons, the existing GSM sites in the network are initially to be reused to provide a UTRAN service, until UTRAN traffic growth justifies additional sites.

This section analyses the downlink coverage that is likely to be achieved from the existing sites, and compares it with the existing 900 and 1800MHz coverage. A similar technique can be used to analyse uplink coverage.

Existing 900MHz coverage

The base station location, output power, and antenna system details were obtained from the existing GSM 900 network site database. This information, together with terrain data, was used with Lucent’s coverage prediction software to generate the following predicted coverage maps:

• City and surrounding area received power bands

• City area received power bands

The maps show that the outdoor received power level is generally greater than –75dBm over the whole city. An outdoor received power level of –75dBm gives a margin of 27dB over the GSM 900 mobile reference sensitivity. In an urban area this can be apportioned as:

• Indoor penetration loss 18dB

• Shadow fade margin 9dB (approximates to a coverage probability of 95% with a standard deviation of 8dB)




Figure 49 City and surrounding area – outdoor 900MHz received power level



243

Figure 50 City area – outdoor 900MHz received power level




Existing 1800MHz coverage

A similar exercise was conducted to predict the existing coverage on the 1800MHz band. The network uses the same base station sites to provide coverage in both bands.

Again, two outdoor coverage maps have been produced:



The maps show that the outdoor received power level is generally greater than –75dBm over the whole city.

Although an outdoor received power level of –75dBm gives a margin of 25dB over the GSM 1800 mobile reference sensitivity, it is likely that the majority of mobiles will achieve similar performance at 1800 as at 900.



245

Figure 51 City and surrounding area – outdoor 1800MHz received power level




Figure 52 City area – outdoor 1800MHz received power level



247

Predicted UTRAN coverage

A similar exercise was conducted to predict the coverage that could be provided if UTRAN BTS equipment was installed at the existing base station sites which are used for both 900 MHz and 1800MHz GSM service.

The predictions are based on the following additional assumptions:

• Single duplex antenna (cross-polarised) used to provide GSM 900, GSM 1800, and UTRAN service

- Characteristics of a prototype multi-band antenna used

• Single duplex antenna feeder for GSM 900, GSM 1800, and UTRAN services

- Tri-band diplexer losses added

- Additional feeder losses incurred at the higher operating frequency added

• For each of the subscriber services link power budget as per section 6.3 Air interface link power budget

• Uplink capacity limited to 50% pole capacity

Again, two outdoor coverage maps have been produced:



The maps show that the outdoor received power level is generally greater than –105dBm over the whole city. A value of 11.6dB was entered for the shadow fade margin, representing a cell area coverage probability of 95%, with an associated standard deviation of 10dB.




Figure 53 City and surrounding area – outdoor UTRAN received power level



249

Figure 54 City area – outdoor UTRAN received power level

Unlike GSM 900 or 1800 networks, different received power levels are required for operation of voice and the various data services. The extent to which the different services can be provided from the existing base station sites is illustrated by maps indicating the area in which a higher received power level is available than the minimum required for each service. Coverage maps for the following services have been produced:

• 13kbs-1 voice circuits

• 64kbs-1 circuit switched data


• 128kbs-1 packet switched data





Figure 55 City and surrounding area – outdoor UTRAN 13kbs-1 voice circuit received power level –111.9dBm



251

Figure 56 City area – outdoor UTRAN 13kbs-1 voice circuit received power Level –111.9dBm




Figure 57 City and surrounding area – outdoor UTRAN 64kbs-1 circuit switched data received power level –107.0dBm



253

Figure 58 City area – outdoor UTRAN 64kbs-1 circuit switched data received power level –107.0dBm




Figure 59 City and surrounding area – outdoor UTRAN 128kbs-1 circuit switched data received power level –106.2dBm



255

Figure 60 City area – outdoor UTRAN 128kbs-1 circuit switched data received power level –106.2dBm




Figure 61 City and surrounding area – outdoor UTRAN 128kbs-1 packet switched data received power level –105.4dBm



257

Figure 62 City area – outdoor UTRAN 128kbs-1 packet switched data received power level –105.4dBm




Figure 63 City and surrounding area – outdoor UTRAN 384kbs-1 packet switched data received power level –100.9dBm



259

Figure 64 City area – outdoor UTRAN 384kbs-1 packet switched data received power level –100.9dBm




Results

The maps indicate that the existing base station sites may be suitable to provide an outdoor UTRAN speech service throughout the city and surrounding area. However the higher data rate services may only be provided in a more restricted area.


Available throughout the city, surrounding towns and main villages, and along the major roads linking the city with neighbouring settlements.


Similar coverage as for 64kbs-1 circuit switched data, as the link power budget is only ~1dB worse. This introduces slight discontinuity in coverage along some of the main roads from the city to the surrounding towns and villages. Coverage within the main surrounding towns and villages is maintained.


Contiguous coverage in the city.

Coverage becomes a little patchy in the following areas:

- Surrounding towns of Nogueiro, Tenoes, and Guaitar (to the north east of the city) and Balazar (to the south east)

- City suburbs of Nogueire and Lomar

- Coverage is very patchy in the roads linking these settlements, and in the suburb of Ferreiros to the south-west of the city.


Coverage in the city has a few holes towards the south and south-east.

Approximately 50% of the surrounding area of the city is covered.

Coverage is no longer contiguous between the city and neighbouring settlements.

Coverage is very patchy in the south-eastern city suburbs such as Ferreiros and Lomar.

Conclusion

The purpose of this example is to illustrate the RF coverage design process, and show how the Lucent software can be used to aid this task.

It has demonstrated that the various subscriber services available from a UTRAN have different coverage areas, and that generally, coverage falls as the subscriber data rate increases.



261

Having established that an adequate UTRAN service can be provided within the city from the existing base station sites, the next step is to examine the EMC problems associated with sharing the same antennas.

8.5. EMC when using a common antenna

This section calculates the in-band and out-of-band power levels presented to each of the receivers involved.

900MHz receiver - in-band power

From 1800MHz transmitter

• 1800MHz transmitter spurii <-98dBm (in 100kHz BW)

• Antenna inter-band isolation >30dB

• Resulting spurii power <-128dBm

This level is below the minimum sensitivity and noise floor of the GSM 900 receiver (-108.5dBm BTS2000 static) and will not affect operation of the GSM 900 receiver.

From UTRAN transmitter

• UTRAN transmitter spurii <-98dBm (in 100kHz BW)



This level is below the minimum sensitivity and noise floor of the GSM 900 receiver (-108.5dBm BTS2000 static) and will not affect operation of the GSM 900 receiver.




1800MHz receiver - in-band power





This level is below the minimum sensitivity and noise floor of the of the GSM 1800 receiver (-109.5dBm BTS2000 static), and will not affect operation of the GSM 1800 receiver.





This level is below the minimum sensitivity and noise floor of the of the GSM 1800 receiver (-108.5dBm BTS2000 static), and will not affect operation of the GSM 1800 receiver.

UTRAN receiver - in-band power



• Three GSM 900 transmitters (using one antenna) correction factor of 6dB



This level is approximately that of the minimum sensitivity value of the UTRAN receiver (-110 to –121dBm depending on the type of service), and therefore will have little effect on the operation of the UTRAN receiver.







263


This level is approximately that of the minimum value of sensitivity of the of the UTRAN receiver (-110 to –121dBm depending on the type of service), and therefore will have little effect on the operation of the UTRAN receiver.

GSM 900MHz receiver – out-of-band power


• 1800MHz transmitter power at antenna input port +39dBm (in 200kHz BW)

• Tri-band diplexer isolation >60dB


• 900MHz BTS duplexer isolation >30dB

• 900MHz RF pre-selector filter >30dB

• Resulting received power <-111dBm

This resulting received power level is far below that where intermodulation, desensitisation, and blocking may occur.


• UTRAN transmitter power at antenna input port +40dBm (in 5MHz BW)

• Bandwidth correction factor (5MHz to 200kHz) -14dB










GSM 1800MHz receiver – out-of-band power










• UTRAN transmitter power at antenna input port +40dBm (in 5MHz BW)

• Bandwidth correction factor (5MHz to 200kHz) -14dB









265

UTRAN receiver – out-of-band power






• UTRAN BTS duplexer isolation >30dB

• UTRAN RF pre-selector filter >30dB








• UTRAN BTS duplexer isolation >30dB

• UTRAN RF pre-selector filter >30dB






Conclusion

The EMC assessment indicates that the three types of BTS may be able to share a common tri-band antenna without any undue impairment to receiver performance.

However, when considering single antenna operation, the following can be considered if additional EMC improvement measures are required:

• In-band interference (such as spurii) generated from an out-of-band BTS can be reduced by adding in-band filtering in the antenna feeder line of the out-of-band BTS

• Out-of-band interference (for example, signals causing intermodulation, desensitisation, and blocking), generated from an out-of-band BTS can be reduced by adding out-of-band filtering in the antenna feeder line of the in-band BTS

However, unnecessary filtering should not be used in order to avoid reduction in output power, reduction in received signal power, and degradation in receiver system noise figure.


Appendix A Silence Duration Parameters


267

Appendix A Silence Duration Parameters

This appendix reproduces an extract from ETSI UMTS xx.15 v1.0.0 1999-02 regarding the Definition and Setting of Silence Duration Parameters.

Definition of silence duration parameters

SDc is defined as the following:

SDc = 2 * tsynth. + (c+1) * TSGSM

where, tsynth. : tsynth. is the maximum allowed delay for a mobile’s synthesizer to switch from one FDD to one GSM frequency. Typically this value could be derived, from the larger frequency difference switch (= 1245 MHz) which is the one from upper FDD downlink frequency (2170 MHz) to lower GSM downlink frequency (925 MHz). The value of tsynth. is set to [500 ms]. c: integer value {0,1,2,3,4,5,6,7,8} TSGSM: GSM time slot duration. 577 ms SDc: silence duration number c. SDC is the necessary time for a dual mode mobile to switch to one GSM frequency, decode c consecutive GSM slots and switch back to the current FDD carrier. In the particular case where the scheduling of silence duration SDc is synchronised with the scheduling of GSM time slots, the UE can decode (c+1) consecutive GSM time slots. SD pattern: a set of consecutive silence durations enabling the capture of at least one time slot 0 of one GSM BCCH carrier. TSDc: delay between two consecutive silence durations within a SD pattern when GSM time slot number i mod(8) is the first time slot to be captured by the first silence duration and GSM time slot number ( i+c) mod(8) is the first time slot to be captured by the second silence duration.




Tpattern: delay between two consecutive patterns, optionally generalised to Tpattern1 and Tpattern2 to be used alternating between consecutive patterns.

Npattern: number of consecutive patterns to be used for scanning a particular GSM frequency before this search attempt is assumed to be unsuccessful.

Figure 65 Illustrations of SD patterns, silence duration SDc ,Tpattern1 ,Tpattern2, TSDc and Npattern. [© ETSI 1998]

Table 41 Illustration of possible SD combinations [© ETSI 1998]


Acronyms


269

Acronyms

The following acronyms are used in this document.

3GPP 3rd Generation Partnership Project

ACELP Algebraic Code Excited Linear Prediction Codec

ACS Active Codec Set

AFS AMR Full Rate

AHS AMR Half Rate

AMPS Advanced Mobile Phone Service

AMR Adaptive Multi Rate

AP Application Processors

ASIC Application Specific Integrated Circuit

BA BCCH Allocation (list)

BBU Baseband Unit

BCH Broadcast Channel

BCCH Broadcast Control Channel

BCCH-C BCCH-Constant

BCCH-V BCCH-Variable

BCS Block Check Sequence

BER Bit Error Rate

BHCA Busy Hour Call Attempts




BLER Block Error Rate

BPSK Binary Phase Shift Keying

BSC Base Station Controller

BSIC Base Station Identity Code

BSS Base Station Subsystem

BTS Base Transceiver Station

C/A Clear/Acquisition

CCCH Common Control Channel

CCPCH Common Control Physical Channel

CCTrCH Coded Composite Transport Channel

CCU Channel Codec Unit

CIR Carrier to Interference Ratio

CPCH Common Packet Channel

CPICH Common Pilot Channel

CRC Cyclic Redundancy Check

CS Circuit Switched

CTCH Common Traffic Channel

DCH Dedicated Channel

DCCH Dedicated Control Channel

DCPH Dedicated Physical Channel

DECT Digital European/Enhanced Cordless Telephone/ Telecommunications

DPDCH Dedicated Physical Data Channel

DPCCH Dedicated Physical Control Channel

DRNS Drift Radio Network Subsystem

DRX Discontinuous Reception

DSCH Downlink Shared Channel

DSP Digital Signal Processing

DTCH Dedicated Traffic Channel

DTX Discontinuous Transmission

EDGE Enhanced Data Rates for GSM Evolution

EFR Enhanced Full Rate

EIRP Effective Isotropic Radiated Power



271

ETACS Extended Total Access Communications System

ETSI European Telecommunications Standards Institute

FACH Forward Access Channel

FAUSCH Fast Uplink Signalling Channel

FBI Feedback Information

FCC Federal Communications Commission (USA)

FEC Forward Error Correction

FDD Frequency Division Duplex

FR Full Rate

GPS Global Positioning System

GMSK Gaussian Minimum Shift Keying

GoS Grade of Service

GPRS General Packet Radio Service

GSM Global System for Mobile Communications

HCS Hierarchical Cell Structure

HR Half Rate

IMT-2000 International Mobile Telecommunications 2000

ICC Intelligent Carrier Card

IF Intermediate Frequency

IP Internet Protocol

IR Incremental Redundancy

ITU-R International Telecommunications Union - Radio

LA Link Adaptation

LMTT Local Maintenance Test Terminal

LNA Low Noise Amplifier

LO Local Oscillator

LRU Line Replaceable Unit

MAC Medium Access Control

MACS Codec Modes

MDS Minimum Discernible Signal

MHA Masthead Low Noise Amplifier

MOS Mean Opinion Score

MS Mobile Station




MSC Mobile Switching Centre

MTBF Mean Time Between Failures

MTTR Mean Time To Repair

NMT Nordic Mobile Telephone

NSS Network Switching Subsystem

OCCCH ODMA Common Control Channel

ODCH ODMA Dedicated Channel

ODCCH ODMA Dedicated Control Channel

ODMA Opportunity Driven Multiple Access

ODTCH ODMA Dedicated Traffic Channel

OMC Operations and Maintenance Centre

ORACH ONMA Random Access Channel

OVSF Orthogonal Variable Spreading Factor

PACCH Packet Dedicated Control Channel

PAGCH Packet Access Grant Channel

PBCCH Packet Broadcast Control Channel

PBX Private Branch Exchanges

PCCH Paging Control Channel

PCCCH Packet Common Control Channel

PCH Paging Channel

PCM Pulse Code Modulation

PCS Personal Communication System

PCU Packet Control Unit

PDCH Packet Data Channel

PDTCH Packet Data Traffic Channel

PHS Personal Handy-phone System

PLMN Public Land Mobile Network

PPCH Packet Paging Channel

PRACH Physical Random Access Channel

PS Packet Switched

PSCH Primary Synchronisation Channel



273

PTM-M Point to Multipoint – Multicast

QoS Quality of Service

QPSK Quadrature Phase Shift Keying

RAB Radio Access Bearer

RACH Random Access Channel

RLC Radio Link Control

RNC Radio Network Controller

RNS Radio Network System

RPE-LTP Regular Pulse Excited Long Term Prediction

RRC Radio Resource Control

SACCH Slow Associated Control Channel

SCCH Synchronisation Control Channel

SCH Synchronisation Channel

SCR Source Controlled Rate

SF Spreading Factor

SGSN Serving GPRS Support Node

SHCCH Shared Channel Control Channels

SHO Soft Handover

SID Silence Information Descriptor

SIR Signal to Interference Ratio

SMS-CB Short Message Service Cell Broadcast Centre

SRNS Serving Radio Network Subsystem

SSCH Secondary Synchronisation Channel

STF Speech Transcoding Frame

STM-n Synchronous Transfer Module

STTD Space Time Block Coding Based Diversity

TB Tail Bits

TBF Temporary Block Flow

TDD Time Division Duplex

TDMA Time Division Multiple Access

TF Temporary Flow Indicator

TFCI Transport Format Combination Indicator

TFO Tandem Free Operation




TrFO Transcoder Free Operation

TPC Transmit Power Control

TRAU Transceiver/Rate Adaptor Unit

TRX Transceiver

TS Time slot

TSTD Time Switched Transmit Diversity

UMTS Universal Mobile Telecommunications System

USCH Uplink Shared Channel

UTRAN Universal Terrestrial Radio Access Network

WCDMA Wideband Code Division Multiple Access

WLL Wireless Local Loop

WRC World Radio Conference

Comments Form

Lucent Technologies – Proprietary See Notice on first page

Comments Form

Document number: 401-380-373 Issue number: 1.1 Date: July 2000

Title: GSM to UMTS Transition RF Engineering Guideline

Lucent Technologies welcomes feedback on this customer Information Product. Your comments are of great value in assisting us to improve our Information Products.

1. Please rate (tick the box) the effectiveness of this Information Product in the following areas:

Excellent Good Fair Poor Not Applicable Ease of use Clarity Completeness Accuracy Organisation Appearance Examples Illustrations Overall Satisfaction

2. Please place a tick against any improvement that could be made to this Information Product:

Improve the overview/introduction Make it more concise/brief Improve the table of contents Add more step-by-step procedures/tutorials Improve the organisation Add more troubleshooting information Include more figures Make it less technical Add more examples Add more detail

3. Please provide details for the suggested improvement in the box below:

Comments Form

Lucent Technologies – Proprietary See Notice on first page

4. What did you like most about this Information Product:

5. Any additional comments:

6. If we may contact you concerning your comments, please complete the following:

Date:

Name:

Company/Organisation:

Address:

Job Title:

Telephone No:

When you have completed this form, please fax it to:

The Technical Manager, CTIP-UK, Fax Number (+44) 1666 824515

Download - Transition from GSM to UMTS

Top Related