Radio Cell planning principles

CELL PLANNING PRINCIPLES

STUDENT TEXTEN/LZT 123 3314

R3A

Cell Planning Principles

',6&/$,0(5

This book is a training document and contains simplifications.Therefore, it must not be considered as a specification of thesystem.

The contents of this document are subject to revision withoutnotice due to ongoing progress in methodology, design andmanufacturing.

Ericsson assumes no legal responsibility for any error or damageresulting from the usage of this document.

This document is not intended to replace the technicaldocumentation that was shipped with your system. Always referto that technical documentation during operation andmaintenance.

This document was produced by Ericsson Radio Systems AB.

• It is used for training purposes only and may not be copiedor reproduced in any manner without the express writtenconsent of Ericsson.

• This document number, EN/LZT 123 3314, R3A supportscourse number LZU 108 3273.

EN/LZT 123 3314 R3A

&RS\ULJKW��E\�(ULFVVRQ�5DGLR�6\VWHPV�$%

Revision Record

5(9,6,21�5(&25'

Date Revision No. ChaptersAffected

EN/LZT 123 3314 R3A

96/09/02 R1A All

96/11/04 R1B 1-3, 6-9, 11, 13

97/03/31 R2A All

97/03/31 R2A All

97/06/05 R2B All

98/06/15 R3A All

Table of Contents

&HOO�3ODQQLQJ�3ULQFLSOHV

Table of Contents

Topic Page

EN/LZT 123 3314 R3A -i -

1. Cell Planning Introduction................................................................ 1

2. System Description......................................................................... 7

3. Radio Wave Propagation.............................................................. 21

4. Traffic ........................................................................................... 41

5. System Balance............................................................................ 47

6. Coverage Predictions ................................................................... 51

7. Channel Planning ......................................................................... 61

8. Surveys......................................................................................... 69

9. Design Projects ............................................................................ 77

10. Implementation ............................................................................. 87

11. System Tuning............................................................................ 127

12. System Growth ........................................................................... 149

13. Radio Network Features ............................................................. 155

14. Index............................................................................................ 257

Cell Planning Introduction

Chapter 1

This chapter is designed to provide the student with anintroduction to cell planning.

OBJECTIVES:Upon completion of this chapter, the student will be able to:

• Explain briefly the major steps in cell planning

• Describe what support Ericsson offers regarding cellplanning services

1 Cell Planning Introduction

EN/LZT 123 3313 R3A – i –


Table of Contents

Topic Page

INTRODUCTION....................................................................................1

CELL PLANNING PROCESS................................................................1

STEP 1: THE CELLPLANNING PROCESSTRAFFIC AND COVERAGEANALYSIS (SYSTEM REQUIREMENTS) ..................................................................... 3

STEP 2: NOMINAL CELL PLAN.................................................................................... 3

STEP 3: SURVEYS (AND RADIO MEASUREMENTS)................................................. 4

STEP 4: (FINAL CELL PLAN) SYSTEM DESIGN......................................................... 4

STEP 5: IMPLEMENTATION......................................................................................... 4

STEP 6: SYSTEM TUNING ........................................................................................... 4

ERICSSON SUPPORT FOR CELL PLANNING....................................5


EN/LZT 123 3313 R3A – 1 –

INTRODUCTION

This course, Cell Planning Principles, is intended to give thestudent an understanding of the Radio Frequency (RF)engineering processes and what elements they contain.

The course is broken down into chapters that explain thedifferent elements of the process.


– 2 – EN/LZT 123 3313 R3A

CELL PLANNING PROCESS

Cell planning can be described briefly as all the activitiesinvolved in determining which sites will be used for the radioequipment, which equipment will be used, and how theequipment will be configured.

In order to ensure coverage and to avoid interference, everycellular network needs planning. The major activities involvedin the cell planning process are depicted in Figure 1-1.

Traffic & Coverageanalysis

Nominal cell plan

Surveys

System design

Implementation

System tuning

System Growth Initial Planning

Sites

Cell Plan

FQ Plan

Traffic CoverageQuality.

..

Traffic Data

Cell design

data

Cov. map

Site conf.

Figure 1-1 The cell planning process


EN/LZT 123 3313 R3A – 3 –

STEP 1: THE CELLPLANNING PROCESS TRAFFIC AND COVERAGEANALYSIS (SYSTEM REQUIREMENTS)

The cell planning process starts with traffic and coverageanalysis. The analysis should produce information about thegeographical area and the expected need of capacity. The typesof data collected are:

• Cost

• Capacity

• Coverage

• Grade of Service (GoS)

• Available frequencies

• Speech Quality Index

• System growth capability

The traffic demand (i.e. how many subscribers will join thesystem and how much traffic will be generated) provides thebasis for cellular network engineering. Geographical distributionof traffic demand can be calculated by using demographic datasuch as:

• Population distribution

• Car usage distribution

• Income level distribution

• Land usage data

• Telephone usage statistics

• Other factors such as subscription charges, call charges, andprice of mobile stations

STEP 2: NOMINAL CELL PLAN

Upon compilation of the data received from the traffic andcoverage analysis, a nominal cell plan is produced. The nominalcell plan is a graphical representation of the network and simplylooks like a cell pattern on a map. However, a lot of work liesbehind it (as described previously).

Nominal cell plans are the first cell plans produced and form thebasis for further planning. Quite often a nominal cell plan,together with one or two examples of coverage predictions, isincluded in tenders.


– 4 – EN/LZT 123 3313 R3A

At this stage, coverage and interference predictions are usuallystarted. Such planning needs computer-aided analysis tools forradio propagation studies, e.g. Ericsson’s planning tool known asthe Ericsson Engineering Tool (EET).

STEP 3: SURVEYS (AND RADIO MEASUREMENTS)

The nominal cell plan has been produced and the coverage andinterference predictions have been roughly verified. Now it istime to visit the sites where the radio equipment will be placedand perform radio measurements. The former is importantbecause it is necessary to assess the real environment todetermine whether it is a suitable site location when planning acellular network. The latter is very important because evenbetter predictions can be obtained by using field measurementsof the signal strengths in the actual terrain where the mobilestation will be located.

STEP 4: (FINAL CELL PLAN) SYSTEM DESIGN

Once we optimize and can trust the predictions generated by theplanning tool, the dimensioning of the RBS equipment, BSC,and MSC is performed. The final cell plan is then produced. Asthe name implies, this plan is later used during systeminstallation. In addition, a document called Cell Design Data(CDD) containing all cell parameters for each cell is completed.

STEP 5: IMPLEMENTATION

System installation, commissioning, and testing are performedfollowing final cell planning and system design.

STEP 6: SYSTEM TUNING

After the system has been installed, it is continually evaluated todetermine how well it meets the demand. This is called systemtuning. It involves:

• Checking that the final cell plan was implementedsuccessfully

• Evaluating customer complaints

• Checking that the network performance is acceptable

• Changing parameters and performing other measures (ifneeded)


EN/LZT 123 3313 R3A – 5 –

The system needs constant retuning because the traffic andnumber of subscribers increases continuously. Eventually, thesystem reaches a point where it must be expanded so that it canmanage the increasing load and new traffic. At this point, acoverage analysis is performed and the cell planning processcycle begins again.


– 6 – EN/LZT 123 3313 R3A

ERICSSON SUPPORT FOR CELL PLANNING

Ericsson offers the Way Forward as a design for cell planning.The Way Forward is a solution concept that combines a numberof techniques, features, and service products. Together theyprovide substantial capacity gain in GSM mobile telephonenetworks without the need for additional radio frequencyspectrum.

The Way Forward method is described in more detail in Chapter12 “System Growth”.

System Description

Chapter 2

This chapter is designed to provide the student with an overviewof the GSM system.


• Explain the basic function of a GSM system

• Describe the network nodes of a GSM system

• Describe general terms used in the GSM system

• Describe the geographical network structure

2 System Description

EN/LZT 123 3314 R3A – i –


Table of Contents

Topic Page

GLOBAL SYSTEM FOR MOBILE COMMUNICATIONS (GSM) ...........7

THE DIFFERENT GSM-BASED NETWORKS .............................................................. 7

NETWORK HARDWARE.......................................................................9

OPERATION AND SUPPORT SYSTEM (OSS) ..................................10

SWITCHING SYSTEM (SS) .................................................................11

BASE STATION SYSTEM (BSS) ........................................................13

BSC.............................................................................................................................. 13

BTS .............................................................................................................................. 14

RBS 200....................................................................................................................... 14

RBS 2000..................................................................................................................... 14

AIR INTERFACE..................................................................................15

FREQUENCY ALLOCATION....................................................................................... 15

CHANNEL CONCEPT ................................................................................................. 15

LOGICAL CHANNELS................................................................................................. 15


EN/LZT 123 3314 R3A – 7 –

GLOBAL SYSTEM FOR MOBILE COMMUNICATIONS (GSM)

In 1982, the Nordic PTT sent a proposal to ConférenceEuropéenne des Postes et Télécommunications (CEPT) tospecify a common European telecommunication service at 900MHz. A Global System for Mobile Communications (GSM)standardization group was established to formulate thespecifications for this pan-European mobile cellular radiosystem.

During 1982 through 1985, discussions centered around whetherto build an analog or a digital system. Then in 1985, GSMdecided to develop a digital system.

In 1986, companies participated in a field test in Paris todetermine whether a narrowband or broadband solution wouldbe employed. By May 1987, the narrowband Time DivisionMultiple Access (TDMA) solution was chosen.

Concurrently, operators in 13 countries (two operators in theUnited Kingdom) signed the Memorandum of Understanding(MoU) which committed them to fulfilling GSM specificationsand delivering a GSM system by July 1, 1991. This opened alarge new market.

The next step in the GSM evolution was the specification ofPersonal Communication Network (PCN) for the 1800 MHzfrequency range, Digital Cellular System (DCS) 1800 (or GSM1800), and Personal Communication Services (PCS) 1900 (orGSM 1900) for the 1900 MHz frequency range.

THE DIFFERENT GSM-BASED NETWORKS

Different frequency bands are used for GSM 900/1800 andGSM 1900 (Figure 2-1). In some countries, an operator appliesfor the available frequencies. In other countries (e.g. UnitedStates), an operator purchases available frequency bands atauctions.


– 8 – EN/LZT 123 3314 R3A

Network type

GSM 900 890 - 915/935 - 960 MHz GSM 900/1800

GSM 1800 1710 - 1785/1805 -1880 MHz

GSM 1900 1850 - 1910/1930 -1990 MHz GSM 1900

Ericsson’simplementations

Frequency bandUL/DL

Figure 2-1 Frequency bands for the different GSM-basednetworks


EN/LZT 123 3314 R3A – 9 –

NETWORK HARDWARE

Every cellular system has hardware that is specific to it and eachpiece of hardware has a specific function. The Ericsson GSM-based systems comply with the GSM standard while varyingfrom it for the purpose of overall system improvement.

The system solutions integrate existing Ericsson hardware andnew technology to provide a "total" solution for the mobiletelephony market. The major systems in the network are:

• Operation and Support System

• Switching System

• Base Station System

The system is normally configured as depicted in Figure 2-2.

BSSBSC

BTS

MS

OSS

Switching System

Information transmission

Call connections and information transmission

Base Station System

MXE

MIN

SOG

BGW

SS

MSC/VLR

EIR

ILR

GMSC

DTI

AUC

HLR

SMS-GMSCSMS-IWMSC

Figure 2-2 Ericsson GSM-based system model


– 10 – EN/LZT 123 3314 R3A

OPERATION AND SUPPORT SYSTEM (OSS)

For GSM system administration, the OSS supports the networkoperator by providing:

• Cellular network administration

• Network operation and maintenance


EN/LZT 123 3314 R3A – 11 –

SWITCHING SYSTEM (SS)

SS

MSC/VLR

EIR

ILR

GMSC

DTI

AUC

HLR

SMS-GMSCSMS-IWMSC

Figure 2-3 Switching System

• 0RELOH�VHUYLFHV�6ZLWFKLQJ�&HQWHU��06&�

The MSC is responsible for set-up, routing, and supervisionof calls to and from mobile subscribers. Other functions arealso implemented in the MSC, such as authentication. TheMSC is built on an AXE-10 platform.

• 9LVLWRU�/RFDWLRQ�5HJLVWHU��9/5�

In the Ericsson GSM based solution, the VLR is integratedwith the MSC. This is referred to as the MSC/VLR. TheVLR contains non-permanent information about the mobilesubscribers visiting the MSC/VLR service area, e.g., whichlocation area the MS is currently in.

• *DWHZD\�06&��*06&�

The Gateway MSC (GMSC) supports the function forrouting incoming calls to the MSC where the mobilesubscriber is currently registered. It is normally integrated inthe same nodes as MSC/VLR.

• +RPH�/RFDWLRQ�5HJLVWHU��+/5�

In GSM, each operator has a database containinginformation about all subscribers belonging to the specificPublic Land Mobile Network (PLMN). This database can beimplemented in one or more HLRs. Two examples ofinformation stored in the database are the location(MSC/VLR service area) of the subscribers and servicesrequested. The HLR is built on an AXE-10 platform.


– 12 – EN/LZT 123 3314 R3A

• $8WKHQWLFDWLRQ�&HQWHU��$8&�

For security reasons, speech, data, and signaling areciphered, and the subscription is authenticated at access. TheAUC provides authentication and encryption parametersrequired for subscriber verification and to ensure callconfidentiality.

• (TXLSPHQW�,GHQWLW\�5HJLVWHU��(,5�

In GSM there is a distinction between subscription andmobile equipment. As mentioned above, the AUC checksthe subscription at access. The EIR checks the mobileequipment to prevent a stolen or non-type-approved MSfrom being used.

• ,QWHUZRUNLQJ�/RFDWLRQ�5HJLVWHU��,/5�

Around the world there are market demands for roamingcapabilities with GSM. The ILR is the node that forwardsroaming information between cellular networks usingdifferent operating standards. This currently exists only inthe GSM 1900 network.

• 6KRUW�0HVVDJH�6HUYLFH��*DWHZD\�06&��606�*06&�

A Short Message Service Gateway MSC (SMS-GMSC) iscapable of receiving a short message from a Service Center(SC), interrogating an HLR for routing information andmessage waiting data, and delivering the short message tothe MSC of the recipient MS. In Ericsson’s GSM system, theSMS-GMSC functionality is normally integrated in theMSC/VLR node.

• 6KRUW�0HVVDJH�6HUYLFH��,QWHU:RUNLQJ�06&��606�,:06&�

A Short Message Service InterWorking MSC (SMS-IWMSC) is capable of receiving a mobile originated shortmessage from the MSC or an ALERT message from theHLR and submitting the message to the recipient SC. TheSMS-IWMSC functionality is normally integrated in theMSC/VLR node.

• 'DWD�7UDQVPLVVLRQ�,QWHUIDFH��'7,�

DTI - consisting of both hardware and software - provides aninterface to various networks for data communication.Through DTI, users can alternate between speech and dataduring the same call. Its main functions include a modemand fax adapter pool and the ability to perform rateadaptation. It was earlier implemented as the GSMInterWorking Unit (GIWU).


EN/LZT 123 3314 R3A – 13 –

BASE STATION SYSTEM (BSS)

The Base Station System (BSS) is comprised of two majorcomponents. They are:

• Base Station Controller (BSC)

• Base Transceiver Station (BTS)

Base Station System

BSC

BTS

Figure 2-4 Base Station System

BSC

The Base Station Controller (BSC) is the central point of theBSS. The BSC can manage the entire radio network andperforms the following functions:

• Handling of the mobile station connection and handover

• Radio network management

• Transcoding and rate adaptation

• Traffic concentration

• Transmission management of the BTSs

• Remote control of the BTSs


– 14 – EN/LZT 123 3314 R3A

BTS

The Base Transceiver Station (BTS) includes all radio andtransmission interface equipment needed in one cell. TheEricsson name for the BTS is Radio Base Station (RBS). TheEricsson RBS corresponds to the equipment needed on one siterather than one cell. Each BTS operates at one or several pairs offrequencies. One frequency is used to transmit signals to themobile station and one to receive signals from the mobilestation. For this reason at least one transmitter and one receiveris needed.

RBS 200

The RBS 200 Base Station family was the first Ericsson GSMbase station developed in the early 1990’s. It exists only in theGSM 900/1800 product line. The RBS 200/203/204 is the GSM900 BTS, and the RBS 205 is the BTS supporting GSM 1800.

In general, the 200 family supports 4 transceiver modules(TRXs) per cabinet.

RBS 2000

The RBS 2000 Base Station family is based on the RBS 200 andcan be used for GSM 900/1800 and GSM 1900. There are fourdifferent models in the series:

• RBS 2101 with 2 Transceiver Units (TRUs)

• RBS 2102 and 2202 with 6 TRUs

• RBS 2103 (GSM 900 only) with 6 TRUs and smallerfootprint

• RBS 2301 is the micro-base station

All models are outdoor versions except RBS 2202.


EN/LZT 123 3314 R3A – 15 –

AIR INTERFACE

FREQUENCY ALLOCATION

Figure 2-1 (shown earlier) lists the band allocations for each ofthe different GSM based networks.

In many countries, the whole frequency band will not be usedfrom the outset.

CHANNEL CONCEPT

The carrier separation in GSM is 200 kHz. That yields 124carriers in the GSM 900 band. Since every carrier can be sharedby eight MSs, the number of channels is 124 times eight = 992channels. These are called physical channels. The correspondingnumber of carriers for GSM 800 and GSM 1900 are 374 and299, respectively.

LOGICAL CHANNELS

On every physical channel, a number of logical channels aremapped. Each logical channel is used for specific purposes, e.g.,paging, call set-up signaling or speech.

There are eleven logical channels in the GSM system. Two ofthem are used for traffic and nine for control signaling.

Traffic CHannels (TCH)

Two types of TCH are used:

• Full rate channel, Bm

This channel can be used for full rate or enhanced full ratespeech (13 kbit/s after speech coder) or data up to 9.6 kbit/s.

• Half rate channel, Lm

This channel can be used for half rate speech (6.5 kbit/s afterspeech coder) or data up to 4.8 kbit/s.


– 16 – EN/LZT 123 3314 R3A

Control channels

Nine different types of control channels are used.

Broadcast CHannels (BCH)

• Frequency Correction CHannel (FCCH)

Used for frequency correction of the MS, downlink only.

• Synchronization CHannel (SCH)

Carries information about TDMA frame number and BaseStation Identity Code (BSIC) of the BTS, downlink only.

• Broadcast Control CHannel (BCCH)

Broadcasts cell specific information to the MS, downlinkonly.

Common Control CHannels (CCCH)

• Paging CHannel (PCH)

Used to page the MS, downlink only.

• Random Access CHannel (RACH)

Used by the MS to request allocation of a Stand AloneDedicated Control Channel (SDCCH), either as a pageresponse or an access to MS call origination/registration,location updating, etc. uplink only.

• Access Grant CHannel (AGCH)

Used to allocate SDCCH to a MS, downlink only.

Dedicated Control CHannels (DCCH)

• Stand alone Dedicated Control CHannel (SDCCH)

Used for signaling during the call set-up or registration, up-and downlink.

• Slow Associated Control CHannel (SACCH)

Control channel associated with a TCH or a SDCCH, up-and downlink. On this channel the measurement reports aresent on the uplink, and timing advance and power orders onthe downlink.


EN/LZT 123 3314 R3A – 17 –

• Fast Associated Control CHannel (FACCH)

Control channel associated with a TCH, up- and downlink.FACCH works in bit-stealing mode, i.e. 20 ms of speech isreplaced by a control message. It is used during handoverwhen the SACCH signaling is not fast enough.

Several logical channels can share the same physical channel orTime Slot (TS). On TS0 (on one carrier per cell, the BCCH-carrier) the broadcast channels and the common control channelsare multiplexed.


– 18 – EN/LZT 123 3314 R3A

Carrier C0 Downlink Uplink

0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7Frame 0 F T D0 T T T T T R T A5 T T T T T

S T D0 R T A5B T D0 R T A5B T D0 R T A5B T D1 R T A6B T D1 R T A6C T D1 R T A6C T D1 R T A6C T D2 R T A7C T D2 R T A7F T D2 R T A7S T D2 R T A7

12 C A D3 I A I A I R A I I A I A I

C T D3 R T I

C T D3 R T I

C T D3 R T D0C T D4 R T D0C T D4 R T D0C T D4 R T D0C T D4 R T D1F T D5 R T D1S T D5 R T D1C T D5 R T D1C T D5 R T D2C T D6 R T D2

25 C I D6 A I A I A R I D2 A I A I A

C T D6 R T D2C T D6 R T D3C T D7 R T D3C T D7 R T D3F T D7 R T D3S T D7 R T D4C T A0 R T D4C T A0 R T D4C T A0 R T D4C T A0 R T D5C T A1 R T D5C T A1 R T D5

38 C A A1 I A I A I R A D5 I A I A I

C T A1 R T D6F T A2 R T D6S T A2 R T D6C T A2 R T D6C T A2 R T D7C T A3 R T D7C T A3 R T D7C T A3 R T D7C T A3 R T A0C T I R T A0C T I R T A0

50 I T I R T A0

Figure 2-5 Mapping of logical channels on the BCCH-carrier.F = FCCH, S = SCH, B = BCCH, C = CCCH, I = Idle,Dx = SDCCH, Ax = SACCH


EN/LZT 123 3314 R3A – 19 –

Eight SDCCHs can share the same physical channel, normallyTS 1 on the same frequency as the BCCHs and the CCCHs. ASACCH will be associated with every SDCCH and they willshare the same TS.

The SDCCH can be mapped together with the BCCH andCCCH on TS 0. TS 1 can then be used as a TCH. In this way weincrease the capacity on the traffic channels, but the capacitywill decrease on the SDCCH. This mapping is useful in cellswith only one carrier

F

F

S

S

B

B

C

C

F

F

S

S

C

C

C

C

F

F

S

S

D0

D0

D1

D1

F

F

S

S

D2

D2

D3

D3

F

F

S

S

A0

A2

A1

A3

-

-

BCCH + CCCH+ 4 SDCCH/4(downlink)

D3

D3

R

R

R

R

A2

A0

A3

A1

D0

D0

D1

D1

BCCH + CCCH+ 4 SDCCH/4(uplink)

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

D2

D2

R

R

Figure 2-6 Multiplexing of BCCH + CCCH + 4 SDCCH/4 on TS0

Radio Wave Propagation

Chapter 3

This chapter is designed to provide the student with an overviewof basic concepts of radio network dimensioning.


• List general properties of electromagnetic waves

• Describe how radio waves are generated

• Describe how information is superimposed on radio waves

• Describe radio wave propagation and attenuation

• Describe the pathloss concept

• Describe the origin of some fast signal variations

3 Radio Wave Propagation

EN/LZT 123 3314 R3A – i –


Table of Contents

Topic Page

WAVES ................................................................................................21

GENERATION OF RADIO WAVES .....................................................23

ISOTROPIC ANTENNA ............................................................................................... 23

HALF-WAVE DIPOLE ANTENNA................................................................................ 23

SUPERIMPOSING INFORMATION ON RADIO WAVES ....................27

AIR INTERFACE DATA.......................................................................29

FREQUENCY SPECTRUM ......................................................................................... 29

DUPLEX DISTANCE.................................................................................................... 30

CHANNEL SEPARATION............................................................................................ 30

ACCESS METHOD AND TRANSMISSION RATE ...................................................... 30

RADIO WAVE PROPAGATION...........................................................31

SIGNAL VARIATIONS.........................................................................34

INTERFERENCE..................................................................................36

INTERSYMBOL INTERFERENCE (ISI)....................................................................... 39


EN/LZT 123 3314 R3A – 21 –

WAVES

There are many seemingly different types of electromagneticwaves. They include radio waves, infrared rays, light, x-rays,and gamma rays among others. Radio waves are one type ofelectromagnetic radiation. They are typically generated asdisturbances sent out by oscillating charges on a transmittingantenna. Other types of electromagnetic radiation are caused byintense heat, atomic reactions, and stimulated emission (lasers).Regardless of its origin, an electromagnetic wave is comprisedof oscillating electric and magnetic fields. For a simple,traveling, plane wave, the electric and magnetic fields areperpendicular to each other and also to the direction ofpropagation. Waves can be described by simple sinusoidalfunctions (Figure 3-1) and are conveniently characterized bytheir wavelength, λ (the length of one cycle of oscillation). Thiscan be calculated as,

λ ⋅ =I F

where:

λ = wavelength in meters per cycle

f = frequency in cycles per second (or hertz)

c = speed of light, a constant approximately equal to3⋅108 meters/second for all electromagnetic waves.

Magnetic Field

Electric Field

H

E90°

Directionof Travel

Figure 3-1 An electromagnetic plane wave “frozen” in time


– 22 – EN/LZT 123 3314 R3A

Propagation properties are different across the frequencyspectrum. Radio waves fall in the frequency spectrum between3 Hz and 3000 GHz. This part of the spectrum is divided intotwelve bands (Figure 3-2). Since the properties of UHF wavesand frequency allocations conform closely to the needs formobile telephony, we will only be considering the Ultra HighFrequency (UHF) band in the rest of this chapter.

FREQUENCY CLASSIFICATION DESIGNATION

3 - 30 Hz

30 - 300 Hz

300 - 3000 Hz

3 - 30 kHz

Extremely Low Frequency

Voice Frequency

Very-Low Frequency

ELF

VF

VLF

30 - 300 kHz Low Frequency LF

300 - 3000 kHz Medium Frequency MF

3 - 30 MHz High Frequency HF

30 - 300 MHz Very High Frequency VHF

300 - 3000 MHz Ultra High Frequency UHF

3 - 30 GHz Super High Frequency SHF

30 - 300 GHz

300 - 3000 GHz

Extremely High Frequency EHF

Figure 3-2 Frequency spectrum bands


EN/LZT 123 3314 R3A – 23 –

GENERATION OF RADIO WAVES

High frequency radio waves are typically generated byoscillating charges on a transmitting antenna. In the case of aradio station, the antenna is often simply a long wire (a dipole)fed by an alternating voltage/current source, i.e. charge is placedon the antenna by the alternating voltage source. We can thinkof the electric field as being disturbances sent out by the dipolesource and the frequency of the oscillating electric field (theelectromagnetic wave) is the same as the frequency of thesource.

Each antenna has a unique radiation pattern. This pattern can berepresented graphically by plotting the received time-averagedpower, as a function of angle with respect to the direction ofmaximum power in a log-polar diagram. The pattern isrepresentative of the antenna’s performance in a testenvironment. However, it only applies to the free-spaceenvironment in which the test measurement takes place. Uponinstallation, the pattern becomes more complex due to the extrafactors affecting propagation under field conditions. Thus thereal effectiveness of any antenna is measured in the field.

ISOTROPIC ANTENNA

An isotropic antenna is a completely non-directional antennathat radiates equally in all directions. Since all practical antennasexhibit some degree of directivity, the isotropic antenna existsonly as a mathematical concept. The isotropic antenna can beused as a reference to specify the gain of a practical antenna (seethe appendix for a general discussion on gain/loss andlogarithmic units). The gain of an antenna referencedisotropically is the ratio between the power required in thepractical antenna and the power required in an isotropic antennato achieve the same field strength in the desired direction of themeasured practical antenna. The directive gain in relation to anisotropic antenna is called dBi.

HALF-WAVE DIPOLE ANTENNA

A half-wave dipole antenna may also be used as a gain referencefor practical antennas. The half-wave dipole is a straightconductor cut to one-half of the electrical wavelength with theradio frequency signal fed to the middle of the conductor.Figure 3-3 illustrates the radiation pattern of the half-wavedipole which normally is referred to as a dipole. Whereas the


– 24 – EN/LZT 123 3314 R3A

isotropic antenna’s three dimensional radiation pattern isspherical, the dipole antenna’s three dimensional pattern isshaped like a donut.

3-D VIEW

OF DIPOLE PATTERNVERTICAL VIEW

OF DIPOLE PATTERN

HORIZONTAL VIEWOF DIPOLE PATTERN(DIPOLE IN CENTER)

Figure 3-3 Dipole radiation pattern

Directive gain in relation to a dipole is expressed in units of“dBd”. For a dipole and an isotropic antenna with the sameinput power, the energy is more concentrated in certaindirections by the dipole. The difference in directive gainbetween the dipole and the isotropic antenna is 2.15 dB.Figure 3-4 illustrates the differences in gain between theisotropic, dipole and practical antenna. The vertical pattern(Figure 3-4) for the practical antenna is that of a directionalantenna.

IDEAL ISOTROPIC RADIATOR

THEORETICAL HALFWAVE DIPOLE

PRACTICAL ANTENNA

dBd

dBi2.15dBi = dBd + 2.15

2.15dB

Figure 3-4 Gain comparison


EN/LZT 123 3314 R3A – 25 –

When choosing an antenna for a specific application, themanufacturer’s data sheet must be consulted. The data sheetcontains information including antenna gain, beamwidth(vertical and horizontal), and graphs showing the vertical andhorizontal patterns. Examples of the graphs normally found in adata sheet are shown in Figure 3-5. The patterns displayed arethose of a directional antenna. The antenna’s gain isapproximately 15 dBd.

45 degrees

-45 degrees

0dB-5-10 5 10 010 degreesat 3dB down

VERTICAL PATTERN

HORIZONTAL PATTERN

0dB

5

10

-5

-10

0 degrees

315 degrees 45 degrees

3dB down 3dB down

3dB down

3dB down

60 degreesat 3dB down

Figure 3-5 Vertical and horizontal antenna patterns for a “real”antenna


– 26 – EN/LZT 123 3314 R3A

The beamwidth, B, is defined as the opening angle between thepoints where the radiated power is 3 dB lower than in the maindirection (Figure 3-6). Both the horizontal and verticalbeamwidths are found using the 3 dB down points, alternativelyreferred to as half-power points.

) B

Antenna lobe

Max gain-3 dB

Main direction

Max gain-3 dB

Figure 3-6 Definition of beamwidth


EN/LZT 123 3314 R3A – 27 –

SUPERIMPOSING INFORMATION ON RADIO WAVES

Information is seldom transmitted in the same frequency rangeas it was generated. The reason is that if, as an example, wewant to broadcast a 2 kHz signal, the antenna would have to be75 km long (half a wavelength). However, by translating thesignal to a much higher frequency band (e.g. the UHF band ofcellular telephony) antenna sizes drop to a few decimeters. Thisis also a prerequisite for having numerous “channels”simultaneously. Frequency translation is implemented bymodulating the amplitude, frequency or phase of a so-calledcarrier wave in accordance with the wave form of the wantedsignal. Several modulation schemes exist (e.g. amplitudemodulation) common for analog radio signals and phasemodulation. Any modulation scheme increases the carrierbandwidth and hence is a limit on the capacity of the frequencyband available. Since the bandwidth of the carrier increases ifthe bit rate increases, a high carrier frequency is necessary toobtain many different “channels”. The cell planner cannotchoose modulation techniques but the consequences of thesystem choice are very important, since carrier bandwidth andcarrier separation effects, e.g. interference properties. Wavepropagation also behaves differently in different frequencybands.

The modulation technique used in GSM 900 is called GaussianMinimum Shift Keying (GMSK). This narrow-band digitalmodulation technique is based on phase shifting. That is, bits arerepresented by continuous positive or negative phase shifts. Bychanging the phase continuously, sharp discontinuities areavoided, thus narrowing the bandwidth of the modulated carrier.GMSK modulation also involves filtering the incoming bitstream with a Gaussian filter to obtain a more narrow bandwidthof the modulated carrier. In fact the full width at half maximumof the carrier becomes 162 kHz, corresponding nicely to the 200kHz carrier separation.

Transmitting the information on the air interface in digitizedform has an advantage over analogue techniques, in that sincechannel coding protects bits, the signal is less sensitive toperturbations. In addition, it enables Time Division MultipleAccess (TDMA) which means that one carrier frequency can beused for several connections. Each connection uses only oneparticular time slot (out of the eight available in GSM). This hasthe advantage that the mobile is released fromtransmitting/receiving continuously and can perform, e.g.measurements on neighboring cells. One main advantage with


– 28 – EN/LZT 123 3314 R3A

TDMA is that it enables Mobile Assisted Hand Over (MAHO)which is essential for effective connection control.


EN/LZT 123 3314 R3A – 29 –

AIR INTERFACE DATA

Below is a summary of some important air interface data forGSM 900, GSM 1800, and GSM 1900.

FREQUENCY SPECTRUM

Different frequency bands are used for GSM 900, GSM 1800,and GSM 1900. In some countries, operators apply for theavailable frequencies. In other countries (e.g. the United States),operators purchase frequency bands at auctions.

In December of 1994 the Federal Communications Commission(FCC) auctioned “broadband” licenses to prospective operatorsoffering personal communications services. Each operator ownsthe rights to the licenses for a period of ten years. The UnitedStates is divided into 51 regions or Major Trading�Areas (MTA)and 493 Basic Trading Areas (BTA). The FCC issued twoGSM 1900 licenses for each MTA and four for each BTA. OneMTA can be geographically as large as a state, while one BTAcan be compared in size to a large city. BTAs are designed foruse in major metropolitan areas.

The FCC has specified the frequency range and output power.The frequency band is divided into six frequency blocks(Figure 3-7): three duplex blocks A, B, and C (90 MHz totalspectrum bandwidth) and three other duplex blocks D, E, and F(30 MHz total spectrum bandwidth).

1850 1910 1930 1990

A D B E F C A D B E F CUnlicensed

MTAsUplink

MTAsDownlink

A,B 2 x 15 MHz MTAC 2 x 15 MHz BTAD, E, F 2 x 5 MHz BTA

60 MHz 60 MHz

Figure 3-7 Spectrum allocation for GSM 1900 in United States.140 MHz for GSM 1900; 120 MHz licensed and 20 MHzunlicensed


– 30 – EN/LZT 123 3314 R3A

DUPLEX DISTANCE

The distance between the uplink and downlink frequencies isknown as duplex distance. The duplex distance is different forthe different frequency bands (Figure 3-8).

Standard GSM 900 GSM 1800 GSM 1900Duplex dist. 45 MHz 95 MHz 80 MHz

Figure 3-8 Duplex differences for different frequency bands

CHANNEL SEPARATION

The distance between adjacent frequencies on the uplink or thedownlink is called channel separation. The channel separation is200 kHz, regardless of the standard chosen from the onesmentioned above. This separation is needed to reduceinterference from one carrier to another neighboring frequency.

ACCESS METHOD AND TRANSMISSION RATE

GSM has chosen the Time Division Multiple Access (TDMA)concept for access. In GSM, there are eight TDMA time slotsper frame (Figure 3-9). Each time slot is 0.577 ms long and hasroom for 156.25 bits (148 bits of information and a 8.25 bitslong guard period) yielding a bit rate on the air interface of270.8 kbits.

0 1 2 3 4 5 6 7

4.615 ms

3 57 1 26 1 57 3

Burst148 Bits

156.25 Bits0.577 ms

DataTrainingData

Figure 3-9 Basic TDMA frame, timeslot, and burst structures


EN/LZT 123 3314 R3A – 31 –

RADIO WAVE PROPAGATION

In this course we are primarily interested in the transmission lossbetween two antennas: the transmitter/emitter and the receiver.Many factors including absorption, refraction, reflection,diffraction, and scattering affect the wave propagation.However, in free space an electromagnetic wave travelsindefinitely if unimpeded. This does not mean there are notransmission losses, as we will see in this first simple modelwhere isotropic emission from the transmitter and line of sightbetween the two antennas separated by a distance, G, in freespace are assumed (Figure 3-10).

Figure 3-10 Radio wave propagation in free space

Since an isotropic antenna by definition distributes the emittedpower, 3W, equally in all directions, the power density, 6U, (powerper area unit) decreases as the irradiated area, �πG�, at distanceG, increases, i.e.

63GU

W=4 2π

If the transmitting antenna has a gain, *W, it means that it isconcentrating the radiation towards the receiver. The powerdensity at the receiving antenna increases with a factorproportional to *W, i.e.

63*

GU

W W=4 2π

The power received by the receiving antenna, 3U, is proportionalto the effective area, $U, of that antenna, i.e.

3 6 $U U U

= ⋅


– 32 – EN/LZT 123 3314 R3A

It can be shown that the effective area of an antenna isproportional to the antenna gain, *U, and the square of thewavelength, λ, of the radio wave involved, i.e.

$*

U

U=λπ

2

4

and hence the received power becomes

( )33* *

GU

W W U=λ

π

2

24

The transmission loss can be calculated as the ratio between thetransmitted power and received power, i.e.

( )

ORVV33

G* *

W

U W U

= =4 2

2

πλ

Radio engineers work with the logarithmic unit dB so thetransmission loss, /, then becomes

( )

( ) ( ) ( )/ ORVVG

* *G

* *W U

U W= =

=

− −��

��

��

�

�log log log log log

πλ

πλ

Radio engineers treat the antenna gains, 10log(*U) and10log (*W), separately, so what is given in the literature as thepathloss, /S, is only the term 20log(�πG�λ). To make it clearer,the pathloss in free space is

/G

S=

20

4log

πλ

Note that the wavelength dependency of the pathloss does notcorrespond to losses in free space as such. It is a consequence ofthe finite effective receiver area.

This expression is fairly general. The only thing which changeswhen we improve our models is the expression for the pathloss.The antenna gain is normally given in dB(i), i.e. as 10log(*),where gain means a reduction of the total transmission loss, /,between a transmitting and receiving antenna.

This model helps us to understand the most important featuresof radio wave propagation. That is, the received power decreaseswhen the distance between the antennas increases and thetransmission loss increases when the wavelength decreases (oralternatively when the frequency increases).


EN/LZT 123 3314 R3A – 33 –

For cell planning, it is very important to be able to estimate thesignal strengths in all parts of the area to be covered, i.e. topredict the pathloss. The model described in this section can beused as a first approximation. However, more complicatedmodels exist. Improvements can be made by accounting for:

• the fact that radio waves are reflected towards the earth’ssurface (the conductivity of the earth is thus an importantparameter).

• accounting for transmission losses due to obstructions in theline of sight.

• the finite radius of the curvature of the earth.

• the topographical variations in a real case as well as thedifferent attenuation properties of different terrain types suchas forests, urban areas, etc.

The best models used are semi-empirical, i.e. based onmeasurements of pathloss/attenuation in various terrains. Theuse of such models are motivated by the fact that radiopropagation can not be measured everywhere. However, ifmeasurements are taken in typical environments, the parametersof the model can be fine-tuned so that the model is as good aspossible for that particular type of terrain.


– 34 – EN/LZT 123 3314 R3A

SIGNAL VARIATIONS

The models described in the previous section can be used toestimate the average signal level (called the “global mean”) atthe receiving antenna. However, a radio signal envelope iscomposed of a fast fading signal super-imposed on a slow fadingsignal (Figure 3-11). These fading signals are the result ofobstructions and reflections. They yield a signal which is thesum of a possibly weak, direct, line-of-sight signal and severalindirect or reflected signals.

The fast fading signal (peak-to-peak distance ≈ λ/2) is usuallypresent during radio communication due to the fact that themobile antenna is lower than the surrounding structures such astrees and buildings. These act as reflectors. The resulting signalconsists of several waves with various amplitudes and phases.Sometimes these almost completely cancel out each other. Thiscan lead to a signal level below the receiver sensitivity. In openfields where a direct wave is dominating, this type of fading isless noticeable.

Short-term fading is Rayleigh distributed with respect to thesignal voltage. Therefore, it is often called Rayleigh fading. Thistype of fading affects the signal quality, and as a result somemeasures must be taken to counter it.

The first and most simple solution is to use more power at thetransmitter(s), thus providing a fading margin. Another way toreduce the harm done by Rayleigh fading is to use spacediversity, which reduces the number of deep fading dips.Diversity means that two signals are received which haveslightly different “histories” and, therefore, the “best” can beused. With the efficient space diversity system used by Ericsson,voice quality is much improved.


EN/LZT 123 3314 R3A – 35 –

SS at Rx-antenna

DistanceVariations due toShadowing

Global means

Variations due toRayleigh fading

Figure 3-11 Short-term (fast) and long-term (slow) fading

The signal variation received, if we smooth out the short-termfading, is called the “local mean”. Its power, often called thelocal average power, expressed in a logarithmic scale isnormally distributed. Therefore, this slow fading is called “log-normal fading”. If we drive through a flat desert without anyobstructions, the signal varies slowly with distance. However, innormal cases the signal path is obstructed.

Obstructions near the mobile (e.g. buildings, bridges, trees, etc.)cause a rapid change of the local mean (in the range of five tofifty meters), while topographical obstructions cause a slowersignal variation. Because log-normal fading reduces the averagestrength received, the total coverage from the transmitter isreduced. To combat this, a fading margin must be used.Problems generated by multi-path reflections are made moresevere by log-normal fading since the direct beam is weakenedby the obstructing object.

Phases between various reflected waves are different. This isdue to the fact that they propagate over different distances orequivalently use different times to reach the receiver. This timedispersion can cause particular problems if the phase differencebetween the reflected waves is very large. For GSM 900, a largephase difference is generally several thousands of wavelengths(i.e. one kilometer or more). In this case, different waves addedtogether in the receiver carry information about differentsymbols (bits). If the direct wave is weak, and consequently thereflected waves are relatively strong, it can be difficult todetermine which symbol (bit) was transmitted.


– 36 – EN/LZT 123 3314 R3A

INTERFERENCE

Cellular systems are often interference limited rather than signalstrength limited, some elementary information about differentproblems associated with the re-use of carriers is provided inthis section.

dB

C

I

Carrier, f1

CI

>0 dB

Distance

Interferer, f1

Figure 3-12 Co-channel interference

Co-channel interference is the term used for interference in acell by carriers with the same frequency present in other cells.Figure 3-12 illustrates this situation. Since the same carrierfrequency is used for the wanted carrier as for the unwantedcarrier, quality problems can arise if the signal from theunwanted carrier is too strong.

The GSM specification states that the signal strength ratio, C/I,between the carrier, C, and the interferer, I, must be larger than 9dB. However, Ericsson recommends using C/I >12 dB as aplanning criterion. If frequency hopping is implemented, it addsextra diversity to the system corresponding to a margin ofapproximately 3 dB, i.e.

C/I > 12 dB (without frequency hopping)

C/I > 9 dB (with frequency hopping)

Adjacent carrier frequencies (i.e. frequencies shifted ±200 kHz)with respect to the carrier cannot be allowed to have too strong asignal strength either. Even though they are at differentfrequencies, part of the signal can interfere with the wantedcarrier’s signal and cause quality problems (Figure 3-13). TheGSM specification states that the signal strength ratio, C/A,between the carrier and the adjacent frequency interferer, A,must be larger than -9 dB.


EN/LZT 123 3314 R3A – 37 –

However, adjacent channel interference also degrades thesensitivity as well as the C/I performance. During cell planningthe aim should be to have C/A higher than 3 dB, according toEricsson, i.e.:

C/A > 3 dB

Adjacent frequencies must be avoided in the same cell andpreferably in neighboring cells as well.

dB

A

C

Carrier, f1CA

<0 dB

Distance

Adjacent, f2f2 = f1 ± 200 kHz

Figure 3-13 Adjacent channel interference

By re-using the carrier frequencies according to well-proven re-use patterns (Figure 3-14 and Figure 3-15), neither co-channelinterference nor adjacent channel interference will causeproblems, provided the cells have isotropic propagationproperties for the radio waves.

Unfortunately, this is hardly ever the case. Cells vary in sizedepending on the amount of traffic they are expected to carry,and nominal cell plans must be verified by means of predictionsor radio measurements to ensure that interference does notbecome a problem.

The re-use patterns recommended for GSM are 4/12- and 3/9-patterns. 4/12 means that each cluster has four three-sector sitessupporting twelve cells (Figure 3-14).


– 38 – EN/LZT 123 3314 R3A

B2B3

C1

C2C3

B1

B2B3

C1

C2C3

B1

B2B3

C1

C2C3

A1

A2A3

D3

D2D1

A1

A2A3

D3

D2D1

A1

A2A3

D3

C2C3

B1

B2B3

C1

C2C3

B1

B2B3

C1

C2C3

B1

B2B3

D3

D2D1

A1

A2A3

D3

D2D1

A1

A2A3

D3

D2D1

A1

B2B3

C1

C2C3

B1

B2B3

C1

C2C3

B1

B2B3

C1

C2C3

A1

A2A3

D3

D2D1

A1

A2A3

D3

D2D1

A1

A2A3

D3

Figure 3-14 4/12 re-use pattern

The re-use pattern in Figure 3-14 is compatible with thecondition C/I>12 dB. A shorter re-use distance, given a smallerC/I-ratio, is used in the 3/9-pattern (Figure 3-15). This re-usepattern is recommended only if frequency hopping isimplemented. It has a higher channel utilization because thecarriers are distributed among nine cells rather than 12.

C1

C2C3

A1

A2A3

B1

B2B3

C1

C2C3

A1

A2A3

A2A3

B1

B2B3

C1

C2C3

A1

A2A3

B1

B2B3

C1

C1

C2C3

A1

A2A3

B1

B2B3

C1

C2C3

A1

A2A3

A2A3

B1

B2B3

C1

C2C3

A1

A2A3

B1

B2B3

C1

C1

C2C3

A1

A2A3

B1

B2B3

C1

C2C3

A1

A2A3

A2A3

B1

B2B3

C1

C2C3

A1

A2A3

B1

B2B3

C1


Other re-use patterns with much higher re-use distances (such asthe 7/21), must be used for systems which are more sensitive tointerference; e.g., analog mobile telephone systems.


EN/LZT 123 3314 R3A – 39 –

INTERSYMBOL INTERFERENCE (ISI)

InterSymbol Interference (ISI) is caused by excessive timedispersion. It may be present in all cell re-use patterns. ISI canbe thought of as co-channel interference. However in this casethe interferer, R, is a time delayed reflection of the wantedcarrier. According to GSM specifications, the signal strengthratio C/R must be larger than 9 dB (compared to the C/I-criterion). However, if the time delay is smaller than 15 µs (i.e.4 bits or approximately 4.4 km), the equalizer can solve theproblem. ISI is not affected by the re-use pattern chosen, but isstill an issue for the cell planner.

How can the cell planner avoid ISI in the cellular network?Normally, the reflected waves are much weaker than the directwave. However, if the direct wave is obstructed (shadowed), orif the reflected wave has a very advantageous path ofpropagation, the C/R ratio may creep down to dangerous valuesif the time delay is outside the equalizer window. Hence, timedispersion may cause problems in environments with, e.g.mountains, lakes with steep or densely built shores, hilly cities,and high metal covered buildings. The location of the BTS canthus be crucial. Figure 3-16 and Figure 3-17 suggest somepossible solutions.

D2

D0

D1

Figure 3-16 Locating the BTS close to the reflecting object tocombat ISI

Mountain Site with antennapointing away

Figure 3-17 Pointing the antenna away from the reflectingobject to combat ISI

Traffic

Chapter 4

This chapter is designed to provide the student with an overviewof basic concepts of network dimensioning.


• Define the terms “traffic” and “Grade of Service” (GoS)

• Use Erlang’s B-table to dimension the number of channelsneeded in the system

• Describe channel utilization

4 Traffic

EN/LZT 123 3314 R3A – i –

4 Traffic

Table of Contents

Topic Page

TRAFFIC AND CHANNEL DIMENSIONING .......................................41

CHANNEL UTILIZATION.....................................................................45

4 Traffic

EN/LZT 123 3314 R3A – 41 –

TRAFFIC AND CHANNEL DIMENSIONING

Cellular system capacity depends on a number of differentfactors. These include:

• The number of channels available for voice and/or data

• The grade of service the subscribers are encountering in thesystem

Traffic theory attempts to obtain useful estimates, e.g. thenumber of channels needed in a cell. These estimates depend onthe selected system and the assumed or real behavior of thesubscribers.

What is traffic? Traffic refers to the usage of channels and isusually thought of as the holding time per time unit (or thenumber of “call hours” per hour) for one or several circuits(trunks or channels). Traffic is measured in Erlangs (E), e.g., ifone subscriber is continuously on the telephone, this wouldgenerate one call per hour or 1 E of traffic.

How much traffic can one cell carry? That depends on thenumber of traffic channels available DQG�WKH�DPRXQW�RIFRQJHVWLRQ�ZKLFK�LV�DFFHSWDEOH (by both the customer and theprovider), the so-called Grade of Service (GoS). Differentassumptions on subscriber behavior lead to different answers tothis question. Erlang’s (a Danish traffic theorist) B-table is basedon the most common assumptions used. These assumptions are:

• no queues

• number of subscribers much higher than number of trafficchannels available

• no dedicated (reserved) traffic channels

• poisson distributed (random) traffic

• blocked calls abandon the call attempt immediately

This is referred to as a “loss system”. Erlang’s B-table relatesthe number of traffic channels, the GoS, and the traffic offered.This relationship is tabulated in Figure 4-1. Assuming that onecell has two carriers, corresponding typically to 2x8-2=14 trafficchannels and a GoS of 2% is acceptable, the traffic that can beoffered is A=8.2003 E (Figure 4-1).

This number is interesting if an estimate on the average trafficper subscriber can be obtained. Studies show that the averagetraffic per subscriber during the busy hour is typically 15-20 mE


– 42 – EN/LZT 123 3314 R3A

(this can correspond to, e.g. one call lasting 54-72 seconds perhour). Dividing the traffic that one cell can offer, Acell=8.20 E,by the traffic per subscriber, here chosen as Asub=0.025 E, thenumber of subscribers one cell can support is derived as8.20/0.025 = 328 subscribers.

Q .007 .008 .009 .01 .02 .03 .05 .1 .2 .4 Q

1 .00705 .00806 .00908 .01010 .02041 .03093 .05263 .11111 .25000 .66667 1

2 .12600 .13532 .14416 .15259 .22347 .28155 .38132 .59543 1.0000 2.0000 2

3 .39664 .41757 .43711 .45549 .60221 .71513 .89940 1.2708 1.9299 3.4798 3

4 .77729 .81029 .84085 .86942 1.0923 1.2589 1.5246 2.0454 2.9452 5.0210 4

5 1.2362 1.2810 1.3223 1.3608 1.6571 1.8752 2.2185 2.8811 4.0104 6.5955 5

6 1.7531 1.8093 1.8610 1.9090 2.2759 2.5431 2.9603 3.7584 5.1086 8.1907 6

7 2.3149 2.3820 2.4437 2.5009 2.9354 3.2497 3.7378 4.6662 6.2302 9.7998 7

8 2.9125 2.9902 3.0615 3.1276 3.6271 3.9865 4.5430 5.5971 7.3692 11.419 8

9 3.5395 3.6274 3.7080 3.7825 4.3447 4.7479 5.3702 6.5464 8.5217 13.045 9

10 4.1911 4.2889 4.3784 4.4612 5.0840 5.5294 6.2157 7.5106 9.6850 14.677 10

11 4.8637 4.9709 5.0691 5.1599 5.8415 6.3280 7.0764 8.4871 10.857 16.314 11

12 5.5543 5.6708 5.7774 5.8760 6.6147 7.1410 7.9501 9.4740 12.036 17.954 12

13 6.2607 6.3863 6.5011 6.6072 7.4015 7.9667 8.8349 10.470 13.222 19.598 13

14 6.9811 7.1154 7.2382 7.3517 8.2003 8.8035 9.7295 11.473 14.413 21.243 14

15 7.7139 7.8568 7.9874 8.1080 9.0096 9.6500 10.633 12.484 15.608 22.891 15

16 8.4579 8.6092 8.7474 8.8750 9.8284 10.505 11.544 13.500 16.807 24.541 16

17 9.2119 9.3714 9.6171 9.6516 10.656 11.368 12.461 14.522 18.010 26.192 17

18 9.9751 10.143 10.296 10.437 11.491 12.238 13.385 15.548 19.216 27.844 18

19 10.747 10.922 11.082 11.230 12.333 13.115 14.315 16.579 20.424 29.498 19

20 11.526 11.709 11.876 12.031 13.182 13.997 15.249 17.613 21.635 31.152 20

21 12.312 12.503 12.677 12.838 14.036 14.885 16.189 18.651 22.848 32.808 21

22 13.105 13.303 13.484 13.651 14.896 15.778 17.132 19.692 24.064 34.464 22

23 13.904 14.110 14.297 14.470 15.761 16.675 18.080 20.737 25.281 36.121 23

24 14.709 14.922 15.116 15.295 16.631 17.577 19.031 21.784 26.499 37.779 24

25 15.519 15.739 15.939 16.125 17.505 18.483 19.985 22.833 27.720 39.437 25

26 16.334 16.561 16.768 16.959 18.383 19.392 20.943 23.885 28.941 41.096 26

27 17.153 17.387 17.601 17.797 19.265 20.305 21.904 24.939 30.164 42.755 27

28 17.977 18.218 18.438 18.640 20.150 21.221 22.867 25.995 31.388 44.414 28

29 18.805 19.053 19.279 19.487 21.039 22.140 23.833 27.053 32.614 46.074 29

30 19.637 19.891 20.123 20.337 21.932 23.062 24.802 28.113 33.840 47.735 30

31 20.473 20.734 20.972 21.191 22.827 23.987 25.773 29.174 35.067 49.395 31

32 21.312 21.580 21.823 22.048 23.725 24.914 26.746 30.237 36.295 51.056 32

Figure 4-1 Part of Erlang’s B-table, yielding the traffic (inErlangs) as a function of the GoS (columns) and number oftraffic channels (rows)

4 Traffic

EN/LZT 123 3314 R3A – 43 –

Dimensioning the network now implies using demographic datato determine the sizes of the cells. The preceding example issimplified, however, it provides an understanding of what ismeant by traffic and traffic dimensioning.

The problem may be that given a number of subscribers in oneparticular area (e.g., an airport), how many carriers do we needto support the traffic if only one cell is to be used?Dimensioning a whole network while maintaining a fixed cellsize means estimating the number of carriers needed in eachcell. In addition, traffic is not constant. It varies between day andnight, different days, and with a number of other factors. Mobiletelephony implies mobility and hence subscribers may movefrom one area to another during the course of a day.

It is important that the number of signaling channels (SDCCHs)is dimensioned as well, taking into account the estimated systembehavior in various parts of the network. For example, cellsbordering a different location area may have lots of locationupdating, and cells on a highway probably have manyhandovers. In order to calculate the need for SDCCHs thenumber of attempts for every procedure that uses the SDCCH aswell as the time that each procedure holds the SDCCH must betaken into account. The procedures are; location updating,periodic registration, IMSI attach/detach, call setup, SMS,facsimile and supplementary services. The number of falseaccesses must also be estimated. This is typically quite a highnumber, but still small compared to the traffic.

When the grade of service that should be used to consult thetraffic tables is chosen, the fact that calls go through twodifferent devices (SDCCH and TCH) must be kept in mind.Figure 4-2 explains this situation.

(1-GoS1)A’

GoS2(1-GoS1)A

(1-GoS1)A (1-GoS2)(1-GoS1)A

GoS1(A+A’)+GoS2(1-GoS1)A

GoS1(A+A’)

A

A’

SDCCH TCH

Figure 4-2 A call goes through two different devices


– 44 – EN/LZT 123 3314 R3A

If A is the traffic on the SDCCH for normal call and fax setupand A’ is the traffic that accounts for the rest of the proceduresperformed on the SDCCH, we obtain the “global” grade ofservice, GoST, for calls which go through an SDCCH and aTCH:

GoST = GoS1 + (1-GoS1)GoS2

where GoS1 is the grade of service on the SDCCH and GoS2 isthe grade of service on the TCH. Since we can assume that thegrade of service given by the operator corresponds to the trafficchannels, i.e. GoS2, it is obvious that a better grade of service isneeded on the SDCCH. This is not only important becausesignaling is performed on it, but also because the “global” gradeof service would otherwise be higher than the grade of serviceexpected by the operator.

The most accurate SDCCH dimensioning is achieved by lookingat the congestion level in the TCHs and the SDCCHs for thespecific cell. The optimum configuration is achieved byselecting a configuration with as many TCHs as possible,without letting the GoS1 exceed 1/4 of GoS2 (with only fourSDCCHs, GoS1 ≤ GoS2/2).

SDCCHs can only be allocated in steps of four or eight, asrequired by the GSM specifications, i.e., one cell can have four,eight, twelve, etc. SDCCHs, but no more than 128 SDCCHs canbe allocated to one cell.

4 Traffic

EN/LZT 123 3314 R3A – 45 –

CHANNEL UTILIZATION

Assume the task is to find the necessary number of trafficchannels for one cell to serve subscribers with a traffic of 33 E.The GoS during the busy hour is not to exceed 2%. Byconsidering the above requirements and consulting Erlang’s B-table, 43 channels are found to be needed (Figure 4-3).

Q .007 .008 .009 .01 .02 .03 .05 .1 .2 .4 Q

43 30.734 31.069 31.374 31.656 33.758 35.253 37.565 42.011 49.851 69.342 43

Figure 4-3 Part of Erlang’s B-table for 43 channels giving theoffered traffic (E) as a function of the GoS (%)

Assume five cells are designed to cover the same area as thesingle cell. These five cells must handle the same amount oftraffic as the cell above, 33 E. Acceptable GoS is still 2%. First,the total traffic is divided among the cells (Figure 4-4). Trafficdistribution over several cells results in a need for morechannels than if all traffic had been concentrated in one cell.

This illustrates the fact that it is more efficient to use manychannels in a larger cell than vice versa. To calculate the channelutilization, the traffic offered is reduced by the GoS of 2%(yielding the traffic served) and dividing that value by thenumber of channels (yielding the channel utilization).

With 43 channels (as in the previous single cell example), thechannel utilization is 33.083/ 43 = 77%, i.e., each channel isused approximately 77% of the time. However, by splitting thiscell into smaller cells, more traffic channels are required hencethe channel utilization decreases.

Cell Traffic (%) Traffic (E) No. ofchannels

Channelutilization (%)

A 40 13.20 21 62

B 25 8.25 15 54

C 15 4.95 10 49

D 10 3.30 8 40

E 10 3.30 8 40

Σ 100 33.00 62

Figure 4-4 What happens when a certain amount of traffic isdistributed over several cells?


– 46 – EN/LZT 123 3314 R3A

As we will see in the following chapter, capacity andinterference problems prevent us from always using the mosteffective channel utilization scheme and so solutions in realnetworks must compromise between efficiency (i.e. cost) andquality.

System Balance

Chapter 5

This chapter is designed to provide the student with an overviewof system balancing.


• Balance a cell in a cellular system

• Describe the concept of effective radiated power

5 System Balance

EN/LZT 123 3314 R3A – i –

5 System Balance

Table of Contents

Topic Page

SYSTEM BALANCING ........................................................................47

EFFECTIVE RADIATED POWER........................................................50

5 System Balance

EN/LZT 123 3314 R3A – 47 –

SYSTEM BALANCING

An area is usually referred to as being covered if the signalstrength received by an MS, around 95% of that area, is higherthan some design value, around -90 dBm (1 pW). However,coverage in a two-way radio communication system is decidedby the weakest transmission direction. As a result, an equivalentdefinition exists for uplink coverage (i.e. the signal received bythe BTS from an MS in (95% of) an area must be higher thansome minimum value). It makes no sense to have differentcoverage on uplink and downlink because this causes an excessamount of energy to be dissipated into the system adding extrainterferences and costs. A system balance must be obtainedbefore coverage calculation can start. However, practiceindicates that it is advantageous to have (if possible) a somewhathigher output (2 to 3 dB) power from the base station than theone strictly calculated from power balance calculations asbelow.

To achieve this balance, it is necessary to make sure that thesensitivity limit, 06VHQV, of the MS (for downlink transmission)is reached at the same point as the sensitivity limit, %76VHQV, ofthe BTS (for uplink transmission). The input power, 3LQ06, atthe MS receiver equals the output power, 3RXW%76, of the BTSplus gains and losses. If input power is set equal to thesensitivity level, for some pathloss (Figure 5-1)

3LQ 3RXW /F /I *D /S *D /I 0606 %76 %76 %76 %76 06 06 VHQV

= − − + − + − =

and equivalently for the input power, 3LQ%76, at the BTS receiver

3LQ 3RXW /I *D /S *D *G /I %76%76 06 06 06 %76 %76 %76 VHQV

= − + − + + − =

a system balance can be found.

For some configurations the duplex loss, /GXSO%76, can beimportant. If polarization diversity is used, it may be necessaryto introduce a slant polarization (±45°) downlink loss, /VODQW%76(if a ±45° pol antenna is used for downlink transmission), ofabout 1.5 dB.


– 48 – EN/LZT 123 3314 R3A

Tx

TxReceiver

Divider

Rx

Feeder

Feeder

Feeder

Rx

Combiner

LCBTS

Feeder

LFBTS

LFBTS

PoutMS PinMS

LFMS

GAMS

GABTS

GDBTS

GABTS

{

Lp

Lp

PinBTS

without TMA

PinBTS

withTMA

Cabinet

PoutBTS

Figure 5-1 Schematics of the components included in a systembalance. Abbreviations have the following meaning: G=Gain,L=Loss, a=Antenna, f=Feeder & jumper, c=Combiner,MS=Mobile Station, BTS=Base Transceiver Station andd=Diversity, TMA=Tower Mounted Antenna

Assuming that the pathloss, Lp, is identical on uplink anddownlink (a good assumption since the difference in frequencyis only on the order of 5%) and that the BTS transmitting andreceiving antennas have the same gain (not necessary)subtracting the second equation from the first

06 %76 3RXW 3RXW /F *GVHQV VHQV %76 06 %76 %76

− = − − −

is obtained and after rearranging

( )3RXW 3RXW /F *G 06 %76%76 06 %76 %76 VHQV VHQV

= + + + −

The BTS output power, 3RXW%76, measured at the TX outputmust be higher than the output power of the MS, 3RXW06, by avalue corresponding to the sum of the diversity gain, *G%76, thecombiner loss, /F%76 , and the difference in sensitivity (06VHQV- %76VHQV). Note that the reference points for the sensitivity maybe different than what is shown in Figure 5-1 when balancing,e.g., a GSM 1800 system using an Antenna Low NoiseAmplifier (ALNA).

5 System Balance

EN/LZT 123 3314 R3A – 49 –

Ericsson uses nominal values of the sensitivity for cell planningpurposes, e.g., balancing the system for GSM 900 class 4 mobilestations, i.e. 3RXW06=2 W or 33 dBm, using *G%76=3.5 dB,/F%76=3 dB, and using cell planning values for the sensitivities06VHQV=-104 dBm and %76VHQV=-107 dBm, an output power ofthe BTS of 42.5 dBm is obtained, i.e.,

3RXW%76 = 33 + 3 + 3.5 + (-104 + 107) = 42.5 dBm

Hence, an 18 W BTS is needed. The output power of the BTSneeds to be higher than the output power of the MS because notonly is the BTS more sensitive (and hence can accept a smallersignal strength), it also has an extra loss when transmitting,/F%76, and an extra gain when receiving, *G%76. The balance isindependent of the BTS antenna gain, therefore the coverage cannow be changed by changing the antenna gain, since it issymmetrical. That is, increasing the coverage downlink byincreasing the antenna gain is matched by a correspondingincrease in coverage on the uplink.

In conclusion, the BTS output power should never be changedonce the system is balanced for a particular configuration andmobile class. However, if “smaller cells” are desired, the powercan be decreased because it can be matched by a corresponding,forced decrease in the output power of the MS.


– 50 – EN/LZT 123 3314 R3A

EFFECTIVE RADIATED POWER

The power radiated by the antenna (Figure 5-1) is given by3RXW%76 - /F%76 - /I%76. However since antennas have gain,concentrating the radiation in a certain direction, the effectiveradiated power, (53, or (L53, is often used to describe theantenna since adding the pathloss (and receiving antenna gainand feeder loss) to the (L53 yields the received signal strength.

(L53 3RXW /F /I *D%76 %76 %76 %76

= − − +

The L (lack of L) indicates that the antenna gain, *D%76, has beengiven with respect to an isotropic (dipole) antenna, i.e. in unitsof dBi (dBd). The effective radiated power can be interpreted asthe power that an isotropic (dipole) antenna would have to befed with if it were to emit the same power density in a givendirection as an antenna with the gain *D%76. As mentioned in anearlier section, the difference between (L53 and (53 is 2.15 dB.Hence, the relation between (L53 and (53 is (L53 = (53 +2.15.

Coverage Predictions

Chapter 6

This chapter is designed to provide the student with an overviewof elaborate radio wave coverage predictions.


• Explain why different models are used in differentenvironments

• Discuss more elaborate radio wave propagation models andthe predictions they can make

6 Coverage Predictions

EN/LZT 123 3314 R3A – i –


Table of Contents

Topic Page

INTRODUCTION..................................................................................51

FLAT CONDUCTIVE EARTH ..............................................................52

KNIFE EDGE DIFFRACTION ..............................................................53

FIELD MEASUREMENTS AND SEMI-EMPIRICAL MODELS ............54

ALGORITHM 9999...............................................................................57

URBAN MODEL...................................................................................58

MICROCELL MODELS........................................................................59

OVERVIEW.................................................................................................................. 59

LINE-OF-SITE MODELLING ....................................................................................... 59

NON LINE-OF-SITE MODELLING .............................................................................. 59

IN-BUILDING MODELLING ......................................................................................... 60

OPEN AREA MODELLING.......................................................................................... 60

POINTS AFFECTED BY MULTIPLE SOURCES......................................................... 60


EN/LZT 123 3314 R3A – 51 –

INTRODUCTION

It is important to be able to estimate cell coverage, not only todetermine the size of the cell, but also to be able to estimateinterference. The definition of coverage is usually the following:an area is considered covered if in 90 percent of that area, thesignal received by the mobile station is larger than some designvalue, e.g., -90 dBm, i.e., SS design = -90 dBm i.e.,

3LQ06�≥�66

GHVLJQ

The signal strength requirement is estimated by adding marginsto the MS receiver sensitivity. These are fast and slow fadingmargins, interference margins, margins for body loss, andpossibly additional margins for in-car and indoor coverage. Themargins depend on the type of environment and operatorrequirements.

It is very important to be able to estimate the signal strength inall parts of the area to be covered, i.e., to predict the pathloss.There are more elaborate models than the one discussed inchapter three, “Radio Wave Propagation”. Improvements can bemade by taking into account:

• the fact that radio waves are reflected towards the earthsurface (the conductivity of the earth is then an importantparameter)

• the transmission losses due to obstructions in the line ofsight

• the finite radius of the curvature of the earth

• the terrain type in a real case, as well as the differentattenuation properties of different land usages such asforests, urban areas, etc.

The best models used are semi-empirical, i.e. based onmeasurements of path loss/attenuation in various terrains. Theuse of such models are motivated by the fact that radiopropagation cannot be measured everywhere. However, ifmeasurements can be performed in typical environments,parameters of the model can be adjusted so that the modelbecomes a good approximation for that particular type of terrain.This chapter briefly discusses a few of these more elaboratemodels.


– 52 – EN/LZT 123 3314 R3A

FLAT CONDUCTIVE EARTH

MobileBase

h1

h2

d

Figure 6-1 Radio wave propagation over flat conductive earth

In Figure 6-1, reflections against the surface of the earth aretaken into account. If we assume an unobstructed propagationthrough free space, the signal at the receiving antenna can beseen as the sum of one direct signal and the reflected signal, ifwe also assume that the earth is a perfect conductor (hardly agood assumption, except possibly for sea water), i.e. loss freereflection, this yields (for the received power at the receivingantenna) the interference term:

( )3

3* *K KG

GU

W U W

=

λ

πλ

π

2 2 1 2

2

2

2

sin

which is the squared sum of the field amplitudes from the directand reflected wave. (See chapter three, “Radio WavePropagation” for the explanation of the symbols.) Assuming thath1h2<<λd (i.e., small angles, the sine function can be replacedwith its argument (radians) in and so

( )3

3* * K K

GU

W U W= 1 2

2

4

or

( ) ( )/33

GK K

* *W

U

U W=

=

− −10 20 10 10

2

1 2

log log log log

The expression 20log(G��K�K�)) corresponds to the path loss.


EN/LZT 123 3314 R3A – 53 –

KNIFE EDGE DIFFRACTION

Additional path loss due to objects obstructing the line of sightcan be taken into account by calculating the (Fresnel) diffractionpattern at the receiver. The intensity is a function of the height ofthe obstruction above (or below) the line of sight as well as thedistances transmitter-object and receiver-object (Figure 6-2).

TX RXh

d1 d2

Figure 6-2 Knife edge diffraction

Derivations of the expression is somewhat lengthy, so here wemust be satisfied with expressing the additional attenuationcaused by these so-called “knife edges” in a diagram (Figure 6-3). The additional attenuation is read as a function of theparameter ν, which is given as

( )ν

λ=

+K

G G

G G

2 1 2

1 2

ν

0

Figure 6-3 Knife edge diffraction loss as a function of ν


– 54 – EN/LZT 123 3314 R3A

FIELD MEASUREMENTS AND SEMI-EMPIRICAL MODELS

The models discussed previously do not take into account thetopographical variations in a real environment nor the differentattenuation properties of different land usages such as forests,urban areas, etc. Although calculations taking all details intoaccount are possible, they are tremendously time consuming andnot practical to use for the cell planner. Indeed, empirical datacan be used. An example of such data is shown in Figure 6-4.This figure is a depiction of measurements made in 1968 by aJapanese engineer, Okumura. It is interesting to note that thefree space model yields consistently higher field strengths. Thatis, it yields lower path loss than the measurements.

Okumura made measurements in various types of terrain, eachyielding a new set of curves. However, the diagrams can only beused as a rough guide since terrain types differ from place toplace, and local variations in the topography as well as in theland usage cannot be accounted for.

Figure 6-4 Okumura’s field measurements displayed as thefield strength 1.5 m above ground as a function of thelogarithmic distance between base and mobile. Curves areshown for different effective base station antenna heights, h1..


EN/LZT 123 3314 R3A – 55 –

Empirical data can be used to improve more elaborate models.In particular, the attenuation as a function of terrain type andland usage, e.g., in 1980, Hata presented a number of semi-empirical formulas based on the field measurements made byOkumura.

As an example the expression for the path loss in urban areas isgiven as:

/S(urban) = 69.55 + 26.16log I - 13.82logK

E + (44.9 - 6.55logK

E)log G

- a(KP),

where

a(KP) = (1.1 log I - 0.7)K

P - (1.56 log I - 0.8)

I = carrier frequency in MHz (150 - 1000 MHz)

KE = the base station antenna height in meters (30 - 200 m)

G = distance in km from the base station (1 - 20 km)

KP = mobile antenna height in meters above ground (1 - 10 m)

Note that this particular formula is strictly valid only for urbanareas in Japanese “quasi-smooth” terrain, but it is still useful forrough estimates of cell coverage. For the same type of terrain,this formula can be adjusted to yield

/S(suburban) = /

S(urban) - 2 [log (I/28)]2 - 5.4

/S�open) = /

S(urban) - 4.78 (log I)2 + 18.33 log I - 40.94

for two other types of land usage.

Another model worth mentioning is the Cost 231-Hata modelwhich can be used when the carrier frequency is in the interval1500-2000 MHz, as in the case of GSM 1900. Here

/S�= 46.3 + 33.9logI - 13.82logK

E�+ (44.9 - 6.55logK

E)logG - a(K

P)

Ericsson has developed simple Okumura-Hata type modelsbased on wave propagation measurements which can be used toestimate the coverage for both GSM 900 and GSM 1800. Here

/S�= A - 13.82logK

E�+ (44.9 - 6.55logK

E)logG - a(K

P)


– 56 – EN/LZT 123 3314 R3A

where

a(KP) = 3.2(log 11.75K

P)2 - 4.97

and

A(900) = 146.8 and A(1800) = 153.8 for urban areas

A(900) = 136.9 and A(1800) = 146.2 for suburban areas

A(900) = 118.3 and A(1800) = 124.3 for open areas


EN/LZT 123 3314 R3A – 57 –

ALGORITHM 9999

The prediction model used by Ericsson, Algorithm 9999 (validbetween 0.2 - 100 km), is based on ideas similar to Hata’s in thatempirical data is used to fit parameters for the attenuationcaused by different land usages (clutters). Regardless of whatmodel is actually used, the accuracy of the prediction dependsheavily on the field measurements performed.

Semi-empirical models are used because radio propagationcannot be measured everywhere. Algorithm 9999 takes intoaccount the land usage by the use of digitized map data as wellas knife-edge diffraction and effects of the earth’s curvature.The algorithm, which is incorporated in the EricssonEngineering Tool (EET) packet calculates the path loss for radiowaves between two spatial coordinates. From the antenna, thepath loss is calculated in all directions for an arbitrary distancefrom the antenna. These predictions give an accuracy of about±5 dB if the parameters in the model have been optimized fromfield measurements of signal strengths. From these predictions,various “arrays” can be calculated. The arrays are based on celldata, e.g., the output power of the TRXs, the antenna gains, andthe frequency allocation to the different TRXs are quite fast tocalculate since the path loss is known and does not changeunless the antenna is moved.

These arrays are very useful and EET can plot them on maps toobtain a graphical display of the predictions. Signal Strengths(SS), Carrier-to-Interferer (C/I), and Carrier-to-Adjacent (C/A)channel ratios and other information can be plotted, thus helpingthe cell planner verify the nominal cell plans and/or improve thesystem design.

In addition to Algorithm 9999, EET also supports other modelsincluding the Cost 231-Walfish-Ikegami model (valid between0.02 - 5 km) which can be used in urban areas because it usesstreet orientation, building heights, building separations, androad widths.


– 58 – EN/LZT 123 3314 R3A

URBAN MODEL

In an urban environment, there are mainly two paths for radiowave propagation:

• Over the roof tops

• Along the street

At a far distance from the site, the first part dominates but nearthe site, the second one dominates. The Urban Model is aconcept of two different wave propagation algorithms:

• Half-screen model

• Recursive microcell model

The half-screen model is used for calculating the propagationabove the roof tops. Obstacles such as buildings and treesbetween transmitter and mobile are placed with a number ofscreens with heights correlated to the heights of the obstacles.The path loss is then calculated by using a multiple knife-edgeapproach.

The recursive microcell model calculates the propagation overopen areas, e.g., along streets. The exact locations of thebuildings are used for defining the propagation paths. The pathloss is calculated by determining the so-called illusory distancebetween transmitter and mobile in a street system.


EN/LZT 123 3314 R3A – 59 –

MICROCELL MODELS

OVERVIEW

A prediction area is centered around a base station, known as thereal source and will normally contain a number of buildings andopen areas, which are defined in EET. This information isentered in what is called vector format.

A source is considered to illuminate a neighboring region andthe path loss resulting at points within this area can be calculatedvia a set of well defined equations. The way in which the pathloss to a point is calculated depends upon whether or not thatpoint has line-of-sight to the real source.

The line-of-sight path loss values use a two slope model andalso take into account antenna masking and height gain factors.To model the path loss values within non line-of-sight regionsthe concept of imaginary, or virtual, sources is introduced. Theseare placed on the limiting bounds of the line of sight region andact in effectively the same way as real sources with theexceptions that only regions that lie in a direction considered tobe forward to the source are affected and points within them arenot masked. If required, further virtual sources may be createdon the line of sight boundaries of a virtual source.

LINE-OF-SITE MODELLING

A real source directly affects the region which has line-of-sightto it. Open areas do not affect the area of line-of-sight.

NON LINE-OF-SITE MODELLING

Regions without line-of-sight to the real source receive radioenergy from it by diffraction around the corners of obstacles.This is modelled by placing a virtual source on the line joining thereal source to the corner of the obstacle and on the far side of thecorner from the real source.

Each of these virtual sources then has an area of influence whichincludes some part of the prediction area lying outside the realsource’s area of influence.


– 60 – EN/LZT 123 3314 R3A

IN-BUILDING MODELLING

When in-building propagation is selected the area of influencefor a source is bounded by the set of non-facing edges of thebuildings that have line of sight to the source. In cases wherebuildings have overlapping non-facing edges preference is givento those nearest the source.

OPEN AREA MODELLING

Open areas are designated regions which have differentpropagation characteristics from street-like regions. They areused to represent large squares, park land, and similar areaswhere vegetation or other clutter affects signal propagation.

To model this, an additional path loss is calculated for the partof the path passing through such a region.

POINTS AFFECTED BY MULTIPLE SOURCES

The creation of virtual sources means that, under certaincircumstances, a point may lie within the area affected by morethan one source. In these cases the signal strength obtained fromeach route is calculated and the strongest signal chosen.

Channel Planning

Chapter 7

This chapter is designed to provide the student with thefundamentals of channel planning.

OBJECTIVES:Upon completion of this chapter the student will be able to:

• Describe the mapping and channel concept

• Define “re-use distance”

• Identify and discuss the various channel plans

• Explain how to avoid co-channel and adjacent channelinterference during channel assignment

7 Channel Planning

EN/LZT 123 3314 R3A – i –

7 Channel Planning

Table of Contents

Topic Page

CELL PLANNING ................................................................................61

TRANSITION REGIONS ......................................................................65

NETWORK COLOR CODE AND BASE STATION COLORCODE (NCC, BCC) ..............................................................................66

7 Channel Planning

EN/LZT 123 3314 R3A – 61 –

CELL PLANNING

The simplest solution to a cell planning problem is to have onecell and use all available carriers in that cell (Figure 7-1).However, such a solution has severe limitations. It is seldomthat coverage can be maintained in the entire area desired. Inaddition, even though the channel utilization may be very high,limited capacity soon becomes a problem due to the limitednumber of carriers available to any operator.

24

Figure 7-1 Example of an area served from one cell by 24carriers

A cellular system is based upon re-use of the same set ofcarriers, which is obtained by dividing the area needing coverageinto many smaller areas (cells) which together form clusters(Figure 7-2).

24

24

2424

f1

f1

f1

f1

Figure 7-2 This is same area as in Figure 7-1 but nowschematically divided into four clusters, each cluster using all(here 24) carriers. The small circles indicate individual cellswhere the frequency f1 is used and a distance between thecorresponding sites, a so-called frequency re-use distance, isindicated by the double arrow


– 62 – EN/LZT 123 3314 R3A

A cluster is a group of cells in which all available carriers havebeen used once (and only once). Since the same carriers are usedin cells in neighboring clusters, interference may become aproblem. The frequency re-use distance (i.e. the distancebetween two sites using the same carrier) must be kept as largeas possible to help prevent interference. At the same time, thedistance must be kept as small as possible from a capacity pointof view. Cellular systems are often interference-limited ratherthan signal-strength-limited.

Re-using the carrier frequencies according to well-proven re-usepatterns (Figure 7-3 and Figure 7-4), neither co-channelinterference nor adjacent channel interference should become aproblem. This is true if the cells have homogenous propagationproperties for the radio waves, and if frequency hopping isimplemented.

The re-use patterns recommended for GSM are the 4/12- and the3/9-patterns. 4/12 means that there are four three-sector sitessupporting twelve cells (Figure 7-3).

B2B3

C1

C2C3

B1

B2B3

C1

C2C3

B1

B2B3

C1

C2C3

A1

A2A3

D3

D2D1

A1

A2A3

D3

D2D1

A1

A2A3

D3

C2C3

B1

B2B3

C1

C2C3

B1

B2B3

C1

C2C3

B1

B2B3

D3

D2D1

A1

A2A3

D3

D2D1

A1

A2A3

D3

D2D1

A1

B2B3

C1

C2C3

B1

B2B3

C1

C2C3

B1

B2B3

C1

C2C3

A1

A2A3

D3

D2D1

A1

A2A3

D3

D2D1

A1

A2A3

D3

Figure 7-3 4/12 re-use pattern (Note: Observe the positions ofthe frequency groups D1 and D3)

7 Channel Planning

EN/LZT 123 3314 R3A – 63 –

The re-use pattern in Figure 7-3 is compatible with the planningcriterion C/I>12 dB. A shorter re-use distance giving a smallerC/I-ratio, is used in the 3/9-pattern (Figure 7-4).

C1

C2C3

A1

A2A3

B1

B2B3

C1

C2C3

A1

A2A3

A2A3

B1

B2B3

C1

C2C3

A1

A2A3

B1

B2B3

C1

C1

C2C3

A1

A2A3

B1

B2B3

C1

C2C3

A1

A2A3

A2A3

B1

B2B3

C1

C2C3

A1

A2A3

B1

B2B3

C1

C1

C2C3

A1

A2A3

B1

B2B3

C1

C2C3

A1

A2A3

A2A3

B1

B2B3

C1

C2C3

A1

A2A3

B1

B2B3

C1


This re-use pattern (Figure 7-4), which has a higher channelutilization (since the carriers are distributed among nine cellsrather than twelve) is recommended only if frequency hopping isimplemented. That is, it is compatible with the planningcriterion C/I >9 dB. In addition, since C/A ≈ 0 close to some ofthe cell borders (Figure 7-5) special care must be taken. Otherre-use patterns such as the 7/21, with much higher re-usedistances, must be used for systems which are more sensitive tointerference, e.g., analogue mobile telephone systems.

As an example, suppose that one operator has been given 5 MHzof bandwidth and distributes the carriers over nine cells, it canlook like Figure 7-5.


– 64 – EN/LZT 123 3314 R3A

Channelgroups A1 B1 C1 A2 B2 C2 A3 B3 C3

RF 512 513 514 515 516 517 518 519 520

Channels 521 522 523 524 525 526 527 528 529

530 531 532 533 534 535

Figure 7-5 24 frequencies in the 3/9 cell plan. The AbsoluteRadio Frequency numbers (ARF) given here correspond to thefrequency interval 1710 - 1715 MHz in GSM 1800. Note thatthe adjacent cells A1 and C3 also have adjacent frequencies200 kHz apart (f(ARFCN) = 17102 + 0.2(ARFCN - 512))

From this example, it can be seen that channels used in the samecell of a 3/9 (4/12) cell plan are always nine (twelve) RFchannels apart. This is beneficial regarding the properties of thecombiners. A filter combiner requires 600 kHz and a hybridcombiner 400 kHz channel separation for GSM 900.

As mentioned above, the 3/9 re-use pattern has adjacentchannels in some pairs of adjacent cells (A1, C3) which calls forspecial attention when using this re-use pattern.

Hence, a nominal cellplan consists of a hexagonal pattern ofcells where the sites typically are distributed equidistant andwhere ideally they can be placed according to a uniform pattern.However, this is seldom the case because cells often vary insize. Therefore, real nominal cellplans must be verified bymeans of predictions or radio measurements, in order to ensurethat interference does not become a problem.

7 Channel Planning

EN/LZT 123 3314 R3A – 65 –

TRANSITION REGIONS

A uniform re-use pattern implies a constant traffic density overthe network’s coverage area. In practice, however, traffic densityvaries considerably over the area (and during the day). Thismeans it is common that cells of different sizes are used indifferent parts of a system, small cells in high-traffic areas(normally urban) and large cells in areas with lower traffic.

Figure 7-6 shows a case where cells have assorted sizes in acoverage area. This poses special problems in channel planning,as the re-use distance will vary for different cell sizes. To avoidhaving a smaller cell size that has half the re-use distance of thelarger cell size interfering in the larger cell, other RF channelsmust be used in these cells. We need a buffer zone where thesame RF channels are not used in the smaller and larger cells,respectively. This is sometimes a costly but unavoidablearrangement. Another possibility that parallels this type ofvaried coverage is in the use of overlaid and underlaid cells or ahierarchical cell structure. There, again, exists the need to haveseparate channel plans.

Figure 7-6 A cellular network


– 66 – EN/LZT 123 3314 R3A

NETWORK COLOR CODE AND BASE STATION COLORCODE (NCC, BCC)

The Base Station Identity Code (BSIC) is composed of twoentities:

• Network Color Code (NCC)

• BTS Color Code (BCC)

NCC is used to discriminate between cells in two differentPLMNs using the same frequency (Figure 7-7). These PLMNswith the same frequencies are always in different countries sincethe PLMNs in one country use different carrier frequencies. Theoperators in different countries must decide between themselveswhat the NCC assignment will be. The operators may use morethan one NCC value as long as they only use their agreed valuein the border areas.

Country A Country B

NCC = 1 NCC = 2

f1f1

Figure 7-7 The use of NCC in two countries.

BCC is used for protection against co-channel interferencewithin the PLMN. The MS reports the BCC value so that theBSC can distinguish among different cells transmitting on thesame frequency. If frequency re-use clusters are used, it isrecommended that all BTSs in a cluster use the same BCC, andthat an adjacent cluster use another BCC (Figure 7-8). If clustersare not used, great care must be taken when planning BCC.

Surveys

Chapter 8

This chapter is designed to provide the student with an overviewof radio network survey of a cellular network as well as someradio measurements.


• Explain what a radio network survey is and what to considerduring a survey

• Describe three different types of radio measurements

8 Surveys

EN/LZT 123 3314 R3A – i –

8 Surveys

Table of Contents

Topic Page

INTRODUCTION..................................................................................69

SITE REQUIREMENTS........................................................................70

RADIO NETWORK SURVEY...............................................................72

BASIC CONSIDERATIONS......................................................................................... 72

POSITION RELATIVE TO NOMINAL GRID ................................................................ 72

SPACE FOR ANTENNAS............................................................................................ 72

ANTENNA SEPARATIONS ......................................................................................... 73

NEARBY OBSTACLES................................................................................................ 73

SPACE FOR RADIO EQUIPMENT.............................................................................. 74

POWER SUPPLY/BATTERY BACKUP....................................................................... 74

TRANSMISSION LINK................................................................................................. 74

SERVICE AREA STUDY ............................................................................................. 74

CONTRACT WITH THE OWNER................................................................................ 74

RADIO MEASUREMENTS...................................................................75

PATH LOSS PARAMETERS ....................................................................................... 75

TIME DISPERSION ..................................................................................................... 75

INTERFERING TRANSMITTERS................................................................................ 76

8 Surveys

EN/LZT 123 3314 R3A – 69 –

INTRODUCTION

The cell planning process results in a cell plan with nominal sitepositions. If the operator has access to existing locations, it isnecessary to adapt the cell plan according to these locations.

For this reason, it is important that the cell planner has a basicknowledge of the locations that can be used.

The on-site cell planning work that takes place is called the“Radio Network Survey”. This is described in the followingsection. A more detailed survey is performed on the base stationsites. This is called the “site investigation” and is not discussedin this course.


– 70 – EN/LZT 123 3314 R3A

SITE REQUIREMENTS

The proposed network design shows only approximate sitelocations. The exact site position depends on the possibilities toconstruct a site on the suggested location.

• Different permits are usually necessary, e.g. a planningpermit from the local council planning committee. Masts ortowers almost always require planning permits and in manycases they are subject to permits from civil aviation ormilitary authorities (i.e. obstruction lighting may be needed).

• Permission to use the site or a lease contract must be agreedupon with the owner of the site.

Besides the need for the permits, the following must also betaken into account:

• Access roads - The site must be accessible to installationpersonnel and heavy trucks and if there is no road leading tothe site, a helicopter might be needed for material transportsand for mast or tower installation.

• Material transport and storage - The site must have an areasuitable for efficient unloading and handling of goods.

• Space requirements - For an outdoor site it is necessary thatthe ground area is large enough for the radio base station andtower or mast foundation. Power cables must be installedand a mains power source must be found in the vicinity ofthe site if mains power is not available at the site. For anindoor site, the RBS equipment room must fulfill a numberof requirements concerning mains power connection such asgrounding, power outlet, and space for transport networkinterface products.

• Antenna support structures - These must be provided. Theycan consist of several short pipes on a roof, a guyed mast, ora self-supporting tower. The term “tower” usually refers to aself-supported structure, while the term “mast” refers to astructure supported with guy wires.

• AC mains supply - The AC mains supply requirements aredescribed in the technical data given for each RBS in chapter10, “Implementation”.

8 Surveys

EN/LZT 123 3314 R3A – 71 –

• Transmission access - A number of Pulse Coded Modulation(PCM) transmission lines are needed. Two types oftransmission network standards may occur. The first case is2 Mbit/s PCM with 75 ohm unbalanced or 120 ohmbalanced lines, the second case is 1.5 Mbit/s PCM using 100ohm balanced lines.

• Antenna feeder routes - Indoors, the antenna feeder pathsmust have proper cable support facilities, preferably a cableladder. The antenna feeder may also be placed in availablecable chutes inside the building. Outdoors, feeder cablepaths from an antenna supporting structure fall into twocategories. Cables can be installed on cable ladders aboveground from the antenna or through underground cableducts.


– 72 – EN/LZT 123 3314 R3A

RADIO NETWORK SURVEY

BASIC CONSIDERATIONS

It is likely that the system operator has a number of alternativebuildings which may be used in the cellular network planningphase. One reason for this is to reduce the initial cost.

The following aspects of site selection must be studied:

• Position relative to nominal grid

• Space for antennas

• Antenna separations

• Nearby obstacles

• Space for radio equipment

• Power supply/battery backup

• Transmission link

• Service area study

• Contract with the owner

POSITION RELATIVE TO NOMINAL GRID

The initial study for a cell system often results in a theoreticalcell pattern with nominal positions for the site locations. Theexisting buildings must then be adapted in such a way that thereal positions are established and replace the nominal positions.The visit to the site is to ensure the exact location(address/coordinates and ground level). It is also possible formore than one existing site to be used for a specific nominalposition.

SPACE FOR ANTENNAS

The radio propagation predictions provide an indication on whattype of antennas can be used on the base station and in whatdirection the antennas should be oriented.

The predicted antenna height should be used as a guideline whenthe on-site study starts. If space can be found within a maximumdeviation of 15% from the predicted height the originalpredictions can be used with sufficient accuracy.

8 Surveys

EN/LZT 123 3314 R3A – 73 –

If it is possible to install the antennas at a higher position thanthe predicted position, the operator must ensure that there is norisk for co-channel interference. If the antennas are to beinstalled at a lower position than predicted, new predictionsmust be carried out based on this position.

It is not necessary that all antennas in one particular cell havethe same height or direction. That is, it is possible to have cellson the same base station with different antenna heights. This canbe the case if space is limited in some directions. There are alsocell planning reasons for placing antennas at different heights.This includes coverage, isolation, diversity and/or interference

Note: Some of these considerations are discussed in the nextsection.

ANTENNA SEPARATIONS

There are two reasons for antennas to be separated from eachother and from other antenna systems:

• To achieve space diversity

• To achieve isolation

The horizontal separation distance to obtain sufficient spacediversity between antennas is 12-18 λ or 4-6 m for GSM 900and 2-3 m for GSM 1800/1900. Typical values of separationdistances between antennas to obtain sufficient isolation(normally 30 dB) are 0.4 m (horizontal) and 0.2 m (vertical) forGSM 900.

NEARBY OBSTACLES

One very important part in the Radio Network Survey is toclassify the close surroundings with respect to influence on radiopropagation. In traditional point-to-point communicationnetworks, a line-of-sight path is required. A planning criterion isto have the first fresnel zone free from obstacles. (NOTE: Thefresnel zone is the area in open space that must be practicallyfree of obstructions for a microwave radio path to functionproperly, some degree of fresnel consideration is required in theimmediate vicinity of the microwave radio RF envelope/field.)

It is not possible to follow this guideline because the pathbetween the base and the mobile subscriber is normally not line-of-sight. In city areas, one cell planning criterion is to providemargins for these types of obstacles.


– 74 – EN/LZT 123 3314 R3A

If optimal coverage is required, it is necessary to have theantennas free for the nearest 50-100 m. The first fresnel zone isapproximately five meters at this distance (for 900 MHz). Thismeans the lower part of the antenna system has to be five metersabove the surroundings.

SPACE FOR RADIO EQUIPMENT

Radio equipment should be placed as close as possible to theantennas in order to reduce the feeder loss and the cost forfeeders. However, if these disadvantages can be accepted, otherlocations for the equipment can be considered. In addition,sufficient space should be allotted for future expansions.

The radio network survey includes a brief study with respect tothis matter. A more detailed analysis takes place when thelocation is chosen to be included in the cellular network.

POWER SUPPLY/BATTERY BACKUP

The equipment power supply must be estimated and thepossibility of obtaining this power must be checked. Space forbattery back-up may be required.

TRANSMISSION LINK

The base station must be physically connected to the BSC. Thiscan be carried out via radio link, fiber cable, or copper cable.Detailed transmission planning is not included in this course.

SERVICE AREA STUDY

During the network survey it is important to study the intendedservice areas from the actual and alternate base station locations.Coverage predictions must be checked with respect to criticalareas.

CONTRACT WITH THE OWNER

The necessary legal documentation must exist between the landowner and the proposed site user, e.g. a contract for site leasing.Even though cost is a major consideration in the site acquisitionprocess, cost is not discussed as a factor in this course.

8 Surveys

EN/LZT 123 3314 R3A – 75 –

RADIO MEASUREMENTS

PATH LOSS PARAMETERS

A radio survey involves installation of a transportable testtransmitter somewhere in the area where the base station is to beinstalled. Using a specially equipped vehicle, signal strength canbe measured. For this purpose, Ericsson has designed acomputerized measurement system. A locating unit, a measuringreceiver with antenna, a control and processing unit, and a taperecorder are among the equipment contained in the unit. Signallevel can be measured on a number of channels and, for eachchannel, samples are taken at an adjustable speed. Normally,samples are taken several times per wavelength traveled. Thedata is pre-processed before it is stored on either the hard driveor a diskette and presented off-line after the survey. Results canbe presented with respect to median value, standard deviation,and number of “measuring squares” along the test routes. Therecorded files can be imported into EET and displayed on themap. The residual values (i.e. the difference between theprediction and the measurement, can also be displayed. If thereis a difference, the path loss parameters in the prediction modelcan be adjusted according to the measurements.

TIME DISPERSION

Measurements must be performed to verify the time dispersionpredictions. In addition, if there are quality problems, timedispersion measurements must be taken to verify that timedispersion is actually causing the poor quality.

The equipment used for time dispersion measurements consistsof a transmitter and a receiver (Figure 8-1). The transmittersends a short pulse, the signal is received and the pulse responseis evaluated in a controller (Figure 8-2). In this way, the timedelay and the carrier to reflection ratio can be found.


– 76 – EN/LZT 123 3314 R3A

Filter

Amplifier

Spectrumanalyzer

Controller unit

RX Antenna

Site equipment

Filter

Pulsegenerator

Van equipment

TXAntenna

Figure 8-1 Time dispersion measurement equipment

Power window 16 µs

Figure 8-2 Impulse response

INTERFERING TRANSMITTERS

For sites where a number of other radio transmitters are co-located, Ericsson recommends that radio spectrummeasurements and a subsequent interference analysis beperformed. Ericsson has developed special equipment andmethods for this purpose. These include a computer controlledspectrum analyzer and computer programs for calculatinginterference levels at different frequencies. The end result of aradio spectrum measurement is to accept the site from aninterference point of view, to accept it with reservations, or toreject the site and find another one.

Design Projects

Chapter 9

This chapter is designed to provide the student with case studyexercises.


• Do system balance calculations and explain the necessity ofperforming it

• Calculate traffic and channel needs for a given area

• Make propagation calculations

• Construct a theoretical cell plan

9 Design Projects

EN/LZT 123 3314 R3A – i –

9 Design Projects

Table of Contents

Topic Page

GSM 1900 STUDY CASE ....................................................................77

DESCRIPTION OF THE AREA.................................................................................... 77

SYSTEM OPERATOR REQUIREMENTS ................................................................... 77

SYSTEM DESCRIPTION............................................................................................. 77

PROPAGATION MODELS........................................................................................... 77

TRAFFIC ESTIMATION............................................................................................... 78

TASK............................................................................................................................ 79

GSM 1800 STUDY CASE ....................................................................80






TASK............................................................................................................................ 82

GSM 900 STUDY CASE ......................................................................83






TASK............................................................................................................................ 85

9 Design Projects

EN/LZT 123 3314 R3A – 77 –

GSM 1900 STUDY CASE

DESCRIPTION OF THE AREA

The area is situated north of Dallas, Texas and covered by themaps Garland and Plano (both suburbs of Dallas).

The built–up areas on the map incorporate several largesuburban residential areas. However, the maps are not up-to-date. Making a visit to the area, one finds that a large number ofindustrial companies have been built in north Richardson. CollinCreek Mall is situated in South Plano.

A classification of the areas to be covered is described inFigure 9-1 through Figure 9-3.

SYSTEM OPERATOR REQUIREMENTS

The operator requires support of 1-watt cellular phones in theentire system. The antenna gain in the mobile stations isassumed to be 0 dBi, while the feeder loss of the mobile is 0 dB.

Antenna effective heights of 30 meters are used everywhere inthis system for simplicity.

The operator requires a grade of service of 2%.

The signal strength requirement is –90 dBm or higher.

SYSTEM DESCRIPTION

The system can be planned according to a 4/12 re-use pattern.The system operator has access to 24 frequencies. Combiner lossis assumed to be 3 dB regardless of the number of channelsused. Feeder loss is assumed to be 3.5 dB at 30 meters antennaheight. Output and received power is presented in dBm.

PROPAGATION MODELS

Ericsson uses a modified Cost-231-Hata model for thepropagation predictions. In this case though, we have no accessto computers. Therefore, the empirical formulas of Hata will beused to manually calculate/predict radio wave propagation.


– 78 – EN/LZT 123 3314 R3A

TRAFFIC ESTIMATION

Statistical studies of the area have shown that the traffic persubscriber is as follows:

• Average call duration: 120 seconds

• Average number of calls per subscriber during a busy hour:1.0

• Subscriber distribution is represented by Figure 9-1 throughFigure 9-3

Figure 9-1 Offered traffic for Open/Rural areas

Figure 9-2 Offered traffic for Suburban areas

9 Design Projects

EN/LZT 123 3314 R3A – 79 –

&ROOLQ�&UHHN�0DOO

Figure 9-3 Offered traffic for a specific location

TASK

Present a theoretical cell plan that fulfills the system operatorrequirements.

The result will be presented as a map with site locations andcoverage for each cell (represented by hexagons of suitable

size). The area covered by a hexagon is approximately 2.6 x 52

where 5 is the cell radius.

A list of the following parameters should be presented for atleast one cell per area:

• Site location

• Antenna directions

• Antenna gain

• Power level of the TRU

• ERP

• Estimated Traffic

• Frequency allocations


– 80 – EN/LZT 123 3314 R3A

GSM 1800 STUDY CASE


The area to be planned is Stockholm, Sweden with surroundingsuburbs covered by the map “Stockholm 10I NO”.

The terrain is somewhat smooth with two major highways: E3and E4. Stockholm and the suburban centers have buildings ofstone and concrete. The buildings are four to seven stories high.

The built-up areas on the map incorporates a large number ofindustrial companies. Globen Sports Arena is situated south ofStockholm.



The operator requires support of 1 W cellular phones in theentire system. The antenna gain in the mobile stations isassumed to be 0 dBi while the feeder loss of the mobile is 0 dB.




SYSTEM DESCRIPTION

The system operator has access to 48 frequencies. The systemcan be planned according to a 3/9 re-use pattern. Combiner lossis assumed to be 3 dB regardless of the number of channelsused. Feeder loss is assumed to be 3.5 dB at 30 meters antennaheight. Output and received power is presented in dBm.

PROPAGATION MODELS

Ericsson uses a modified Okumura/Hata model for thepropagation predictions. In this case though, we have no accessto computers. Therefore, the empirical formulas of Hata will beused to manually calculate/predict radio wave propagation.Remember to use antenna gain values specified in dBi.

9 Design Projects

EN/LZT 123 3314 R3A – 81 –

TRAFFIC ESTIMATION


• Average call duration: 120 seconds

• Average number of calls per subscriber during a busy hour:1.0

• Subscriber distribution is represented by Figure 9-4�throughFigure 9-6

20

10


500



– 82 – EN/LZT 123 3314 R3A

Globen

1000

500


TASK






• Site location


• Antenna gain

• Output power of the TRX/TRU

• ERP



9 Design Projects

EN/LZT 123 3314 R3A – 83 –

GSM 900 STUDY CASE


The area is situated north of Stockholm and covered by the map“Uppsala 11I SV”.

The terrain is somewhat smooth with two major highways, E4and E18. There are a few suburbs in Stockholm with concretehouses in the central parts within the area. In the lower left partthere is a military exercise area.

The built-up areas on the map incorporate a large number ofindustrial companies and several large suburban residentialareas. Arlanda International Airport is also situated in this area.



The operator requires support of Class 4 cellular phones in theentire system. The antenna gain in the mobile stations isassumed to be 0 dBi while the feeder loss of the mobile is 0 dB.




SYSTEM DESCRIPTION

The system operator has access to 27 frequencies. The systemcan be planned according to a 3/9 re-use pattern. Combiner lossis assumed to be 3 dB regardless of the number of channelsused. Feeder loss is assumed to be 3.5 dB at 30 meters antennaheight. Output and received power is presented in dBm.

PROPAGATION MODELS

Ericsson uses a modified Okumura/Hata model for thepropagation predictions. In this case though, we have no accessto computers. Therefore, the empirical formulas of Hata will beused to manually calculate/predict radio wave propagation.


– 84 – EN/LZT 123 3314 R3A

TRAFFIC ESTIMATION


• Average call duration: 120 seconds.

• Average number of calls per subscriber during a busy hour:1.0.

• Subscriber distribution is represented by Figure 9-7 throughFigure 9-9�


40

50

60


9 Design Projects

EN/LZT 123 3314 R3A – 85 –

Arlanda International airport

100

200


TASK






• Site location


• Antenna gain

• Output power of the TRX/TRU

• ERP



Implementation

Chapter 10

This chapter is designed to provide the student with an overviewof the site equipment; in particular, the radio base stations andthe antenna system.


• Recognize the base stations in the Ericsson RBS 200 and2000 series

• Briefly describe Maxite

• Describe the two combiners used in the RBSs

• Describe how antenna gain is achieved

• Describe some basic antenna types

• Describe the CDUs used

• Describe diversity

• Explain BSC capacity

• Describe different transmission methods

10 Implementation

EN/LZT 123 3314 R3A – i –

10 Implementation

Table of Contents

Topic Page

GENERAL............................................................................................87

THE RBS 200 SERIES.........................................................................88

RBS 200 ARCHITECTURE.......................................................................................... 88

THE RBS 2000 SERIES.......................................................................95

GENERAL.................................................................................................................... 95

EASY TO INSTALL AND MAINTAIN ........................................................................... 95

MAXITE..............................................................................................106

MAXITE APPLICATIONS........................................................................................... 107

COMBINERS......................................................................................109

FILTER COMBINER .................................................................................................. 109

HYBRID COMBINER ................................................................................................. 111

COMBINER AND DISTRIBUTING UNIT (CDU) ........................................................ 112

SENSITIVITY .....................................................................................116

MICRO BASE STATION............................................................................................ 116

MOBILE STATION..................................................................................................... 116

ANTENNAS........................................................................................118

BASIC ANTENNA TYPES ......................................................................................... 118

OMNIDIRECTIONAL ANTENNAS............................................................................. 118

UNIDIRECTIONAL ANTENNAS ................................................................................ 119

SPECIAL ANTENNAS ............................................................................................... 119

MULTI ANTENNA SYSTEMS.................................................................................... 119

DIVERSITY ........................................................................................120

SPACE DIVERSITY................................................................................................... 120

POLARIZATION DIVERSITY..................................................................................... 121

ANTENNA TILT.......................................................................................................... 124

10 Implementation

EN/LZT 123 3314 R3A – 87 –

GENERAL

The Base Station System (BSS) consists of Base StationControllers (BSCs) with a number of base stations connected toeach. The BSS is mainly responsible for all radio relatedfunctions in the system. In the GSM specifications, thedenotation BTS (Base Transceiver Station) is used for the basestation.

The Ericsson implementation of the GSM BTS is the RadioBase Station (RBS). This comes in two series: 200 and 2000.

The distinction between the Ericsson RBS and the GSM BTS isthat while the BTS includes all equipment needed to maintainthe traffic in one cell, the RBS includes all radio andtransmission interface equipment needed on-site to provide radiotransmission for one or several cells.

The first generation of RBSs belonging to the RBS 200 familyoffers indoor (RBS 200) and outdoor (RBS 203/204) versionsfor GSM 900 as well as products for GSM 1800 (RBS 205/206).

The RBS 2000 family is Ericsson’s second generation of radiobase stations. The series provides products for both indoor andoutdoor installations and is available for GSM 900, GSM 1800,and GSM 1900.


– 88 – EN/LZT 123 3314 R3A

THE RBS 200 SERIES

RBS 200 ARCHITECTURE

RBS 200 includes all radio equipment needed on-site and iscomprised of the following major functional units (Figure 10-1):

• Transmission Radio Interface (TRI)

• Transceiver System (TRS), including Transceiver Group(s)(TG)

TRI is a switch which enables a flexible connection to be madebetween the BSC and TRS.

TRS is the RBS 200 (including hardware and software) butwithout TRI.

The TG encompasses all radio equipment connected to one bussystem.

BSC Base Station ControllerTRI Transmission Radio InterfaceTG Transceiver GroupTRS Transceiver SystemTRX Transceiver

BSC

TRI

TG

TG

TRX TRX TRX

TRS

RBS 200

Figure 10-1 RBS 200 block diagram

A common feature of base stations in the RBS 200 series is thatthey are transceiver-oriented. That is, they are designed to haveminimal equipment common to several transceivers. The mostsignificant benefit of this approach is the redundancy that can beachieved. A hardware fault at the radio site only affects a singletransceiver. All critical equipment common to the cell or the sitecan be duplicated for reasons of redundancy.

10 Implementation

EN/LZT 123 3314 R3A – 89 –

RBS 200/205

RBS 200 is the base station for GSM 900 indoor applications(Figure 10-2), and the RBS 205 is the base station for GSM1800 indoor applications. Omni or three-sector sites can be setup. Indoor installations offer large configurations of up to 16transceivers (TRXs) per cell.

The base stations are completely remote controlled by softwarestored in RAM to facilitate upgrading. Remote controlled outputpower and auto-tuned transmitter combiners (TXCMBs) enablesystem retuning without on-site visits.

Technical Data

Number of TRXsper cabinet: 4

Dimensions(D x W x H): 405 x 602 x 1972 mm

Weight: 250 kg

Frequency band: RBS 200 RX 890 to 915 MHzTX 935 to 960 MHz

RBS 205 RX 1710 to 1785 MHzTX 1805 to 1880 MHz

TX output power fromPower Amplifier (PA): RBS 200 45 W per channel

RBS 205 30 W per channel

RX sensitivity: RBS 200 < -105 dBmRBS 205 < -104 dBm

Diversity gain: up to 5 dB

Power consumption: < 1400 W per cabinet

Operating temperature: + 5° to + 40° C

Input power: 230 VAC, + 24 VDC or - 48 VDC

Battery backup time: selection from 15 minutes to 8 hours


– 90 – EN/LZT 123 3314 R3A

Figure 10-2 A fully equipped RBS 200 for GSM 900 indoorapplications

10 Implementation

EN/LZT 123 3314 R3A – 91 –

RBS 203/204

These base station versions from the RBS 200 series aredesigned for outdoor applications for GSM 900 (Figure 10-3 andFigure 10-4). Just like in the RBS 200/205 case, the basestations are completely remote controlled by software stored inRAM to facilitate upgrading. Remote controlled output powerand auto-tuned TXCMBs enable system retuning without on-sitevisits.

The RBS 203 is designed for omnicell configurations with up totwo TRXs. Omni or three-sector sites can be set up in one RBS204 cabinet.

Figure 10-3 A fully equipped RBS 203 for GSM 900 outdoorapplications (omnicell configurations with up to two TRXs)


– 92 – EN/LZT 123 3314 R3A

Technical Data for RBS 203



Weight: 400 kg without batteries460 kg with batteries

Frequency band: RX 890 to 915 MHzTX 935 to 960 MHz

TX output power fromPower Amplifier (PA): 45 W per channel

RX sensitivity: < -105 dBm


Power consumption: 2400 W per cabinet

Operating temperature: - 33° to + 40° C

Input power: 230 VAC Nominal

Battery backup time: 55 minutes at full operationadditional 45 minutes for essentialequipment

10 Implementation

EN/LZT 123 3314 R3A – 93 –

Figure 10-4 A fully equipped RBS 204 for GSM 900 outdoorapplications (omni- or three-sector sites)

Technical Data for RBS 204



Weight: 750 kg without batteries 900 kg with batteries

Frequency band: RX 890 to 915 MHzTX 935 to 960 MHz

TX output power fromPower Amplifier (PA): 45 W per channel

RX sensitivity: < -105 dBm


Power consumption: < 3100 W per cabinet



– 94 – EN/LZT 123 3314 R3A

Input power: three single-phase 3 * 230 VAC

Battery backup time: 60 minutes at full operationadditional 1.5 hours for essentialequipment

10 Implementation

EN/LZT 123 3314 R3A – 95 –

THE RBS 2000 SERIES

GENERAL

The RBS 2000 product family for GSM 900, GSM 1800, andGSM 1900 is specially designed to offer rapid and cost-effectiveroll-outs and low total life-cycle costs.

Superior radio performance with high output power and goodreceiver sensitivity render good coverage and high capacity. Thehigh coverage means money saved with less radio base stationsin a given area if the system is not capacity limited.

EASY TO INSTALL AND MAINTAIN

The RBS 2000 is especially well suited for rapid roll-outs, as itallows simple installation with on-site testing andcommissioning in just one hour. This is easily accomplishedbecause the cabinets have already been pre-assembled andsoftware has been downloaded and tested at the factory prior toshipment. Reduced size is a result of innovative new technologywhich has drastically reduced the number of systemcomponents, meaning minimal maintenance costs.

The RBS 2000 is designed to meet extremely high demands onreliability. All critical system elements are duplicated to ensurehigh Mean Time Between Failures (MTBF). This minimizescostly site visits.

The modular design makes repairs very quick and easy. Should ahardware failure occur, an indicator shows the failing unit. Thetime it takes to replace the unit at the site is less than 15minutes.


– 96 – EN/LZT 123 3314 R3A

RBS 2101

The RBS 2101 is an outdoor self-contained or indoor cabinetwith up to two transceivers (Figure 10-5). It can be configuredfor omni or one sector cells.

Figure 10-5 A fully equipped RBS 2101 (up to two transceivers)

10 Implementation

EN/LZT 123 3314 R3A – 97 –

Technical data for RBS 2101

Frequency band: GSM 900, GSM 1800, or GSM1900TX: 935-960, 1805-1880 or1930-1990 MHzRX: 890-915, 1710-1785 or1850-1910 MHz

Number of transceivers: 1-2

Number of sectors: 1 - 3 (with more than one cabinet)

Transmission interface: 1.5 Mbit/s (T1), 2 Mbit/s (E1)

Dimensions(H x W x D): 1267 x 705 x 450 mm

Weight - with heatexchanger: 193 kg

Power into antenna feeder: 28 W / 44.5 dBm (GSM 900)(recommended cell 22 W / 43.5 dBm (GSM 1800 /GSM planning value) 1900)

Receiver sensitivity: a-107 dBm (GSM 900)a-109 dBm with TMA (GSM1900) a-110 dBm with TMA(GSM 1800)

Power supply: 100-127/200/250 VAC50/60 Hz

Battery backup time: min. 3 minutes

Operating temperaturewith heat exchanger: - 33° to + 40° Cwith air-conditioner: - 33° to + 55° C

Weather proofing: Min level IP55 in IEC 529


– 98 – EN/LZT 123 3314 R3A

RBS 2102

The RBS 2102 is an outdoor self contained cabinet with up tosix transceivers (Figure 10-6). It can be configured for omni orthree sector cells.

Figure 10-6 A fully equipped RBS 2102 (up to six transceivers)

10 Implementation

EN/LZT 123 3314 R3A – 99 –


Frequency band: GSM 900, GSM 1800 or GSM 1900TX: 935-960, 1805-1880 or1930-1990 MHzRX: 890-915, 1710-1785 or1850-1910 MHz


Number of sectors: 1-3



Weight: 605 kg


Receiver sensitivity: a-107 dBm (GSM 900)a-109 dBm with TMA (GSM1900)a-110 dBm with TMA (GSM1800)

Power supply: 100-127/200-250 VAC50/60 Hz

Battery backup: Minimum 1 hour




– 100 – EN/LZT 123 3314 R3A

RBS 2103

The RBS 2103 is an outdoor self contained cabinet with up tosix transceivers (Figure 10-7). It can be configured for omni orthree sector cells and has a smaller footprint than the RBS 2102.

P000703B

PSU

1

ACCU 1

Transportmodule

PSU

2

PSU

3

PSU

4

ECU

1

DXU

1

CDU1

CDU2

CDU3

TRU

1

TRU

2

TRU

3

TRU

4

TRU

5

TRU

6

IDM 1

Temp 1

Hum 1

Temp 3

Temp 2

P000703D

Transportmodule

Mountingbase

Space for 2DC/DC(optional)

Space forsplice tray(optional)

Fan unit and IDM

Serviceoutlet

Earthbar

EACU

TMfuse panel

Mains inputfilter

Base frame

PSU

BFU

Fan tray

Batteries ACCU

Batteries

TRU

CDU

ECU

DXU

Cabinet

BFU

Bus bar

P000703D

Transportmodule

Mountingbase

Space for 2DC/DC(optional)

Space forsplice tray(optional)

Fan unit and IDM

Serviceoutlet

Earthbar

EACU

TMfuse panel

Mains inputfilter

Base frame

PSU

BFU

Fan tray

Batteries ACCU

Batteries

TRU

CDU

ECU

DXU

Cabinet

BFU

Bus bar

Figure 10-7 A fully equipped RBS 2103 (for GSM 900applications only)

10 Implementation

EN/LZT 123 3314 R3A – 101 –


Frequency band: GSM 900TX: 935-960 MHzRX: 890-915 MHz






Power into antenna feeder: 22 W / 43.5 dBm(guarantee minimal value)

Receiver sensitivity: a-105 dBm

Power supply: 188 - 275 VAC45 - 65 Hz

Battery backup time: 45 minutes




– 102 – EN/LZT 123 3314 R3A

RBS 2202

The RBS 2202 is an indoor cabinet with up to six transceivers(Figure 10-8). It can be configured for omni or three sector cells.

Figure 10-8 A fully equipped RBS 2202 for indoor applications

10 Implementation

EN/LZT 123 3314 R3A – 103 –









Receiver sensitivity: a-107 dBm (GSM 900)a-109 dBm with TMA (GSM1900)a-110 dBm with TMA (GSM1800)


Battery backup: optional

Operating temperature: + 5° to + 40° C



– 104 – EN/LZT 123 3314 R3A

RBS 2301

The RBS 2301 micro base station cuts the site costs by up to 70percent (Figure 10-9). The micro base station is the result ofinnovative new technology, e.g. the cooling system has no fansor moving parts which makes the base station totally silent.Many technical functions have been reduced to ApplicationSpecific Integrated Circuits (ASICs) which greatly reduces theheat generation inside the cabinet.

Figure 10-9 A fully equipped RBS 2301 for micro cellapplications

10 Implementation

EN/LZT 123 3314 R3A – 105 –




Number of sectors: 1-3 with more than one cabinet

Transmission interface: 1.5 Mbit/s (T1), 2 Mbit/s (E1) LTE(4-wire) or CSU Long haul (4-wire)


Weight: < 30 kg

Volume: < 33 liter

Power into antenna feeder: 2W

Receiver sensitivity: -107 dBm (GSM 900)-106 dBm (GSM 1800/GSM 1900)


Battery backup time: 3 minutes



– 106 – EN/LZT 123 3314 R3A

MAXITE

Today, Ericsson has one of the smallest and most efficientmicrobase stations for micro and indoor cells available on themarket. RBS 2301 (mentioned previously) was the first productin the micro concept from Ericsson. Now the next step, usingmicrobase stations for coverage in macro cells, is Maxite.

Maxite combines the smallest GSM two transceiver RBS (the2301) with an active antenna system (with distributed poweramplifiers for transmitted signals) and a battery-backed lowoutput power supply. This gives a micro RBS solution with amacro cell coverage that can virtually be mounted anywhere.

There are obvious advantages in using distributed amplifiersintegrated in the antenna patches. Many of them are highlypositive with respect to environmental impact. Two advantagesare the total power consumption is reduced by more than 65%compared to traditional macro sites and the amount of materialused is also significantly reduced.

Technical data for Maxite

Frequency band: GSM 1800 or 1900

TX: 1805-1880 or1930-1990 MHzRX: 1710-1785 or1850-1910 MHz

Number of transceivers: 2

Dimensions Radio cabinet= 535x408x160mm(H x W x D): Power cabinet= 535x408x210mm

Antenna unit (500 W EIRP)= 1350x250x150mmAntenna unit (1250 W EIRP)=2500x600x150mm

Weight: Radio cabinet= 33kgPower cabinet= 40kgAntenna unit (500 W EIRP)= 20kgAntenna unit (1250 W EIRP)= 50kg

10 Implementation

EN/LZT 123 3314 R3A – 107 –

Power supply: Radio cabinet= 230 VACPower cabinet= 230 VACAntenna unit (500 W EIRP)= -48 VAC (from Power cabinet)Antenna unit (1250 W EIRP)= -48 VAC(from Power cabinet)

Maximum output power from the Power cabinet is 500 Watts.

MAXITE APPLICATIONS

The following applications for Maxite in radio networks areforeseen:

• Coverage in macro-cells, including highway applicationswith up to 500 W EIRP

• Wide area coverage with up to 1250 W EIRP in ruralapplications

Sub-urban coverage with roof-top sites

Getting access to sites in the suburban areas might be both timeconsuming and costly. Using conventional macro base stations,it may sometimes be difficult to negotiate an advantageous siterental agreement. With Maxite on the other hand, the need ofphysical space for the base station cabinet is minimized. Spaceon a wall or on a pole is sufficient, perhaps outdoors on a roof-top.

Macrocell with BTS in the basement

Sometimes it might be possible to find an antenna space on arooftop, and a suitable BTS location in the basement. Longfeeders are needed. The operator can then benefit from usingMaxite, thus avoiding radio performance degradation caused bythe long feeders.

Highways

The special version called Maxite Highway provides theoptimum solution for highway coverage. There are considerablesavings in site costs, installation costs and running costscompared to a conventional highway solution.


– 108 – EN/LZT 123 3314 R3A

Macrocell with long feeders on a large building

A two or three sector site can be positioned on a rooftop. Afternegotiations with the building owner it is clear the antennas canbe places as required (on each side of the building). However,there is only one place for positioning the BTS equipment. Longfeeders are needed to one of the antennas. The operator can thenbenefit from using Maxite. The weight of the feeders for Maxiteare 20% of a similar macro RBS solution.

Very high towers

At a site with a very high tower (more than 50 meters), the longfeeders will degrade the performance of the radio signalsconsiderably. Using Maxite, and in particular the 1250 W EIRPversion, is then the best way to maximize coverage.

10 Implementation

EN/LZT 123 3314 R3A – 109 –

COMBINERS

Combiners are needed to enable more than one transmitter to beconnected to one common transmitting antenna. Without acombiner the output from one transmitter would loop back intothe output of another since both of them are physicallyconnected to the same antenna.

In GSM, two different TX-combiners can be used: filter andhybrid.

FILTER COMBINER

The filter combiner (Figure 10-10) can combine the output of upto eight transmitters. In order to use nine to sixteen transmittersin a cell, two filter combiners and two antennas must be used. Itis a narrow band combiner where the frequency of each of theconnected transmitters (TRXs) must be tuned by adjusting afilter. This is done automatically by the system but neverthelesstakes some time.

The total loss in a filter combiner is around 3-4 dB.

The combiner output is connected to the transmitting antenna ontop of the cabinet via a Measuring Coupling Unit (MCU) and abandpass filter TXBP.

The transmitter divider (TXD) distributes the loopback signalfrom the MCU to the different transmitters to enable automatictuning of the filters in the combiner. It also enables Voltage-Standing-Wave Ratio measurements (VSWR).


– 110 – EN/LZT 123 3314 R3A

TX

To TRXs, for adjustmentof the center frequencyof the filter

3 dB loss

935-960 MHz

Filter CMB

MCU

TXBP

TDX

RTX 16

RTX 1

Figure 10-10 TX combiner system, filter combiner, blockdiagram

10 Implementation

EN/LZT 123 3314 R3A – 111 –

HYBRID COMBINER

Hybrid combiners (Figure 10-11) can only combine twotransmitter outputs to one common output. If the BTS has morethan two TRXs, the combiners can be connected in a cascadingfashion. The hybrid combiner is a broadband combiner and doesnot need tuning. The loss in the hybrid combiner is around 3 dBfor each step in the cascade. This means a loss of 3 dB in a BTSwith two TRXs and 6 dB in a BTS with four TRXs. A third stepin the cascade results in total combiner loss of 9 dB. This loss istoo high and gives a maximum of four TRXs combined to oneoutput if hybrid combiners are used.

MCU

TXBP

Hybrid CMB

Hybrid CMBHybrid CMB

RTX 2RTX 1 RTX 4RTX 3

TX

935 - 960 MHz

3 dB loss

3 dB loss

Figure 10-11 TX combiner system, hybrid combiner, blockdiagram


– 112 – EN/LZT 123 3314 R3A

COMBINER AND DISTRIBUTING UNIT (CDU)

Currently, there are four different types of combiners that existfor RBS 2000:

• Combiner and Distributing Unit type A (CDU-A) has nohybrid combiner (Figure 10-12)

• Combiner and Distributing Unit type C (CDU-C) withhybrid combiner (Figure 10-13)

• Combiner and Distributing Unit type D (CDU-D) which is afilter combiner (Figure 10-14)

• Combiner and Distributing Unit type C+ (CDU-C+) is thenew combiner which will replace CDU-C. This will beexplained further.

CDU TYPE A

DUPLEXOR

ANT A

TX 1 RX A

DUPLEXOR

ANT B

TX 2 RX B

Figure 10-12 CDU-A. TXs with diverse receiver antennas

10 Implementation

EN/LZT 123 3314 R3A – 113 –

Duplexfilter

Splitter

Hybridcombiner

TX1 TX2RXA

A-antenna

Duplexfilter

Splitter

Hybridcombiner

TX4TX3RXB

B-antenna

Figure 10-13 Example of CDU-C. 4 TXs into two diversereceive antennas


– 114 – EN/LZT 123 3314 R3A

TRU 1TX

RX (a)

RX (b)

TRU 2

TX

RX (a)

RX (b)

CDU_D19DL

TXBP MCU

TRU 3 TX

RX (a)

RX (b)

CDU_D19UL

RXBPRXDA

RXD1:6

12

6

RXD1:6

12

6

FCOMB

FCOMB

FCOMB

FCOMB

FCOMB

FCOMB

RXBPRXDA

RXD1:6

12

6

RXD

1:6

12

6

ALNA

ALNA

TRU 5 TX

RX (a)

RX (b)

TRU 4 TX

RX (a)

RX (b)

TRU 6 TX

RX (a)

RX (b)

Figure 10-14 CDU-D configuration

As mentioned before, the new combiner, CDU-C+, will take theplace of the old CDU-C combiner in production. CDU-C+ istotally backwards compatible with CDU-C, i.e. it will mimic allthe functions of CDU-C in all prior software revisions.

Some of the new features associated with CDU-C+ are asfollows:

• An extra RX-chain - With the introduction of an extra RX(Receiver)-path, the most requested 2102 configuration(2,2,2) in one cabinet will now be possible. CDU-C+requires only one CDU per sector instead of two per sectorrequired by CDU-C. This will allow customers wanting todeploy sites in a 2,2,2 configuration to do so with as fewcabinets as possible, upgrading to two cabinets later whenthere is a need for higher capacity.

10 Implementation

EN/LZT 123 3314 R3A – 115 –

• Higher sensitivity – When using CDU-C+ with the requiredsoftware release, a higher sensitivity is attained. Specifically,the improvement is 2 dBm over CDU-C. This only applies,however, when used without the TMA (Tower MountedAmplifier). If the TMA is used, it is the sensitivity of theTMA that defines the system sensitivity.

• Built in duplexer – This is used to reduce the number ofrequired antennas, but it is important to note that this featurecan be bypassed if needed.

• Common antenna terminal – CDU-C+ can be used with acommon antenna terminal for both RX and TX or withseparate connections. The configuration using separateconnections is designed mainly for use with a TMA system.

With the advent of CDU-C+, CDU-A will be seen as thecoverage solution, CDU-D as the large capacity solution, andCDU-C+ will be considered as the “standard” configurationsolution, ultimately replacing CDU-C.


– 116 – EN/LZT 123 3314 R3A

SENSITIVITY

In Figure 10-15, the sensitivity figures at the RX reference pointare shown. The sensitivity reference points are shown inFigure 10-16. The figures are valid for both RBS 2000 and RBS200/205.

System Cellplanningsensitivity

Worst casesensitivity

GSM 900 -107 dBm -105 dBm

GSM 1800 with TMA -109 dBm -107 dBm

GSM 1800 without TMA -106 dBm -104 dBm

Figure 10-15 Base station receiver sensitivity

BTS

Cabinet

TMAFeeder & Jumpers

Without TMA:RX ref point 2

With TMA:RX ref point 1

Figure 10-16 Rx reference points with and without TMA (TowerMounted Amplifier)

MICRO BASE STATION

The micro base station RBS 2301 is introduced inCME 20 R6/CMS 40 R2.

Cellplanningpower

Worst casesensitivity

1.6 W, 32 dBm -104 dBm

Figure 10-17 Micro base station performance

MOBILE STATION

There are four GSM 900 or two GSM 1800 MS power classesspecified in the GSM 900 recommendations. The maximumpeak power is defined at the antenna connector.

The values in Figure 10-18 are from the GSM 900specifications. MS manufacturers tend to tune down the peakpower within the limits of type approval. However, most MSsare still very close to the nominal output.

10 Implementation

EN/LZT 123 3314 R3A – 117 –

The maximum peak power can be reduced down to 5 dBm(GSM 900) or 2 dBm (GSM 1800) for the MS power classes, ifthe BSC so decides.

System MS power class Cellplanning

power

Worst case

power

GSM 900 2 39 dBm (8 W) 37 dBm

GSM 900 3 37 dBm (5 W) 35 dBm

GSM 900 4 (handheld) 33 dBm (2 W) 31 dBm

GSM 900 5 (handheld) 29 dBm (0.8 W) 27 dBm

GSM 1800 1 30 dBm (1 W) 28 dBm

GSM 1800 2 24 dBm (0.25 W) 22 dBm

Figure 10-18 The MS power classes

In the GSM 900 recommendations, a reference sensitivity levelis defined. The actual sensitivity level of the MSs should bebetter or equal to this reference sensitivity level. Most MSs(including Ericsson’s) have a performance >2 dB better thanrequired, justifying the presence of a nominal sensitivity figure.

System MS type Cellplanning

sensitivity

Worst case

sensitivity

GSM 900 Handheld -104 dBm -102 dBm

GSM 900 All other types -106 dBm -104 dBm

GSM 1800 Handheld -102 dBm -100 dBm

Figure 10-19 MS reference sensitivity

No loss or antenna gain should be used for the MSs.


– 118 – EN/LZT 123 3314 R3A

ANTENNAS

BASIC ANTENNA TYPES

Each base station has its own antenna system consisting oftransmitting and receiving antennas. As an antenna is passive,the only way to obtain a gain in any direction is by concentratingthe radiation. The gain is therefore not in the transmitted powerbut in the power density transmitted in a particular direction. Ifthis direction is in the direction of communication, gain isobtained. In order to achieve a directional antenna (an antennawith gain), a reflector can be used. The radiated power can alsobe concentrated by stacking dipoles on the same vertical line andfeeding them in such a way that correct power and phase isobtained at each dipole element. The combined field from alldipoles add together more or less constructively, and gain isobtained in directions where constructive interference is found.Each doubling of the number of dipole elements (which alsocorresponds to a doubling of the length of the antenna) increasesthe gain by 3 dB.

Depending on which radiation pattern is wanted, different typesof antennas can be chosen. Below are the most commonly usedantenna types for mobile telephony base stations:

• Omnidirectional antennas

• Unidirectional antennas

• Special antennas

• Multi antenna systems

OMNIDIRECTIONAL ANTENNAS

Omnidirectional antennas (often referred to as omni antennas)have a uniform radiation pattern with respect to horizontaldirections. Looking at vertical directions, though, the radiationpattern is concentrated. Typical gain values are 6 to 9 dBd. Thelimiting factor is mainly the physical size of the antenna. As anexample, an omni antenna with a gain of 9 dBd has a height of 3meters.

10 Implementation

EN/LZT 123 3314 R3A – 119 –

UNIDIRECTIONAL ANTENNAS

This type of antenna has a non-uniform horizontal and verticalradiation pattern and is often used in sector cells. Thus they arealso called sector antennas. The radiated power is concentrated(more or less) in one direction.

Since the radiation is concentrated in the horizontal plane bymeans of reflectors, some gain is already realized. However,antenna elements can also be stacked (similar to omni antennas)to increase the resulting vertical gain.

Typical gain values for unidirectional antennas are 9 to 16 dBd.

SPECIAL ANTENNAS

An example of a special antenna is the slotted coaxial cable. It isoften referred to as the “leaky cable”. Slots in the corrugatedcopper outer allow a controlled portion of the transmitted powerto radiate along the entire length of the cable. Conversely, asignal transmitted near the cable will couple into these slots andbe carried along the cable. Because of its broadband capability,such a cable system can handle two or more communicationsystems at a time. A leaky cable is a good application for an areaof any shape (open or enclosed) which requires localizedcoverage. When using a leaky cable, no gain is achieved. Onenegative aspect of the leaky cable is that it is quite expensive touse.

For most applications, a typical value of sufficient transmittedpower is 20 - 30 W.

MULTI ANTENNA SYSTEMS

As the name indicates, a multi antenna system is a number ofindividual antennas forming a combined radiation pattern. Oneof the simplest types of such a system is a pair of directionalantennas mounted in opposite directions on a tower and fedthrough a power splitter. The purpose of such an arrangement isto cover a large area with one cell (i.e. along a road using alower number of channels than would be the case if using twocells).


– 120 – EN/LZT 123 3314 R3A

Multi antenna systems can also be used to form anomnidirectional pattern (i.e., around a building or a large towerconstruction in cases where an omni antenna cannot be used), orwhen more gain is needed than what an omni antenna systemcan offer in order to achieve larger coverage area.

The typical gain would be the gain of the individual antennasused minus losses due to the power splitter (3 dB).

DIVERSITY

There is a need for receiver diversity in GSM systems toimprove the uplink. The conventional method has been spacediversity where the two RX antennas are separated by a certaindistance but based on experience from measurements andsimulations and in the view point of installation advantagespolarization diversity can be used in some configurations.

If space diversity is used, the two RX signals are demodulated,and decoded, and the best signal based on Bit Error Rate (BER)is used. The effect is an increase in signal strength of three to sixdB.

SPACE DIVERSITY

Figure 10-20 shows a traditional configuration with spacediversity. The horizontal space needed for the antennas isdependent on the required diversity separation. Thisconfiguration is used for RBS 2000 with CDU-A (2 TRUs/cell),CDU-C (3 - 4 TRUs/cell), CDU-D (>6 TRUs/cell) and RBS 200(>8 TRXs/cell). Note that in some cases the duplex filters areexternal.

10 Implementation

EN/LZT 123 3314 R3A – 121 –

TX1/RXA

BTS equipment

CommonTX/RXantenna

dd

CommonTX/RXantenna

Horizontal separation, dd, for diversity = 12-18 λHorizontal separation, dd, for isolation 30 dB = 2 λ(antennas with 65 degrees beamwidth, all gain values)

TX2/RXB

TX1/RXA

TX1/RXA

TX2/RXB

TX1/RXA

TX2/RXB

Top View

1 λ = 0.33 meter at 900 MHz1 λ = 0.17/0.16 meter at 1800/1900 MHz

3-sector site withantennas on thesame height

TX2/RXB

Figure 10-20 Antenna configuration with space diversity

POLARIZATION DIVERSITY

A dual-polarized antenna is an antenna device with two arrayswithin the same physical unit. The two arrays can be designedand oriented in different ways as long as the two polarizationplanes have equal performance with respect to gain andradiation patterns.


– 122 – EN/LZT 123 3314 R3A

Vertical + Horizontal Polarization +-45 Degrees Polarization

Vertical array Horizontal array+45 degrees -45 degrees

Antennahousing

Antennahousing

Connectors

Feeders

Connectors

Feeders

Figure 10-21 Dual polarized antennas

The two most common types are vertical/horizontal arrays andarrays in +/-45 degree slant orientation (Figure 10-21). The twoarrays are connected to the respective RX branches in the BTS.The two arrays can be used as combined TX/RX antennas(Figure 10-22), and then the number of antenna units is reduced,compared with space diversity. The use of a duplex filterreduces the number of antenna units to only one per celldepending on configuration.

10 Implementation

EN/LZT 123 3314 R3A – 123 –

TX1/RXA

BTS equipment

Required isolation >30 dB betweenthe two antenna parts

TX1/RXA +TX2/RXB

Top View

3-sector site withantennas on thesame height

TX2/RXB

TX1/RXA +TX2/RXB

TX1/RXA +TX2/RXB

Figure 10-22 Antenna configuration with polarization diversity

The diversity gain obtained from polarization diversity isslightly less then the gain from space diversity. In the mostcritical environments (such as indoors and inside a car) the gainis, however, almost as good as if space diversity were used.

A dual polarized antenna offers very low correlation between thetwo received signals but the power reception of each branch isslightly better with space diversity. This implies a small benefitfor space diversity in noise-limited environments. For mostapplications, the difference is negligible. In interference-limitedenvironments, on the other hand, the low correlation obtained bypolarization diversity is advantageous.

Due to slightly different propagation characteristics for differentkinds of polarization, the downlink from a +/-45 degree dualpolarized antenna suffers from about 1.5 dB extra loss comparedto a vertically polarized antenna. This loss only affects thedownlink.

The isolation between the two polarization planes needs to be30 dB according to the GSM specifications. The size of theantenna must remain small, as the intention with polarizationdiversity is to reduce the outlook of the antenna installation.


– 124 – EN/LZT 123 3314 R3A

ANTENNA TILT

When the antenna is mounted vertically, the main lobe of theantenna radiation pattern will follow a horizontal line starting atthe center point of the antenna.

For reasons, such as co-channel interference and time dispersionproblems, it may be useful to tilt the antenna and let the mainlobe point a few degrees downwards.

Down tilting should be done with great care, as the cell patternis disturbed and unpredictable reflections might be generated.Coverage at the cell border is also reduced.

10 Implementation

EN/LZT 123 3314 R3A – 125 –

BSC CAPACITY

There are basically four methods to measure the capacity of aBSC, i.e. how much can a BSC handle. These all work togetherto give the most accurate figure when measuring capacity:

• Traffic capacity

• Call capacity

• Subscriber capacity

• Addressing capacity

Traffic capacity tells how many simultaneous calls one BSC canhandle. However, this figure does not give a hint on non-callactivities, such as registrations. Thus this measure is not, byitself, well suited for expressing BSC capacity.

Call capacity tells how many call attempts a unit can handleduring busy hours. This measure still does not take into accountnon-call related activities, such as registrations.

Subscriber capacity tells how many subscribers that can beserved by a unit. This figure is strongly dependent on subscriberbehavior. What sets the limit is real-time capacity (or size of theradio network).


– 126 – EN/LZT 123 3314 R3A

Addressing capacity tells how many hardware or softwaredevices that can be connected and/or defined. This is also knownas system limits. Here, no considerations to real-time processingneeds or amount of traffic are made. For typical traffic models,the processing limits hit the ceiling before the addressing limitsare reached. However, high addressing capacity can be useful,e.g. for coverage of low traffic areas.

System Tuning

Chapter 11

This chapter is designed to provide the student with an overviewof the Ericsson tools that are available and recommended for usein a network optimization process. It addresses the tools, theirfunctions, features, and requirements.


• Describe EET and explain the benefits and capabilities of thecell planning tool

• Explain TEMS functionality and advantages

• Discuss the use of OSS for the purpose of viewing,reconfiguring, and implementing cells

• Explain how the Performance Measurement Report in OSScan be used for performance analysis of the network

• Identify the capabilities of the Cellular Network Analyzer

• Describe how parameter settings affect network performance

11 System Tuning

EN/LZT 123 3314 R3A – i –

11 System Tuning

Table of Contents

Topic Page

INTRODUCTION................................................................................127

ERICSSON ENGINEERING TOOL (EET) .........................................128

NETWORK DIMENSIONING..................................................................................... 128

FREQUENCY PLANNING ......................................................................................... 128

PREDICTING............................................................................................................. 128

TOOLS....................................................................................................................... 129

TEST MOBILE SYSTEM (TEMS) ......................................................131

TEMS TRANSMITTER............................................................................................... 132

TEMS RECEIVER...................................................................................................... 133

HOT SPOT FINDER...........................................................................134

OPERATIONS SUPPORT SYSTEM (OSS).......................................135

INTRODUCTION ....................................................................................................... 135

CELLULAR NETWORK ADMINISTRATION (CNA) .................................................. 135

CELLULAR NETWORK ADMINISTRATION INTERFACE (CNAI)............................ 136

CELLULAR NETWORK MEASUREMENT AND RECORDINGS.............................. 137

CELLULAR NETWORK ANALYZER (CeNA) ...................................143

CELL PARAMETER ADJUSTMENT .................................................146

CELL PARAMETERS ................................................................................................ 146

11 System Tuning

EN/LZT 123 3314 R3A – 127 –

INTRODUCTION

After an initial cell plan has been compiled and approved, it istime to begin the installation of the network equipment. As atime-saving measure, we can begin to optimize the performanceof the radio network as it is being built up. This chapter broadlycovers some of the tools that Ericsson recommends fordiagnosing a network. The major benefit of using these toolscomes not only from their initial use but through their continueduse to monitor and improve network performance. At the end ofthis chapter, radio network tuning by means of parameteradjustment is discussed.


– 128 – EN/LZT 123 3314 R3A

ERICSSON ENGINEERING TOOL (EET)

During the initial phases of the network design process, areliable radio wave propagation tool is necessary. This needcontinues to exist even for the most mature radio networks. Oneof the primary responsibilities of an RF engineer is to improvethe radio network when required to do so. This could be theresult of growth or decreased performance. EricssonEngineering Tool (EET) is based on experience and continualdevelopment adapted to a rapidly changing technology.

EET is based on Planet by Mobile Systems International Ltd.(MSI). It is a UNIX open-windows-based software packagedesigned to simplify the process of planning and optimizing acellular network. Some of the more important features of EETare discussed in the following sections.

NETWORK DIMENSIONING

In the software , it is easy to create new sites or move old ones.All information about the sites is stored in the site database. It ispossible to make changes to one site, a group of sites, or allsites.

A height path profile can be displayed between any two pointson the map. This is very useful for microwave link planning.

FREQUENCY PLANNING

EET allows the allocation of channels or frequency groups to acell. It is possible to do this manually or automatically. Thefrequency assignments are stored in the carrier database. Thefrequencies can be displayed by labeling the cell with theAbsolute Radio Frequency Channel Number (ARFCN), thegroup name, or by color coding the coverage areas according tothe frequency groups.

PREDICTING

When the sites are created it is time to initiate a prediction. It ispossible to predict one site, a group of sites, or all sites. Theresult of the prediction is the pathloss from the sites.

After predicting, arrays for coverage and interferences (C/I andC/A) can be created. The signal strength and interference levelsare calculated for each pixel. The advantage of having both

11 System Tuning

EN/LZT 123 3314 R3A – 129 –

prediction and array steps in this procedure is that it speeds upthe calculations. If the user would like to change, e.g. the outputpower at one site, there is no need for a new prediction becausethe change does not affect the pathloss. The user only has tocreate a new array. Creating arrays is just a matter of adding dB,so it is not very time-consuming. On the other hand, predictionsare more complicated.

The basic propagation model in EET is Ericsson’s Algorithm9999. Out of the topographical database, the profile between thetransmitter and receiver is extracted. The pathloss is calculatedbased on terrain variations in height along the profile (includingcontributions due to knife-edge diffraction), the earth’scurvature, the land usage, and empirical corrections. Thecorrections are a result of Ericsson’s unique experience of wavepropagation and numerous measurements all over the world. Inaddition to Algorithm 9999, EET includes the Okumura-Hata,the Cost 231, and the Walfish-Ikegami models.

TOOLS

Using EET, the user can spread traffic on the map to plan forcapacity. The traffic can be displayed with different colors for

different amounts of Erlangs/km2 or the user can highlight the

cells that do not meet the specified GoS.

It is possible to import data from a test mobile and display theinformation on the map.

EET can import radio survey files which can be used to tune theprediction model for the area where the network is to beplanned.

Data can be imported and exported to OSS.


– 130 – EN/LZT 123 3314 R3A

Figure 11-1 Graphical user interface, the EET main window

11 System Tuning

EN/LZT 123 3314 R3A – 131 –

TEST MOBILE SYSTEM (TEMS)

The TEst Mobile System (TEMS) is a test tool used to read andcontrol the information sent over the air interface between thebase station and the mobile station in a GSM system. It can alsobe used for radio coverage measurements. Furthermore, TEMScan be used both for field measurements and post processing.

TEMS consists of a mobile station with special software, aportable PC, and optionally a GPS receiver (Figure 11-2).

MS

GPS

PC

Figure 11-2 TEMS Hardware

The mobile can be used both in active state and idle mode,additionally, it can be use in any GSM network, depending onthe SIM card. Both layer two and layer three messages can bemonitored and recorded. The MS can simulate GSM 900 powerclass 2 to 4. It is possible to lock on a single frequency. The MScan test each time slot on a selected frequency to verify that allTCHs are available and functioning.

The PC is used for presentation, control, and storage of themeasurements. For the serving cell, it is possible to display, e.g.RxLev, Rxqual, TX power, TA, Base Station Identity Code(BSIC), and ARFCN. For the six strongest neighboring cells, itis possible to display RxLev, BSIC, and ARFCN. Theinformation can be displayed in real-time or recorded andreplayed.


– 132 – EN/LZT 123 3314 R3A

The GPS receiver gives the position of the measurements. Whenthe satellite signals are shadowed by obstacles, the GPS systemmay be used for dead reckoning.

The TEMS measurements can be imported to EET with the useof File and Information Converting System (FICS). This meansthat the measurements can be displayed on the map so that, e.g.the measured handovers can be compared with the predicted cellboundaries. FICS can also convert to EXCEL and wordprocessing packages.

Figure 11-3 TEMS graphical user interface

TEMS TRANSMITTER

For the generation of test signals, it is suitable (however notmandatory) to use one or several TEMS Transmitters. TheTEMS Transmitter is a small unit that transmits in the GSMdownlink band. The output power is adjustable between 17 and27 dBm. A complete editable BCCH is transmitted while theother 7 time slots contain an unmodulated carrier.

In absence of TEMS Transmitters, a Test TransMitter (TTM)can also be used. This is a narrow band Continuous Wave (CW)transmitter with a maximum output power of 43 dBm.

Additionally, the regular transmitter can be used for thisfunction.

11 System Tuning

EN/LZT 123 3314 R3A – 133 –

TEMS RECEIVER

The recommended receiver is TEMS Light equipment. This is aTEMS mobile station connected to a small Fujitsu PC operatedwith a pen. The TEMS Light program is a reduced version ofnormal TEMS but with the possibility to log fixpoints bymarking them with the pen on a scanned map. The informationin the log files is displayed on the scanned map as color marksassociated with a window containing more information abouteach mark.

If TEMS Light is not available, the standard TEMS equipmentor a Test Measurement Receiver (TMR) can be used.

An even faster coverage verification can be made by usingTEMS Pocket. This is a test mobile station with some TEMSfunctions available on the mobile display. TEMS Pocket cannotbe operated from a computer. Areas where the signal may beweak are checked by locking TEMS Pocket to the used AbsoluteRadio Frequency Channel Number (ARFCN) and Base StationIdentity Code (BSIC) and reading the signal from the display.There is also an audible warning to indicate a low signal.


– 134 – EN/LZT 123 3314 R3A

HOT SPOT FINDER

It is important to deploy microcells where the heaviest traffic islocated (also known as “hot spots”). One way to find suitablelocations for microcells is Ericsson’s Hot Spot Finder. The HotSpot Finder is a GH388 mobile modified to transmit aBCCH/BSIC combination signal. Basically, it acts as a dummycell. The mobiles in the surrounding cells will treat the Finder asa neighbor and include BCCH/BSIC combination signals in themeasurement reports. Different locations and antenna types andpositions can be tested prior to the implementation of themicrocell. The potential traffic is estimated by looking at themeasurement reports for the mobiles in the surrounding cells.

11 System Tuning

EN/LZT 123 3314 R3A – 135 –

OPERATIONS SUPPORT SYSTEM (OSS)

INTRODUCTION

The GSM Operations Support System (OSS) is a UNIX basedtool that enables the supervision, planning, and engineering of anetwork from one central location. The capabilities of the OSSthat concern RF Engineers are discussed in this chapter(Figure 11-4).

Figure 11-4 OSS Main menu

CELLULAR NETWORK ADMINISTRATION (CNA)

One of the most important aspects of managing a cellular radionetwork is that of managing the individual cells. The cellsrepresent the infrastructure from which the mobile subscriberaccesses the network. Hence, a poorly managed infrastructurewill most likely be reflected by dissatisfied customers and asubsequent loss of revenue.

The purpose of the Cellular Network Administration (CNA)feature is to provide a user-friendly interface from which a usercan manage the cells in an efficient and controlled manner.


– 136 – EN/LZT 123 3314 R3A

Normally, there is a multitude of radio related parameters thatneed to be set in several different network elements in aconsistent manner in order to achieve a well-balanced, properlyfunctioning radio network. Default parameters are used whenthe operator does not enter a parameter value. Parameters can becopied from one cell and pasted into another. It is also possibleto create profile areas collecting all cell parameters commonlyused for different types of cells. Cell parameters are validated atthe time of the entry. This particular feature helps to reduce thepossibility of incorrect cell parameters and increases theefficiency of personnel as the number of cells in the networkincreases (Figure 11-5).

Figure 11-5 Part of OSS CNA menu

CELLULAR NETWORK ADMINISTRATION INTERFACE (CNAI)

The Cellular Network Administration Interface (CNAI) is anexternal interface to Cellular Network Administration. TheCNAI allows for an external cell planning tool, e.g. EET, toexchange information with the CNA database. The data isexchanged between the two via ASCII coded text files. Theessence of this interface is to provide simplified data import andexport capabilities to CNA for ease of user handling of the datatransfer mechanism.

Cell planning data can be used as an example. The OSS interactswith the Ericsson Engineering Tool (EET). Such externalsystems can retrieve data from the actual radio network, re-engineer the new cell data, and transfer back the new cell data ina simple manner (Figure 11-6). This avoids time-consumingmanual entry.

11 System Tuning

EN/LZT 123 3314 R3A – 137 –

CNA

OSSOSS

Consistency Checking

Planned Areas Valid Area Physical Network

Update

AdjustInte

rfac

e

Figure 11-6 CNA Interface (CNAI)

CELLULAR NETWORK MEASUREMENTS AND RECORDINGS

Cellular network measurements are a way for the telecomcompany to supervise the quality of the services provided by thenetwork. The measurements are used for early detection of faultsand for planning future extensions. For the measurements to beof any use, however, they must be presented in a way that clearlyshows the conditions or trends that are of interest.

Long-term Measurements

These performance measurements present statistics collected inOperation & Maintenance Subsystem (OMS) and Statistics andTraffic measurement Subsystem (STS). They provide a pictureof how the network is working. This requires a certain amountof subscribers generating traffic in the network.

• Data from OMS measures traffic on routes, traffic types, andtraffic dispersion on different routes. All recordings relate toone piece of AXE equipment.

• Data from STS collects data regarding statistics within theradio environment.


– 138 – EN/LZT 123 3314 R3A

Statistical Report Packages (SRP)

The Statistical Reports Package (SRP) is a new set of reportswhich focuses on collecting and presenting data used formanaging, planning, and engineering a cellular network. Thereports are divided into three categories:

• Management reports

• Planning and Engineering reports

• Operation reports

This provides different target group reports specially designedfor specific needs.

The SRP reports are primarily based on the statisticalinformation from the AXE subsystems STS and OMS, (i.e. fromthe performance measurements described earlier) and is part ofthe performance measurement functions in OSS. This data iscomplemented with configuration data retrieved from the CNAdatabase and printouts obtained from interrogation of thenetwork elements.

Management Report

The Management Report and its subreports are customized toprovide comprehensive information about the cellular networkin both tabular (alphanumeric) and graphical form. This enablesmanagement to increase the control of the network size andbehavior. The following subreports are included in the tabularpart of the Management Report:

• Subscriber Data

• Call Processing

• Infrastructure Data

• System Performance

• Cell Performance

The following reports are included in the graphical trends part ofthe Management Report:

• HLR Subscribers Distribution Trend

• Network Subscriber Distribution Trend

• Traffic per Subscriber Trend

11 System Tuning

EN/LZT 123 3314 R3A – 139 –

• System Calling Minutes and Traffic Trend

• Switch Peak Hour and Peak Processor Load Trend

• Restart Statistics Trend

• Availability Trend

• Radio Network Performance Trend

Planning and Engineering Reports

The Planning Reports are specifically designed to identify thoseareas in the radio network and the switching system wherefuture network expansions are needed. The reports assist infocusing network improvements on critical problem areas andsupport the operator in prioritizing the planned networkexpansions. These are a set of reports intended to be used byplanning and engineering personnel and consist of the reportslisted below:

• Planning Report Summary

The Planning Report Summary provides an overview of thenetwork size and performance in both tabular and graphicalform. Information presented in tabular form is:

Call processing per node:

– System Performance (switched network, radio network,and traffic channel availability)

– Cell Performance (total dropouts, top and bottom 10 ontraffic, top and bottom 10 on congestion and utilization)

– Handover Performance

The graphical trend reports in the Planning Report Summaryare:

– Switch Peak Hour and Peak Processor Load

– System Calling Minutes Trend

– Restart Statistics Trend

– Availability Trend

– Radio Network Performance Trend

– SDCCH Performance Trend

• Radio System Performance

• Cell Traffic reports

• Neighboring Cell Analysis Performance report


– 140 – EN/LZT 123 3314 R3A

• Route Traffic report

• Traffic Dispersion report

• Traffic Profile

• Subscriber Activity Profile

Operation Reports

Operation Reports include reports for both the radio networkand the switching network. The reports are used to quicklydetect cells and routes with unacceptable performance enablingthe operator to take immediate action to preserve the quality ofthe network. In particular, the exception reports combined withSRPs scheduling function provide an invaluable tool foridentifying and analyzing vital performance situations. Thereports listed below are generated only if the thresholds set bythe operator are passed.

• Radio Exception report

• Detailed Radio Statistics report

• Handover Matrix report

• Route Exception report

• Call Rejection Exception report

Short-term Measurements

The traffic recording part of Radio Network Recording (RNR)serves as a compliment to regular statistics collection as well asto so-called test mobile systems, e.g. TEMS. RNR monitors thebehavior of the mobile from the network side and from themobile telephone side as well as the result of the chosen channelallocation strategy. This is enabled by three recording facilitiesin the BSC:

• Mobile Traffic Recording (MTR)

• Cell Traffic Recording (CTR)

• Channel Event Recording (CER)

Mobile Traffic Recording (MTR)

The mobile-related recordings trace practically everythinghappening to a mobile, specified by International MobileSubscriber Identity (IMSI) number, by recording events andmeasurements. Example of events are Layer 3 messages,

11 System Tuning

EN/LZT 123 3314 R3A – 141 –

assignment, handover, and disconnection. Measurementsinclude, e.g. signal strength on uplink and downlink, signalquality on uplink and downlink, signal strengths for neighboringcells, and timing advance. The recorded material is sent from theBSC to OSS where the mobile-related recording data can beanalyzed and presented in tables or graphs (Figure 11-7). Theresults can be used for tuning cell parameters for better systemperformance, forecast, cell planning, and verifying poor mobilestations.

Figure 11-7 MTR

Cell Traffic Recording (CTR)

The Cell Traffic Recordings (CTRs) record the same type ofinformation as MTR but for a number of connections originatingin a specific cell. The recording (which can last minutes orhours) is actually a collection of short recordings, each up to 30seconds. Up to 16 different MSs can be recordedsimultaneously. The recordings are triggered on an event, e.g.handover and record an equal amount of time, e.g. five secondsbefore and after the event.


– 142 – EN/LZT 123 3314 R3A

Channel Event Recording (CER)

The Channel Event Recording (CER) is a new BSC feature thatcan be initiated from OSS. It assists in monitoring the followingBSC functions:

• Channel Administration

• Differential Channel Allocation

• Idle Channel Measurements

CER classifies TCHs and SDCCHs into different interferencelevels based on measurements.

11 System Tuning

EN/LZT 123 3314 R3A – 143 –

CELLULAR NETWORK ANALYZER (CeNA)

The Cellular Network Analyzer (CeNA) is a cellular qualityinformation system that enables optimization of a digitalnetwork’s performance. The system is comprised of severalfeatures that assist the network operator in achieving maximumperformance from the equipment (Figure 11-8).

MTU

Taxi

MTU

Taxi

MTU

Taxi

Public SwitchedTelephone Network

(PSTN)

Digital CellularNetwork

TCP/IP

Fixed TestUnit Server(FTU)

PresentationStation (PS)

ReportGenerator(RG)

DatabaseServer(DBMS)

MobileTest Unit(MTU)

Figure 11-8 Cellular Network Analyzer (CeNA)

The CeNA system is able to carry out a number ofmeasurements. The basic measurements used include:

• Received signal level (Figure 11-9 and Figure 11-10)

• Signal quality

• Timing advance

• Transmission power

• Call events (drop, block, connection, and handover)

• Protocol signaling (Layers 1 and 3)

Using CeNA, it is easy to generate reports, present measuredresults from a conversation (Figure 11-9), or present statistics(Figure 11-10).


– 144 – EN/LZT 123 3314 R3A

All measurement activities, result collection, result storing, andresult transition are based on unattended operation. Mobile units(MTUs) are mounted in vehicles and are assigned subscribernumbers in the cellular network (Figure 11-8). A subscribernumber in the public network is assigned to a fixed unit (FTU).In accordance with a measurement order, MTUs regularly callFTU, execute measurements, and transfer the results to a database for storage and transition (DBMS). All measurement-ordersetups, result presentations, and report generations are executedfrom an operator terminal (PS, RG).

All measurement results are stored in an open database system,allowing easy access to and sharing of information throughoutthe organization via LAN or WAN. The open database systemarchitecture also enables the operator to access and useinformation in the operator’s own application and in conjunctionwith information from the same or other database systems.

Figure 11-9 Conversation route presented by CeNA in map,table, and graph form (symbol key not shown)

11 System Tuning

EN/LZT 123 3314 R3A – 145 –

Figure 11-10 Statistics presentation by CeNA in map and tableform

Some of the main benefits of CeNA are as follows:

• Increased revenue with less call failure

• Improved call quality

• Accurate network information

• Maximized geographical coverage with existinginfrastructure

• Visual network performance

• Acquisition of network data and performance informationwithout a manpower requirement

• Detection of interference and improve frequency planning

• Improved traffic routing


– 146 – EN/LZT 123 3314 R3A

CELL PARAMETER ADJUSTMENT

When all measurements have been analyzed, the user maydetermine that there is an inconsistency in the parameter setting.The user can then tune hysteresis and offset parameters toimprove the network quality.

OSS provides a graphical user interface for changingparameters. Using OSS also limits the possibility for humanerrors by performing validation and consistency checks on theparameter settings. This means that the user can run checks onthe parameter settings before updating the network. OSS alsoprovides a storage and fallback area that can be loaded if errorsoccur during network update.

After the changes are made, the user can continue monitoringthe network performance in their STS-based reports in order tofollow up on the changes in network performance.

CELL PARAMETERS

When a new system is built or when new cells are added orchanged in an existing system, the cell planner provides theoperator with a document for each cell containing data forinsertion of the cell in the radio network. This document iscalled Cell Design Data (CDD). The data from all suchdocuments is then converted into Data transcript Tape (DT) andloaded into the corresponding BSC. A DT tape contains not onlyCDD information but also other data needed for the completeconfiguration of the BSC.

The reason for having so many parameters is so the operator canadjust and tune the network to fit their specific requirements. Allparameters are permitted to be set within a certain range andusually have a default value.

The default values provide a good basis to start with. Parameterscan be changed later if, e.g. measurements indicate thatadjustments are necessary. Several parameters should not bechanged at the same time because it is more complicated toknow which parameter setting change effected the system.

Some of the parameters are system specific and some are set persite, cell, or subcell.

11 System Tuning

EN/LZT 123 3314 R3A – 147 –

Offset

An offset is used to make a cell appear better (or worse) than itreally is by increasing/decreasing measured signal strength. Theoffset is a cell-to-cell relation and is always unsymmetrical.

Hysteresis

A hysteresis is used to prevent the ping-pong effect, meaningseveral consecutive handovers between two cells. The ping-pongeffect can be caused by fading, the MS zigzagging between thecells, or by non-linearities in the receiver. The hysteresis is acell-to-cell relation and is always symmetrical.

Control of Radio Network Features

Other parameters are used to control radio network features likeDiscontinuous Transmission (DTX), frequency hopping, andpower control.

Timers and Filters

There are some timers and filters which can be set byparameters. Depending on the timer settings or length of filters,the system responds faster or slower to the change. A fast systemis less stable than a slower system. A fast system is necessary ifmicro cells are used because handovers are frequent in this case.

Identification

Parameters used to identify, for example, a cell or a location areain the network.

Penalties

The penalties are used to punish a cell in the locating algorithm.When a cell is punished, it appears worse then it really is. This isto avoid handback in case of an urgent handover or to avoidseveral repeated handover attempts in case of signaling failure.Refer to Chapter 13 “Radio Network Features” for detailedinformation on penalties.


– 148 – EN/LZT 123 3314 R3A

Thresholds

Thresholds for cell ranking, call release, and access can be set.

12 System Growth

12 System Growth

Table of Contents

Topic Page

SYSTEM GROWTH ...........................................................................149

INTRODUCTION ....................................................................................................... 149

THE ERICSSON WAY FORWARD ........................................................................... 149

CELL SPLIT ............................................................................................................... 150

MULTIPLE RE-USE PATTERNS............................................................................... 152

SYSTEM GROWTH

INTRODUCTION

If the number of subscribers in a system continues to increase, atsome point it becomes necessary to increase the capacity of thesystem. There are several ways to do this:

• Increase the frequency band (e.g. a GSM 900 operator mightbuy GSM 1800 licenses)

• Implement half-rate

• Make frequency re-use tighter (e.g. going from a 4/12 re-usepattern to a 3/9 re-use pattern by implementing frequencyhopping)

• Make the cells smaller and smaller

After a description of the regular procedure for adding new sites(cell split), tightening of the re-use pattern by means of MultipleRe-use Pattern (MRP) is briefly discussed.

These methods of adapting to system growth will directly affectthe cell planning process as described throughout this chapter.

THE ERICSSON WAY FORWARD

Ericsson’s concept for increased capacity in GSM radionetworks is also known as “The Way Forward”.

As mentioned in Chapter 1, The Way Forward is a VROXWLRQFRQFHSW that combines a number of techniques, features, andservice products. Together they provide substantial capacity gainin GSM mobile telephone networks without the need foradditional radio frequency spectrum.

The focus of The Way Forward lies on tight frequency reuse andthe implementation of micro cells which together provide almostunlimited possibilities of capacity expansion.

The Way Forward solution concept has been developed in co-operation with GSM operators to ensure the fulfillment ofcustomer needs and requirements. Each time The Way Forwardis implemented, it is adapted to the local environment and thecustomer’s individual requirements.

The following procedures (cell split and multiple re-usepatterns) are directly involved in The Way Forward method.

CELL SPLIT

It is clear that a smaller cell size increases the traffic capacity.However, a smaller cell size means more sites and a higher costfor the infrastructure. Obviously, it is preferable not to workwith an unnecessarily small cell size.

What is needed is a method that matches cell sizes to thecapacity requirements. The system is started using a large cellsize, however, when the system capacity needs to be expanded,the cell size is decreased in order to meet the new requirements.

This normally also calls for using different cell sizes in differentareas. This method is called cell split, and is illustrated inFigure 12-1 through Figure 12-4.

([DPSOH�

Initially, the largest possible cell size is used consideringcoverage range (Figure 12-1). Next step is to introduce threecells per site (Figure 12-2), using the original sites and feedingthe cells from the corners. This represents a cell split of 1 to 3,(Figure 12-3). Now the number of sites is still the same, but thenumber of cells are three times as many as before. The followingstep is to do a cell split of, e.g. 1 to 4 (Figure 12-4).

As seen from the figure, the old sites are still used in the newcell plan, but additional sites are now required.

Figure 12-1 Cell split (phase 0)

Figure 12-2 Cell split (phase 1)

Figure 12-3 Cell split 1:3 (phase 2)

Cell split 1 to 3 (Figure 12-3) requires three times as many cells.After the split, the capacity is three times higher per area unit,and the cell area is three times smaller. The antenna directionson the site that existed before the split must be changed by 30degrees.


– 151– EN/LZT 123 3314 R3A

Figure 12-4 Cell split 1:4 (phase)

Cell split 1 to 4 (Figure 12-4) requires four times as many sites.After the split, the capacity is four times higher per area unit,and the cell area is four times smaller. There is no need tochange the antenna directions in a 1:4 cell split.

MULTIPLE RE-USE PATTERNS

Multiple Re-use Pattern (MRP) is Ericsson’s frequency re-usemethod and it is a scheme to gradually tighten the frequency re-use in a cellular network. It is also very well suited to handlingnetworks with uneven traffic distribution, i.e. different numberof transceivers (TRXs) in each cell.

A tighter frequency re-use means an increased interference leveland, therefore, the scheme means to handle this.

The idea is that instead of organizing the TCH carriers accordingto a single re-use scheme, the available frequencies could besplit into a number of segments each representing a re-usecluster. The BCCH carriers are already planned separately.These re-use clusters are of different sizes providing different re-use situations on each of the carriers in the cell. By thenapplying frequency hopping over all the carriers averaging theinterference situation in each overlaid re-use cluster, a veryefficient system is be built.

12 System Growth

EN/LZT 123 3314 R3A – 152–

1 2 3 4 5 6 7 8 9 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 3712 13

1. Frequency Allocation

3. Allocate to cell= Frequency plan

To microcellBCCH-carriers

To macrocelltransceivers

BCCH Macro TCH group 1 TCH 2 TCH 3

10 11

BCCH Micro

2. Divide into bands

Figure 12-5 Building up a Multiple Reuse Pattern (MRP)

In the example in Figure 12-5, the 37 carriers (7.5 MHz) couldbe split into 31 carriers for the macrocell layer, leaving sixcarriers for microcell BCCH frequencies. The macrocellfrequencies can be split into, e.g. the following bands 12/9/6/4,where the first twelve frequencies are used for the macrocellBCCH carrier frequencies (TRX1). The second group of ninefrequencies are then re-used on TRX2, the third group of sixfrequencies re-used on TRX3, etc. Each cell is only allocatedwith the necessary number of carriers (starting from the mostrelaxed re-use) given by the traffic requirements per cell up to amaximum of four TRXs per cell. The average re-use for12/9/6/4 is (12+9+6+14)/4=7.75 for the cell and (9+6+4)/3=6.3for the TCH frequencies.

This method allows a gradual tightening of the re-use as moretransceivers are installed in the cell. The re-use group on the lasttransceiver can be tight since it probably not will be used inevery cell. This can be traded into tighter re-use.

It will be possible to combine two transceiver cells using a 12/9allocation with three transceiver cells using 12/9/6, and fourtransceiver cells using 12/9/6/4.

Radio Network Features

Chapter 13

This chapter is designed to provide the student with an overviewof the basic, indispensable, network features.


• List the different processes that the idle mode task can besubdivided into

• Describe each of the idle mode tasks

• Explain the purpose and main flow of the locating algorithm

• Describe the difference between K-cells and L-cells

• List the main situations when a channel is allocated

• Describe what can be achieved by the different radionetwork features

13 Radio Network Features

EN/LZT 123 3314 R3A – clv –


Table of Contents

Topic Page

IDLE MODE BEHAVIOR....................................................................153

WHAT CAN BE ACHIEVED....................................................................................... 153

SHORT TECHNICAL DESCRIPTION ....................................................................... 154

LOCATING.........................................................................................166

INTRODUCTION ....................................................................................................... 166

BACKGROUND ......................................................................................................... 167



CHANNEL ADMINISTRATION..........................................................195

DYNAMIC MS (BTS) POWER CONTROL.........................................198



DTX ....................................................................................................205



FREQUENCY HOPPING ...................................................................208



INTRA-CELL HANDOVER.................................................................213



ASSIGNMENT TO ANOTHER CELL.................................................215



DYNAMIC OVERLAID/UNDERLAID SUBCELLS.............................216


– clvi – EN/LZT 123 3314 R3A



HIERARCHICAL CELL STRUCTURES.............................................220



EXTENDED RANGE ..........................................................................223



IMMEDIATE ASSIGNMENT ON TRAFFIC CHANNEL .....................226


DOUBLE BA LISTS ...........................................................................228


IDLE CHANNEL MEASUREMENTS .................................................230



CELL LOAD SHARING .....................................................................233



MULTIBAND OPERATION................................................................236



DIFFERENTIAL CHANNEL ALLOCATION.......................................239



ADAPTIVE CONFIGURATION OF SDCCH.......................................240

INTRODUCTION ....................................................................................................... 240



FREQUENCY ALLOCATION SUPPORT (FAS)................................242

INTRODUCTION ....................................................................................................... 242



EFFICIENT PRIORITY HANDLING (EPH) ........................................247


EN/LZT 123 3314 R3A – clvii –

INTRODUCTION ....................................................................................................... 247



NEIGHBORING CELL SUPPORT (NCS) ..........................................250

INTRODUCTION ....................................................................................................... 250




EN/LZT 123 3314 R3A – 153 –

IDLE MODE BEHAVIOR

WHAT CAN BE ACHIEVED

High Signal Strength When Accessing the System

The MS will at all times try to obtain the highest possible signalstrength when accessing the system. This is achieved by meansof the idle mode cell selection and reselection algorithms. Thesealgorithms enable the MS to choose the most suitable cell tocamp on (based on signal strength). A cell is suitable if certaincriteria are satisfied. Camping on the most suitable cell providesthe MS with a high probability of good communication with thesystem.

The cell selection and reselection algorithms are governed byparameter settings. Using these parameters an operator can (on aper cell basis) make a specific cell more or less attractive for theMS to camp on. This makes it possible for the operator toachieve similar behavior for MSs in idle mode as in dedicatedmode. Well-designed parameter settings for cell selection andreselection in idle mode makes the MS camp on the cell thatwould have been chosen by the BSC if the MS had been indedicated mode.

Control of the Paging Load

In idle mode, the MS notifies the network whenever it changeslocation area by the location updating procedure. Thus, thenetwork is kept up-to-date concerning which location area theMS is presently located. When the system receives an incomingcall, it knows in which location area it should page the MS. Thesystem does not need to page the MS throughout the entirenetwork, this reduces the load on the system. If the MS does notrespond on the first paging, the network can send a secondpaging.

In addition, the MS can periodically notify the network of itspresent status by the location updating procedure and when it ispowered on or off. This prevents the network from unnecessarilypaging MSs that have been powered off or have left thecoverage area. This may otherwise cause unnecessary load onthe system.


– 154 – EN/LZT 123 3314 R3A

Low Idle Mode Power Consumption

In idle mode, the MS occasionally monitors the systeminformation being transmitted in the current cell or performsmeasurements on neighboring cells to see if a cell change shouldbe initiated. However, most of the time the MS is in “sleepmode”, hence, the power consumption during idle mode is low.This is also referred to as discontinuous reception (DRX).

SHORT TECHNICAL DESCRIPTION

General

While the MS is in idle mode, it continuously performsmeasurements on the BCCH-carriers of serving and neighboringcells to determine which cell to camp on. If necessary, it willalso register its presence in the location area of the chosen cellby performing a location update.

The purpose of camping on a cell is threefold:

• It enables the MS to receive system information from thePublic Land Mobile Network (PLMN)

• The MS can initiate a call by accessing the network on theRandom Access CHannel (RACH) of the cell on which it iscamped

• The PLMN knows the location area of the cell on which theMS is camped (unless the MS has entered a “limitedservice” state) and can, therefore, page the MS when anincoming call is received

The idle mode tasks can be subdivided into 4 processes:

• PLMN selection

• Cell selection

• Cell reselection

• Location updating

The relationship between these processes is illustrated inFigure 13-1.


EN/LZT 123 3314 R3A – 155 –

PLMN selection

Cell selection

Cell reselection

Location updating

Automatic/ManualMode Selection

Indication to User

User Selectionof PLMN

PLMNAvailable

PLMNSelected

Initial Cell Selected

LU ResponsesCell & LAChanges

New LA

PeriodicRegistration

ServiceIndicationto User

Figure 13-1 Overall idle mode process

The concepts in the figure as well as the overall idle modeprocesses are explained in the following sections.

PLMN Selection

General

The MS selects a PLMN when it is powered on or uponrecovery from a lack of coverage. It first tries to select andregister on the registered PLMN (if one exists). If there is noregistered PLMN or if the registered PLMN is unavailable, theMS tries to select another PLMN either automatically ormanually depending on its operating mode.

The MS normally operates on its home PLMN. However,another PLMN may be selected if, e.g. the MS loses coverage.The MS registers on another PLMN if the MS finds a suitablecell to camp on and if a location updating request is accepted.Registration must be successful in order for the MS to be able toaccess that network, however, it does not need to performlocation updating if it is in the same location area belonging tothe same PLMN as it was before becoming inactive state.


– 156 – EN/LZT 123 3314 R3A

The MS can select and register on another PLMN in its homecountry if national roaming is permitted. In this case, the MSattempts periodically to return to its home PLMN.

At power on, the MS selects a registered PLMN regardless ofselection mode. However, if there is no registered PLMN themobile acts according to the selection mode, which can be set bythe mobile subscriber. There are two modes for PLMNselection: automatic and manual. The automatic mode utilizes alist of PLMNs in an order of priority; the manual mode lets theuser decide by indicating which PLMNs are available.

Automatic Mode

In automatic mode, the MS selects a PLMN (if available andallowable) in the following order if no registered PLMN existsor is available:

1. Home PLMN

2. Each PLMN that has been stored in the Subscriber IdentityModule (SIM) in priority order

3. Other PLMNs with received signal level above −85 dBm inrandom order

4. All other PLMNs in order of decreasing signal strength

Manual Mode

In manual mode, the MS first tries to select the registered PLMNor home PLMN if no registered PLMN exists. If this registrationfails or if the user has initiated a PLMN reselection, the MSinforms the user of all available PLMNs. The user can thenselect a desired PLMN which causes the MS to initiate aregistration on this PLMN. If the selected PLMN is notpermitted, the user is told to select another PLMN.

The user can at any time request the MS to initiate reselectionand registration to an alternative available PLMN. This is doneeither using automatic or manual mode (depending on the modeselected by the user).


EN/LZT 123 3314 R3A – 157 –

Cell Selection

General

The cell selection algorithm tries to find the most suitable cell ofthe selected PLMN according to specific requirements. If nosuitable cell is found and all available and allowable PLMNshave been tried, the MS tries to camp on a cell irrespective ofPLMN identity and enter a limited service state enabling it tomake only emergency calls (if necessary). If the MS losescoverage, it returns to the PLMN selection state and selectsanother PLMN.

Two different strategies can be used during cell selection: storedlist cell selection or normal cell selection. Stored list cellselection utilizes a stored BCCH Allocation (BA) list to speedup the cell selection procedure; normal cell selection isperformed when no such list is available.

Algorithm

1RUPDO�FHOO�VHOHFWLRQ

During normal cell selection, the MS tries to select the mostsuitable cell to camp on.

A cell is considered suitable if:

• it belongs to the selected PLMN

• it is not barred (Note: When a cell is barred it will not becamped on by an MS in idle mode, however, an active MScan perform handover to it)

• it does not belong to a location area which is in the list of“forbidden location areas for national roaming”1

• the cell selection criterion is fulfilled

1 Only valid for mobiles supporting GSM phase 2. National roaming may be allowed only to certain location areas of a another PLMN than thehome PLMN. Those location areas that are forbidden will, after an attempt to do a location updating has failed, be stored in the SIM asforbidden location area for national roaming. This list will be cleared when the mobile is powered off or the SIM is removed.


– 158 – EN/LZT 123 3314 R3A

When the MS has no information on which BCCH carriers areused in the network, it follows the procedures in Figure 13-2.

Scan 124 RF channels and measure signal strength for 3-5 seconds

Tune to the RF channel with the high-est received average signal strength

Determine if it is a BCCH carrier bysearching for frequency correction

bursts

Tune to the RF-channelwith the highest signal

strength not already tried

Is it a BCCHcarrier?

The MS shall attempt to synchronizeto this carrier and read BCCH info

Is it the wantedPLMN?

Is the cellbarred for access?

Is C1 > 0?

Camp on this cell

Yes

Yes

No

Yes

No

Yes

No

No

Figure 13-2 Normal cell selection

&HOO�VHOHFWLRQ�FULWHULRQ

While in idle mode, the MS continuously calculates the cellselection quantity, &12. The cell selection criterion is satisfied if&1 > 0.

The quantity &1 is calculated as follows:

&1 = (>UHFHLYHG�VLJQDO�OHYHO@ − $&&0,1) − max (&&+3:5 −3, 0)

2 The name of this quantity in the GSM Technical Specifications is “path loss criterion parameter”. As the C1-criterion is based only on signalstrength and not on path loss, the term used in this document is “cell selection quantity”.


EN/LZT 123 3314 R3A – 159 –

where:

$&&0,1 is the cell parameter that indicates the minimumreceived signal level at the MS required for accessing thesystem.

&&+3:5 is the cell parameter that indicates the maximumtransmitting power that an MS is allowed to use when accessingthe system.

3 is the maximum output power of the MS according to itsclass.

Cell Reselection

Algorithm

&HOO�UHVHOHFWLRQ�PHDVXUHPHQWV

After a cell has been successfully selected, the MS starts the cellreselection tasks. It continuously makes measurements on itsneighboring cells to initiate cell reselection if necessary.

The MS continuously monitors all neighboring BCCH carriersas shown in the BA list and the serving cell BCCH carrier todetermine if it is more suitable to camp on another cell. At leastfive received signal level measurement samples are required foreach defined neighboring cell. A running average of the receivedsignal level is maintained for each carrier in the BA list.

All system information messages sent on BCCH must be read atleast once every 30 seconds in order to monitor changes in cellparameters. The MS also tries to synchronize to and read theBCCH information that contains parameters affecting cellreselection for the six strongest non-serving carriers (in the BAlist) at least every five minutes.

The MS also attempts to decode the %6,& parameter for each ofthe six strongest surrounding cells (at least every 30 seconds) toconfirm that it is still monitoring the same cells. The %6,&parameter consists of two parts. The Network Color Code(NCC) and the Base station Color Code (BCC). If another %6,&is detected, it is treated as a new carrier and the BCCH data forthis carrier will be determined. If the MS detects a PLMN colorcode that is not permitted, the carrier is ignored.

The MS only takes measurement samples when listening to itsown paging group. The rest of the time it is in sleep mode.


– 160 – EN/LZT 123 3314 R3A

Figure 13-3 summarizes how often the BSIC and the BCCH datamust be decoded for the serving cell and neighboring cells whilein idle mode.

BSIC BCCH data

Serving cell - At least every 30 s

Six neighbors At least every 30 s At least every 5 min

Figure 13-3 Decoding of BSIC and BCCH data

&HOO�UHVHOHFWLRQ�FULWHULD

In order to control the traffic distribution between cells, theoperator can favor certain cells in active mode. Examples of thisare Locating and Hierarchical Cell Structures (HCS). In somesituations, there can be a need for a similar behavior in idlemode. Additionally, in a microcell environment, there is a needfor controlling the cell reselection rate specifically for fastmoving mobiles. For these purposes, additional cell reselectionparameters &52, 72, and 37 are broadcast on the BCCH ofeach cell. Before an MS can change cells to camp on, it mustread the offset that should be applied in the cell reselectionalgorithm from the BCCH of the potential target cell.

The cell reselection algorithm consists of five different criteria.If any one of the criteria is satisfied it causes a cell reselection tooccur. The cell reselection process employs a cell reselectionquantity &23. Whenever a cell reselection criterion is satisfied,the MS changes to the cell with the highest &2 value.

&2 is calculated as follows:

&2 = &1 + &52 − 72 * +( 37 − 7) for 37 ≠ 31(2)

C2 = &1 − &52 for 37 = 31(3)

where:

&1 is defined by equation 1,

+ [[

[( ) =

<≥

0 0

1 0,

3 The cell reselection quantity, C2, is only valid for MSs that support GSM phase 2. Phase 1 MSs will use C1�for cell reselection instead of C2.


EN/LZT 123 3314 R3A – 161 –

7 is a timer and

&52, 72, and 37 are parameters

The MS continuously calculates the value of &1 and C2 for theserving and neighboring cells. It reselects and camps on anothercell if any of the following criteria are satisfied:

• The serving cell becomes barred.

• The MS has unsuccessfully tried to access the network theallowed number of times.

• The MS detects a downlink signaling failure.

• &1 for the serving cell falls below zero for a period of fiveseconds which indicates that the path loss to the cell hasbecome too high and that the MS needs to change cell.

• The value of &2 for a non-serving cell exceeds the value of&2 for the serving cell for a period of five seconds.

Location Updating

General

To make it possible for the mobile subscriber to receive a call,the network needs to know where the MS is. The system isinformed about where the MS is located at regular intervals viathe Location Updating function.

There are three different types of location updating defined:

• Normal location updating

• Periodic registration

• IMSI attach

The MS may also inform the network when it enters an inactivestate known as IMSI detach.

Normal

Normal location updating is initiated by the MS when it detectsthat it has entered a new location area. When the MS is listeningto the system information transmitted on the BCCH carrier forthe serving cell, it compares the broadcast Location AreaIdentity (LAI) with the one stored in the MS. If the broadcastLAI differs from the one stored, a location updating type normalwill be initiated and the new LAI will be stored on the SIM. If


– 162 – EN/LZT 123 3314 R3A

the location updating fails, e.g. due to entering of a forbiddenlocation area, the MS either tries to select another cell or returnto the PLMN selection state.

Periodic Registration

To avoid unnecessary paging of a mobile that has left thecoverage area or has run out of battery power, there is a type oflocation updating called periodic registration.

When the MS listens to the system information on the BCCHcarrier, it is informed if periodic registration is used in that celland how often it must inform the network that it is still attached(reachable). This is controlled by the 7�� parameter.

IMSI Attach/Detach

The IMSI attach/detach operation is an action taken by an MS toindicate to the network that it has entered into an active/inactivestate. When an MS is powered on, an IMSI attach message issent to the MSC/VLR. When an MS is powered off, an IMSIdetach message is sent. A flag is set in VLR in order to indicatethe present state of the MS. This prevents unnecessary paging ofpowered off mobiles. The $77 parameter, broadcast in thesystem information messages, informs the MS whether or not itis requested to send a message to the system every time it isturned on or off.

The MS may also be marked as detached (implicit detach) by theMSC. This happens when there has been no successful contactbetween the MS and the network for a time determined by atime-out value (%7'0) plus a guard period (*7'0). Thesupervision time is the sum of these two periods. The base timeduration (%7'0) must be coordinated with the periodiclocation updating time in the interworking BSC (7��).Otherwise, the mobile will be unexpectedly removed from thesystem before a periodic location updating is performed.


EN/LZT 123 3314 R3A – 163 –

Combinations of Control Channels

Only certain combinations of control channels are allowed. Thefollowing three types of BCCHs are available:

Non-combined: BCCH and CCCH4

Combined: BCCH, CCCH, andSDCCH/4

Combined including a CBCH: BCCH, CCCH, SDCCH/4,and CBCH (the CBCHreplaces SDCCH subchannelnumber 2)

There are also four combination types for SDCCH:

SDCCH/8: Each physical channel consists of eightSDCCH sub-channels. That is, eight MSscan be assigned dedicated channels at thesame time.

SDCCH/8including a CBCH: The CBCH sub-channel replaces SDCCH

sub-channel number 2. Only seven MSscan thus be assigned dedicated channelssimultaneously.

SDCCH/4: This is a combination of four SDCCHsub-channels with BCCH and CCCH.Four MSs can be assigned dedicatedchannels at the same time.

SDCCH/4including CBCH: This combination consists of three

SDCCH sub-channels, BCCH, CCCH anda CBCH sub-channel. The CBCH sub-channel replaces SDCCH sub-channelnumber 2. Only three MSs can thus beassigned dedicated channelssimultaneously.

Only one CBCH can be specified. That is, either the channelcombination SDCCH/4 (including a CBCH) or SDCCH/8(including a CBCH) can be defined in a cell.

4 CCCH is a combination of PCH, AGCH and RACH.


– 164 – EN/LZT 123 3314 R3A

Paging

Paging Groups

After an MS tunes to the BCCH carrier and decodes the systeminformation data, it performs an evaluation, taking into accountthe IMSI number that determines the paging group to which itbelongs. The particular method by which an MS determines towhich paging group it belongs, hence, which particular CCCHblock of the available blocks on the paging channel that is to bemonitored, is defined in the GSM specifications. All MSs thatare listening to a particular paging block are defined as being inthe same paging group. When there are no paging messages tobe transmitted to MSs in a certain paging group dummy pagingsare sent instead.

The MS stays in sleep-mode to minimize power consumption inthe time gap between when its own paging group occurs. This isillustrated in Figure 13-4 However, the MS must still read theBCCH data sent by the serving cell at least once every 30seconds.

Sleep mode

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

Measuringon neighbors Sleep mode

Listeningto PCH

F S F S F SB C C

BCCH + CCCH(downlink)

TDMAframes

C

F (FCCH): Frequency Correction ChannelS (SCH): Synchronisation ChannelB (BCCH): Broadcast Control ChannelC (CCCH): Common Control Channel;

Paging Channel (PCH) orAccess Grant Channel (AGCH)

Measuringon neighbors

Listeningto PCH

Listeningto PCH

}

Paging group

Figure 13-4 Idle mode measurements during own paging group


EN/LZT 123 3314 R3A – 165 –

Paging Strategies

The paging procedure in GSM is managed by the MSC.Different paging strategies are possible, e.g. not sending secondpaging or sending the second paging as global paging, i.e. withina whole MSC service area. An operator can control the pagingprocedure with parameter settings in the MSC.


– 166 – EN/LZT 123 3314 R3A

LOCATING

INTRODUCTION

The locating algorithm is the software algorithm that providesthe basis for handover decisions. It is implemented in softwarein the BSC. It can be controlled via parameters and works outthe cell selection for active mobile stations.

Locating serves as the basis for other radio network featureswhich will be mentioned later. These include hierarchical cellstructure (HCS), overlaid/underlaid subcells, intra-cellhandover, assignment to another cell, and cell load sharing.

The locating features enables an operator to offer two basicbenefits to end-users:

• Quality and continuity of calls

• “Cell size” control in order to minimize total interference inthe network

The input to the algorithm is signal strength and qualitymeasurements from the mobile and from the base stationcurrently serving the connection. The output is a list of cells thatthe algorithm judges to be possible candidates for handover. Thecells in the list are ranked and sorted in descending order ofpreference for handover.

The algorithm works continuously, completing a calculationcycle approximately every 480 ms. The algorithm recommendsnot to perform a handover most of the time.

There are several reasons why a handover should berecommended:

• Field strength relations between the current carrier and theneighboring cells’ carriers (the cell with the highest signalstrength of the lowest path loss is selected)

• Signal quality (when the bit error rate is too high, a handoveris suggested)

• Timing advance (as the mobile station moves to the cellborder, the locating algorithm proposes a handover when thetiming advance threshold value is exceeded)

• Other radio network features


EN/LZT 123 3314 R3A – 167 –

Another type of measurement is based on the signal when it isdecoded in the channel decoder (frame error rate). Thesemeasurements are the input to a separate algorithm (“leakybucket”) which is used for recommending disconnection at lowtransmission quality in the base station and the MS, respectively.

There is a large number of parameters in the locating algorithm.The majority of these relate to quantities connected to individualcells and cell-to-cell neighbor relationships. The purpose ofthese parameters is to adapt the locating algorithm to the realityof the cellular network. In the final analysis, it is the locatingalgorithm with its parameter settings, that puts into practice thecell structure that has been planned by the operators for theircustomers.

BACKGROUND

A mobile telephony connection must be handed over betweencells as the person using the phone moves around. There areseveral criteria that can be used for making a decision about ahandover. The criteria serve different purposes, which in turnarise from a range of requirements placed on a mobile telephonysystem.

The basic requirements are coverage, speech quality, andcapacity. Therefore, the purposes of the criteria are to provide aconnection with the maximum signal strength obtainable at eachtime (coverage), to avoid disturbances (speech quality), andmaximize the C/I ratio (speech quality and capacity).


Precise Handover Borders

The handover decisions are based on signal strength and signalquality measurements performed by the mobile station, so-calledMobile Assisted Hand-Over (MAHO). The mobile stationmeasures the strength of the radio signal from the base stationthat serves the connection. In addition, it measures the signalstrength of the BCCH signals from the surrounding cells. Thesemeasurements are used in a comparison in order to find the“best” server.

The advantage of using comparisons (as opposed to fixedhandover thresholds) is that the handover border is fixed inspace and independent of the direction that the mobile is


– 168 – EN/LZT 123 3314 R3A

moving. A safety margin against fluctuating handovers (ahysteresis) can be adjusted to suit the needs of the network.There are a number of reasons for such fluctuations, e.g.

• Measurement noise insufficiently filtered

• Fading due to movements of the mobile station ormovements of objects in the surroundings

A low hysteresis yields a sharp handover border but a largeramount of fluctuating handovers. The amount of tolerableoscillating handovers must be balanced against such things assignaling times, handover failure probability, and probablesystem load.

In addition to hysteresis, timers regulate the minimum timeallowed between handovers.

The advantage of using the mobile station as the measurementprobe for the measurements used in the comparisons is thatsystematic errors in the measurement devices are canceled.Thus, the handover borders are constant and independent of themobile station. However, handover borders may appeardifferently for different mobiles due to inaccuracies in the outputpower.

Handover Borders Adapted to the Radio Environment

The signal strength measurements provided by a mobile to theBSC allows comparisons between the serving cell and theneighboring cells which the mobile monitors. The comparisonscan be made with a signal strength criterion (signal strengthmode) and with a path loss criterion (path loss mode).

In the signal strength mode, the handover borders are influencedif the output Effective Radiated Power (ERP) of one or severalbase stations changes. An increase in the output power in onecell thus means an increase in the area of that cell. However, ifthe output power changes equally everywhere, the borders arenot affected.

In the path loss mode, it is the path loss from the serving basestation and each of the neighboring base stations that are used inthe algorithm. The path loss is calculated by subtracting thereceived signal strength from the ERP in each base station. Inthis way, the output powers in the different base stations do nothave any significance to the handover borders. Thus, if theoutput ERP of one or several base stations is changed


EN/LZT 123 3314 R3A – 169 –

independently of each other, the handover borders are notaffected.

Handover Borders Yielding Low Interference

The comparisons in the locating algorithm serve to find the cellwith the highest signal strength (signal strength mode, or Kmode) or the lowest path loss (path loss mode, or L mode). Bothmodes can have the effect that the C/I ratio in the system isincreased, as compared to handover borders based on fixedthresholds.

The effect in the signal strength mode relies on the fact that Ccan be seen as an approximation to C/I. Thus maximizing C foreach connection is approximately the same as maximizing C/Ifor each connection. This is true at least in a statistical sense.

The effect in the path loss mode relies on the same effect as thesignal strength mode, plus the fact that the path loss mode favorsbase stations of low output ERP. Less energy is emitted into theair, as compared to the signal strength mode, thus decreasing the“I” part of C/I. However, the path loss mode can result in areaswhere the C/I level is locally lower than it would have beenusing the signal strength mode.

Flexible Cell Planning

The base station output power for the frequency that carries theBCCH can be set to a different level than that for all otherfrequencies. This provides an opportunity to plan the BCCHfrequencies according to a different plan or strategy than theother frequencies.

All cell borders can be individually moved by cell-to-cell relatedhandover border offset parameters. Reasons for doing this canbe to adjust them to the topography or to the traffic situation e.g.to place a cell border in the middle of a lake or a mountain, toeliminate cell corners protruding over busy roads, or to movesome of the traffic in a heavily loaded cell to a neighboring cell.In addition, cell borders have individual hysteresis set by cell-to-cell definable parameters. Separate hysteresis values can be usedat low received signal strength (in the signal strength criterion)and at high signal strength (in the path loss criterion).


– 170 – EN/LZT 123 3314 R3A

Bad Quality Urgency Handover

Signal quality measurements are provided by the mobile(downlink) as well as by the base station (uplink). Qualitymeasurements are only provided for the radio path connected tothe serving cell. When poor quality is detected, the locatingalgorithm may propose a handover.

A handover may not be allowed to disrupt the C/I pattern.Therefore, an urgency handover will only be carried out if themobile station is close to the cell border.

Urgeny Handover at Excessive Timing Advance

In order for mobiles to be able to synchronize the transmissionof their bursts at the time the base station expects to receivethem, the time it takes for the radio signals to travel from themobile to the base must be taken into account. The proper timeoffset is found by the base station which sends an order to themobile to start transmitting its bursts in advance of what itwould have done if the speed of the radio signals had beeninfinite. This time interval is called “timing advance”. The GSMTDMA protocol allows for a maximum timing advance thatcorresponds to a base-to-mobile distance of approximately 35km.

The timing advance value calculated by the base station isavailable in the locating algorithm and is used as a measure ofthe base-to-mobile distance. In this way, a maximumgeographical cell radius can be specified. If exceeded, thelocating algorithm proposes a handover.

Auxiliary Radio Network FunctionsLocating is the basic algorithm for determining the best cell toserve a connection. However, locating also incorporatesalgorithms for a number of other radio network features, namelythose radio network features that involve a cell, subcell, orchannel change:

• Assignment to another cell

• Hierarchical cell structures

• Overlaid/underlaid subcells

• Intra-cell handover

• Extended range

• Cell load sharing


EN/LZT 123 3314 R3A – 171 –


Handover Control in the MSC

The MSC has a number of exchange properties in the MSC atintra- and inter-MSC handover. These properties include timesupervision of the switching in the group switch, handoverallowed or not allowed in certain situations, etc.

Measurement Procedure

Every 120 ms (once in a 26-frame multiframe) there is an idleframe in the TCH channel that allows a longer time for the MSto tune again to the measured ARFCNs and decode thesynchronization bursts.

The synchronization burst contains the %6,& which includes theNCC. If the MS can detect the synchronization burst and decodeit, it checks if the NCC is permitted as defined by the parameter1&&3(50. If it is not, the measured signal strength for thatfrequency is discarded. If the NCC is permitted, the measuredvalues are reported to the BSC.

Algorithms

Ericsson1 Overview

The Ericsson1 locating algorithm serves the purpose ofproviding a list of possible cell candidates (in descendingranking order) for handover. The channel allocation andhandover signaling is not considered part of the locatingalgorithm.

Ericsson1 consists of eight stages corresponding to the eightboxes in the main flow chart (Figure 13-5). The stages areprocessed roughly in chronological manner.


– 172 – EN/LZT 123 3314 R3A

Penalty list

Measurementreports

emptylist

Organizing the list

Basic ranking

Auxiliary radionetwork functions

evaluations

Allocation reply

Urgency conditions

Filtering

Initiations

Sending the list

Figure 13-5 Main flow of the Ericsson1 locating algorithm

,QLWLDWLRQV

A locating individual is created. If there has been a previouslocating individual handling the same connection, a penalty listcan be received.

)LOWHULQJ

Measured values are filtered (smoothed) by averaging of anumber of consecutive measurements for each type of value.

%DVLF�UDQNLQJ

A basic ranking list of cell candidates is prepared according tothe minimum level condition and the K and L locatingconditions.

8UJHQF\�FRQGLWLRQV

Two types of urgency conditions are evaluated: poor signalquality and excessive timing advance. The signal quality isevaluated on the uplink as well as on the downlink.


EN/LZT 123 3314 R3A – 173 –

$X[LOLDU\�UDGLR�QHWZRUN�IXQFWLRQV�HYDOXDWLRQV

The criteria for overlaid/underlaid subcell change, hierarchicalcell structures, intra-cell handover, assignment to another cell,extended range, and cell load sharing are evaluated.

2UJDQL]LQJ�WKH�OLVW

All cells are organized into one final candidate list according torules that are defined by the outcome of the urgency conditions,the overlaid/underlaid evaluations, the hierarchical cellstructures evaluations, intra-cell handover evaluations, and cellload sharing evaluations. Additional locating criteria may beapplied in order to remove unsuitable candidates.

6HQGLQJ�WKH�OLVW

The candidate list is sent to the central processor to be used forthe channel allocation.

$OORFDWLRQ�UHSO\

The outcome of the channel allocation determines the action. Ifsuccessful, the connection is transferred to another channel andthe locating processing is transferred to a new software process.If failure occurs, the connection is not transferred, however, anumber of safety measures are taken.

Ericsson3

2YHUYLHZ

The first three stages of the algorithm (output power correction,minimum signal strength evaluation and subtraction of signalstrength penalties) are performed in exactly the same way as forEricsson1. The ranking, however, is much simpler. Ericsson3takes only signal strength into account and does not considerpath loss.

5DQNLQJ

In the ranking algorithm an offset value and a hysteresis valueare used when ranking neighboring cells. The offset value isused to displace the cell border as compared to the borderstrictly given by signal strength. The hysteresis reduces the riskfor ping pong handovers.


– 174 – EN/LZT 123 3314 R3A

Initiations

A locating individual is the software process that handleslocating and the auxiliary radio network functions. It is activatedas a result of immediate assignment, assignment, or handover.This means that all types of connection, traffic and signaling, arehandled by locating. Handover during the signaling part of theconnection may be inhibited.

At a change of channel (assignment, handover, subcell change,and intra-cell handover) a new locating individual is created andtakes over the handling of the connection. The old locatingindividual is terminated.

If the new locating individual was activated as a result of ahandover, a list of penalties is transferred from the old locatingindividual. Limited penalty information is also transferred aturgency handover to a cell in another BSC.

Immediately after an assignment, a handover or a subcellchange, it is desirable to remain on the same channel for a while.The reason is that the filtering of measurements needs sometime to produce reliable estimates on which to base furtheraction. Therefore, at initiation of a locating individual, a timer isstarted. The timer inhibits further action until it expires.

Filtering

0HDVXUHPHQW�SUHSDUDWLRQ

Locating is based on a number of quantities that are reported tothe BSC (Figure 13-6).


EN/LZT 123 3314 R3A – 175 –

Data description Source

signal strength downlink own cell full set MS

signal strength downlink own cell subset MS

signal strength downlink six strongestneighbors

MS

quality downlink own cell full set MS

quality downlink own cell subset MS

quality uplink own cell full set BTS

quality uplink own cell subset BTS

timing advance BTS

MS power capability (according to classmark) BSC(MS)

DTX used by base station BTS

DTX used by mobile MS

Figure 13-6 Data that is used for the locating evaluations

The signal strength, signal quality, and timing advancemeasurements are made and reported once for each SACCHperiod, i.e. every 0.48 s.

The mobile can measure signal strength for up to 32 neighborcells but can only report the six strongest in each measurementreport.

The signal strength measurements from the serving cell and thequality measurements (from the MS as well as from the BTS)are available in two sets: the full set and the subset. The full setis measurements taken from all bursts in the entire SACCHperiod. The subset is measurements taken from bursts wheretransmission is guaranteed even when DTX is active. Thelocating algorithm selects either the full set or the subset. Ingeneral, the full set is used if DTX has not been used during themeasurement period (SACCH period), and the subset if DTXhas been used during the period.

All signal strength measurement reports are delivered as integervalues from 0 to 63, corresponding to signal strength of –110dBm to –47 dBm. Measurement values above –47 dBm are setto 63, and below –110 dBm are set to 0.

The quantity used as the measure of quality is the Bit Error Rate(BER) that is estimated in the signal decoding process mappedon a logarithmic scale. The quality measurement reports aredelivered from the MS and the BTS as an integer value between0 and 7, where 0 corresponds to good quality (low BER) and 7


– 176 – EN/LZT 123 3314 R3A

to bad quality (high BER). The locating algorithm transformsthese values linearly to a 0 to 70 scale.

The timing advance values are delivered from the BTS as valuesfrom 0 to 63 bits, where 63 corresponds to a BTS-MS distanceof approximately 35 km.

The BSC receives the measurement results and stores them.Neighbor cells from which a signal strength measurement reporthas arrived are marked “valid”. These neighbors are theneligible for ranking.

6LJQDO�VWUHQJWK�DQG�TXDOLW\�ILOWHULQJ

The signal strength and quality measurement values that arrivelatest are filtered in order to smooth out measurement noise. Inaddition, some fading components of a duration of about thesame as the filter response time are filtered out.

Five types of filter are available:

• General FIR filters

• Recursive straight average

• Recursive exponential

• Recursive 1:st order butterworth

• Median

7LPLQJ�DGYDQFH

One single timing advance filter is used for all cells in the BSC.The filter is a straight averaging filter.

Basic Ranking

2YHUYLHZ

The purpose of basic ranking is to produce a list of all neighborcell candidates ranked in order of preference. The basic rankingprocedure consists of seven stages:

• Correction of base station output power for downlinkmeasurements

• Evaluation of the minimum signal strength condition

• Subtraction of signal strength penalties

• Evaluation of the sufficient signal strength condition


EN/LZT 123 3314 R3A – 177 –

• Signal strength evaluation (K-criterion)

• Path loss evaluation (L-criterion)

• Combination of a basic ranking list

&RUUHFWLRQ�RI�EDVH�VWDWLRQ�RXWSXW�SRZHU�IRU�GRZQOLQNPHDVXUHPHQWV

When the mobile station measures the signal level from aneighbor cell, that cell transmits on a control channel (BCCH).This channel may have a different output power (ERP) than thetraffic channel that will be seized in the case of a handover. Thelocating algorithm, however, aims at controlling cell borders asthey are for channels on traffic frequencies. Therefore, the signalstrength measurements U[OHY for all cells (including serving cell)are corrected for the difference in ERP on the transmittertransmitting on the BCCH frequency %63:5 and thetransmitter transmitting on the other frequency or frequencies,%67;3:5.

66B'2:1P

= U[OHYP

+ %67;3:5P

– %63:5P

, (1)

where P refers to neighbor cells and serving cell.

Serving cell correction is only performed if necessary, that is, iftransmission occurs on the BCCH frequency. The correction forneighbor cells is always done, even though a traffic channel on aBCCH frequency may be seized after the handover.

0LQLPXP�OHYHO�FRQGLWLRQ�IRU�QHLJKERU�FHOOV

The output from the neighbor cell signal strength filtering, U[OHY,is tested against two minimum level thresholds. One thresholdfor the downlink signal strength and one for an estimated uplinksignal strength. These levels are defined for each cellindividually. The weakest of uplink and downlink signal levelsis used for the minimum level criterion selection. This is to seeif the signals can be considered high enough above thesensitivity level for the corresponding cells to be in question forhandover.

Neighbor cells fulfilling the uplink as well as the downlinkminimum level condition are eligible for further processing.


– 178 – EN/LZT 123 3314 R3A

3HQDOWLHV

• Penalties, or punishments, aggravate handover to cells thatfor some reason are temporarily undesirable. Theaggravation consists of subtracting a signal strength value (apenalty) from the signal strength estimate U[OHY for theundesirable cell so that it appears “worse” than it really is.

A locating penalty is given by one of the following threereasons:

• Handover failure:

If there is a signaling failure at handover, the failure mightbe repeated if a handover to the same cell is attempted toosoon. Therefore, the cell to which the handover failed ispunished.

• Bad quality urgency handover:

The cell that is abandoned due to bad quality urgency isnormally the best cell from a signal strength point of view.This means that without precautions, it would head thecandidate list at the next locating evaluation and thus cause ahandback.

• Excessive timing advance urgency handover:

Excessive timing advance urgency handover is similar to badquality urgency handover. It is handled in the same manner,with its own set of parameters. However, when a cell ispunished due to excessive timing advance, all other co-sitedneighboring cells are tested for excessive timing advanceand punished if necessary.

• Temporary penalty:

This is a punishment which is associated with the featurehierarchical cell structures.

Each penalty is applied to the punished cell for a specific timeknown as the penalty duration. It is applied only for theconnection that experienced the handover failure, the urgencycondition, or the condition related to hierarchical cell structures.

Before cell ranking, a check is performed to see whether thereare any penalties for the neighbor cells in question. If a cell has apenalty (and if the associated penalty duration has expired), thatpenalty is removed.


EN/LZT 123 3314 R3A – 179 –

6XIILFLHQW�OHYHO�FRQGLWLRQ

For all neighbor cells selected according to the minimum levelcondition (plus serving cell) a sufficient level condition isapplied. This condition separates cells for which a high signalstrength has been reported (“high signal strength cells”) fromthose for which a low signal strength has been reported (“lowsignal strength cells”). The condition can also be seen as a wayto separate areas in the cells into “high signal strength areas” and“low signal strength areas”.

The minimum signal strength level and the sufficient signalstrength level can be seen as delineating areas around the basestation. Figure 13-7 shows an example of how the minimum andsufficient levels can appear in an idealized geographical plane(i.e. in a flat geography without shadow fading).

A

B

uplink minimum level

downlink minimum level

uplink effective sufficient level

downlink effective sufficient level

“high signalstrength”

“low signalstrength”

Figure 13-7 Minimum and sufficient levels around a basestation

The solid lines represent the final (i.e. the combined uplink anddownlink) conditions. The sufficient level shown is valid onlyfor the neighbor cell B.

As described previously, the sufficient level for a serving cellcan be different in relation to different neighbor cells. Theevaluation of the sufficient level condition for serving cell willbe related to the best neighbor cell (the neighbor with thehighest ranking in the basic ranking list, see next section).Therefore, the neighbor cells must be ranked before thisevaluation can take place.


– 180 – EN/LZT 123 3314 R3A

5DQNLQJ�DFFRUGLQJ�WR�WKH�VLJQDO�VWUHQJWK�FULWHULRQ��.�FULWHULRQ�

Cells that do not fulfill the sufficient level condition (the “lowsignal strength” cells) are called K-cells, and are rankedaccording to a relative signal strength ranking (the K-criterion).Cells that do fulfill the sufficient level condition (the “highsignal strength” cells) are called L-cells. L-cells are consideredgood enough to be eligible for a path loss ranking (the L-criterion).

The K-criterion is a relative signal strength criterion, since theK-values are signal strength relative to the sufficient level.

The total effective K-value for a serving cell is calculated as thelowest of the uplink and downlink K-values.

For neighbor cells, the total effective K-value is also the lowestof the uplink and downlink K-values but modified with an offsetand a hysteresis.

The hysteresis�is used to decrease the ranking values forneighbor cells which become somewhat underrated incomparison to the serving cell. This is to prevent ping-ponghandovers. The hysteresis is defined as a cell-to-cell relation,and is always symmetrical (i.e. the same value for twoneighboring cells).

The offset is used to change the ranking value. This has theeffect that the cell border is displaced away from the cell forwhich the parameter has a positive value. It is defined as a cell-to-cell relation and is always anti-symmetrical (i.e. the samevalue but different sign in two neighboring cells).

The function of hysteresis and offset is illustrated in Figure 13-8in a theoretical signal strength diagram, where it is assumed thatthe sufficient values are equal in cell A and cell B.


EN/LZT 123 3314 R3A – 181 –

handover border, B to A

signal strength from cell A

hysteresis nominal cell border

offset

original cell borderwithout offset

handover border, A to B

signal strength from cell B

hysteresis corridor

hysteresis

Figure 13-8 Hysteresis and offset

After a handover, the serving cell becomes the neighbor cell andvice versa. The nominal cell border in Figure 13-8 remains inthe same position, but the old handover border (the one from Ato B in the figure) is replaced by the handover border from B toA. Thus an area (or a corridor) is created around the nominalcell border (the shaded area), i.e. the so-called hysteresiscorridor. In this area, the connection can belong to either cell.

Figure 13-9 shows an example of how a cell border, togetherwith its hysteresis corridor, can be moved with the offsetparameter. The figure shows handover borders in a geographicalplane in a more realistic manner.

offset

original cell border

hysteresiscorridor

AB

nominal cell border

Figure 13-9 Handover borders, hysteresis, and offset


– 182 – EN/LZT 123 3314 R3A

K-cells are ranked according to their K-value. The highest K-value ranks first.

5DQNLQJ�DFFRUGLQJ�WR�WKH�SDWK�ORVV�FULWHULRQ��/�FULWHULRQ�

L-cells are ranked according to their path loss, calculated as thedifference between base station ERP and the received signalstrength in the mobile. The path loss criterion is independent ofmobile as well as base station power ratings. At a certainlocation, every mobile will then evaluate the path loss equally,irrespective of power class or base station power changes. Thisis not the case with the relative signal strength criterion.Thereby, calls are transferred from a big cell (which causesstrong interference) to a small cell causing weaker interferencedue to its lower power rating. Thus, the usage of the path losscriterion is expected to lower the total statistical interferencelevel over the entire network. If two cells have equal ERP, theK- and L-criterion give the same ranking result with respect toeach other.

For serving L-cell, the effective L-value is equal to the path loss.

For neighboring cells, the effective L-value is modified with anoffset and a hysteresis in the same manner as the K-rankingvalue, i.e. L-cells are ranked according to their L�value, and thelowest L-value ranks first.

The offset and hysteresis�have the same symmetry properties astheir K-counterparts, that is, anti-symmetry and symmetry,respectively. They are, thus, used to control cell borders thatappear when the serving cell and the strongest neighbor cell (theone the handover is usually performed on) are both L-cells.

5DQNLQJ�RI�VHUYLQJ�FHOO

At this point, the rankings of neighboring cells are known. Inorder to rank the serving cell, it must be determined whether it isa K- or L-cell. As with neighboring cells, this is done byevaluating the sufficient level condition.

If the serving cell fulfills the sufficient level condition, it is anL-cell.

If the serving cell is a K-cell, it is ranked among the other K-cells. If the serving cell is an L-cell, it is ranked among the otherL-cells.


EN/LZT 123 3314 R3A – 183 –

%DVLF�UDQNLQJ�OLVW

Finally, the basic ranking list is compiled by joining together L-cells and K-cells. L-cells are put on top, with the cell having theranking value corresponding to the lowest path loss first. K-cellsare put at the bottom, with the cell having the ranking valuecorresponding to the lowest relative signal strength last. Thismeans that an L-cell is always ranked better than a K-cell nomatter what the original signal strength values are that are usedin the calculations.

An overview of the basic ranking procedure is given inFigure 13-10.


– 184 – EN/LZT 123 3314 R3A

Minimumlevel condition

fulfilled?

Penalty evaluation

Sufficientlevel condition

fulfilled?

Rank n according toK-critierion

Rank n according toL-critierion

L-cells

K-cells

Highest L-rank

Lowest L-rankHighest K-rankLowest K-rank

For all reported neighbor cells n

Discard nNo

Yes

Yes No

moreneighbors?

Wait untilranking ofneighborcells isdone

Sufficientlevel condition

fulfilled?

For serving cell s

L-cells

K-cells

Yes No

Serving cellis K-cell

Serving cellis L-cell

No

Yes

Figure 13-10 Overview of the basic ranking procedure


EN/LZT 123 3314 R3A – 185 –

Urgency Conditions

Two criteria are used for urgency detection: excessive timingadvance and bad quality.

The information about bad quality and excessive TA urgency isused in the organization phase to indicate when the cell needs tobe abandoned urgently. In those cases, it is allowed to perform ahandover to a cell that has a lower basic ranking value than theserving cell. However, it is not allowed to make a bad qualityurgency handover to a worse cell from anywhere in the servingcell. This might cause a call to be connected to a cell that is faraway from the correct cell according to the prediction that wasdone in the cell planning. Thereby, the call may cause excessiveuplink co-channel interference to another connection andcorrespondingly, may experience excessive downlink co-channelinterference. Therefore, an extra signal strength criterion isapplied.

An offset defines how far away from the nominal cell border amobile is allowed to be located in order to be eligible for badquality handover. It is applied to the K-ranking values and the L-ranking values.

The main cause of bad quality is normally co-channelinterference. However, it can also appear as a result of adjacentchannel interference, excessive time dispersion, or low signalstrength. In all cases, a handover may lead to an improvement ofthe connection quality. If no suitable handover candidate isfound, the call remains connected to the serving cell.

Timing advance can be used as a measure of the base-to-mobiledistance. This can thus be used as a soft cell limit. If a suitablecandidate is not found at excessive TA urgency, the call remainsconnected to the serving cell. An excessive TA urgency hasnone of the urgency region limitations that bad quality urgencyhas.

The radio conditions that gave rise to the urgency condition maywell prevail in the original cell. In addition, the original cellwould be the best cell from a strictly signal strength or path losspoint of view. In order to prevent an immediate handback to theoriginal cell after an urgency handover, the original cell ispunished with a penalty during a certain amount of time.


– 186 – EN/LZT 123 3314 R3A

Auxiliary Radio Network Functions

*HQHUDO

Six auxiliary radio network functions are incorporated in thelocating software:

• Assignment to another cell

• Hierarchical cell structures

• Dynamic overlaid/underlaid subcells

• Intra-cell handover

• Extended range

• Cell load sharing

Below are some brief descriptions of these auxiliary functions inconjunction with locating.

$VVLJQPHQW�WR�DQRWKHU�FHOO

The normal locating algorithm is used to find the most suitablecell at call connection. This is possible since locating is alreadystarted at immediate assignment when a signaling channel isestablished between the MSC and the mobile station. When atraffic channel is assigned (i.e. when the signaling is completedand the total path is established), locating has had theopportunity to evaluate the radio environment.

In the case where a better cell than the serving cell (i.e. the onethat provided the downlink idle connection) is found, that cellwill be the first one in the locating candidate list. This is called“Assignment to better cell”.

At this stage of the locating evaluations, cells worse than theserving cell may also remain in the candidate list. At congestionin the serving cell or in a better cell, the call can be set up inanother worse cell. This is called “Assignment to worse cell”.However, in the same manner as with bad quality urgency, cellsat a large radio distance from the nominal cell border are notallowed to remain as candidates. The reason for this extra signalstrength criterion is the same as with the bad quality urgencyregion (i.e. not to cause, and not to become subject to, excessiveco-channel disturbance).

Assignment to a better cell as well as assignment to a worse cellcan only be performed during a certain time after immediateassignment.


EN/LZT 123 3314 R3A – 187 –

+LHUDUFKLFDO�FHOO�VWUXFWXUHV

The hierarchical cell structures make it possible to organize cellsin up to three layers. One way of utilizing the hierarchical cellfeature is to apply it to a network with one microcell layer andone umbrella cell layer that covers the microcells.

The layer can also be seen as a priority designation. Layer 1 thenequals priority 1, which means the first priority. This impliesthat the feature hierarchical cell structures can have other usagethan the one mentioned above. It can be used if there is a needfor prioritizing certain kinds of cells, irrespective of the size ofthe cells. One important example is a combined GSM900 andGSM1800 system, where the two frequency bands can beassigned to different layers.

The strategy employed for layered cells is to direct traffic fromcells in a higher layer to cells in a lower layer wheneverpossible. Higher layer cells will only pick up traffic at call setupcongestion, in radio coverage holes, and at radio disturbances.

In order to implement this strategy, cells in different layers arenot ranked together. If they were, a strong umbrella cell wouldpick up all traffic in its vicinity, and the microcells under itwould not receive the traffic they were intended to serve.

'\QDPLF�RYHUODLG�XQGHUODLG�VXEFHOOV

This feature can be used to increase the traffic capacity in acellular network without adding new sites. It provides amechanism to cover temporary peaks in the traffic load in a cellwith a retained interference level.

The strategy behind the dynamic overlaid/underlaid subcell is toonly use the overlaid subcell as a last resort when the traffic loadin the underlaid subcell becomes too high

,QWUD�FHOO�KDQGRYHU

The intra-cell handover evaluations may result in arecommendation to change the channel within the subcell.

([WHQGHG�UDQJH

Extended range allows the use of cells with a radius ofmaximum 72 km (in contrast to the normal 35 km). Forextended range cells, the value range of those parameters thatconcern TA-related quantities is larger.


– 188 – EN/LZT 123 3314 R3A

&HOO�ORDG�VKDULQJ

Cell load sharing employs the locating calculations torecalculate the ranking value for certain neighbors. If the load isabove a given threshold in a cell, all active connections in thatcell become subject to load sharing recalculations.

After the basic ranking procedure in locating has beencompleted, the ranking values of selected neighbor cellcandidates are recalculated with a reduced hysteresis. Theselection of neighbor cells is based on the following criteria:

• The neighbor is a worse cell

• It belongs to the same hierarchical level

• It belongs to the same BSC

• The load in the neighbor cell is below a certain threshold

The hysteresis reduction applies to all three hysteresis. It isapplied in a gradual manner over a certain time, starting fromzero to a specified percentage of each normal hysteresis value.After the recalculations, the new ranking list replaces the old infurther treatment. The result is that a worse neighbor mayeventually change status and become better, thus initiating ahandover.

Organizing the List

*HQHUDO

Cells in the basic ranking list are divided into eleven categories.Cells within each category retain the ranking order they had inthe basic ranking list. The different categories are organized in asequence that is given by the combination of indications thatoriginate in locating and in the auxiliary radio network functionsevaluations. The selection of categories and the order in whichthey are put together reflect the functionality of the features andtheir priorities. The final cell candidate list can contain sevencandidates: maximum six neighbor cell candidates and oneserving cell candidate.

Information about the cause for the action is attached to eachcell candidate. For neighbor cells, the cause value indicates thehandover cause. For the serving cell the cause values indicatewhich action to perform: assignment, intra-cell handover, orsubcell change.


EN/LZT 123 3314 R3A – 189 –

&DWHJRULHV

The cells are divided into the eleven categories (Figure 13-11).The basis for the categorization of cells in the candidate list arethe following four criteria:

• Layer:

− ��±��, where layer 1 is the lowest layer (highest prioritycells)

• Higher or lower ranking than serving cell in the basicranking list:

− E (better) denotes cells with higher ranking

− Z (worse) denotes cells with lower ranking

• Signal strength above or below the threshold for layerchange:

− R (over) indicates that the layer threshold criterion isfulfilled

− X (under) indicates that the layer threshold criterion is notfulfilled

• Serving cell or neighboring cell:

− V denotes serving cell

− all other cells are neighbor cells.


– 190 – EN/LZT 123 3314 R3A

Description Category

Better layer 1 cells above their ownthreshold

1bo

Better layer 1 cells below their ownthreshold

1bu

Worse layer 1 cells above their ownthreshold

1wo

Worse layer 1 cells below their ownthreshold

1wu

Better layer 2 cells above their ownthreshold

2bo

Better layer 2 cells below their ownthreshold

2bu

Worse layer 2 cells above their ownthreshold

2wo

Worse layer 2 cells below their ownthreshold

2wu

Better layer 3 cells 3b

Worse layer 3 cells 3w

Serving cell s

Figure 13-11 Cell categories

3ULRULWLHV�RI�UDGLR�QHWZRUN�DFWLRQV

The way the categories are put together reflects a priorityscheme among the radio network actions (Figure 13-12). Priority1 is the highest priority. Figure 13-12 is valid for the assignmentphase as well as for an ongoing connection. In the assignmentphase, “Handover” means “Assignment to better cell”.

3ULRULW\ 5DGLR�QHWZRUN�DFWLRQ

1 Change to a lower cell layer

2 Excessive TA urgencyhandover

3 Handover to better cell

4 Overlaid/underlaid subcellchange

5 Intra-cell handover

6 Bad quality urgency handover

7 Change to a higher cell layer

Figure 13-12 Priority of radio network actions


EN/LZT 123 3314 R3A – 191 –

/LVW�SUHSDUDWLRQ

The cell categories are organized to realize the order of priority(Figure 13-12) to the maximum extent possible. This is done atall combinations of the indications that are given by locating andthe auxiliary radio network functions evaluations. Theindications that are used are listed in Figure 13-13.

Indication Description

1 Assignment request arrived

2 AW (Assignment to Worse cell)state

3 Excessive TA urgency detected

4 Bad quality urgency detected

5 Overlaid/underlaid subcellchange requested

6 Intra-cell handover requested

Figure 13-13 Indications used to organize the candidate list

In extensive tables (not included here) the completespecification is shown of how the candidate list is built from thecell categories (Figure 13-11) in the presence of the indications(Figure 13-13). The selection of categories and their orderingmust be different depending on which layer the serving cellbelongs to. Thereby, the priority scheme and the strategy todirect traffic to the lowest cell layer whenever possible can befulfilled. Within each category, the cells retain their rankingorder from the basic ranking list. An overview of thecompilation of the final handover candidate list is found inFigure 13-14.


– 192 – EN/LZT 123 3314 R3A

Cell ACell BCell C

Cell DServing cell

Cell ECell F

L-cells

K-cells

Better cells (b)

Worse cells (w)

Basic candidate list Cell load sharing

Recalculation with reduced hysteresis

Ordering into categories

1bo1bu1wo1wu2bo2bu2wo2wu3b3ws

Hierarchical cell structures

Layer 1 cell => 1,Layer 2 cell => 2,Layer 3 cell => 3,

above own treshold (o)below own treshold (u)

Handover candidatelist in ranking order

Locating indications

Assignment request arrivedAW stateExcessive TA urgency detectedBad quality urgency detectedOverlaid/underlaid subcell changeIntra-cell handover requested

Example: C(1bo), B(1wo), E(1wo), A(2bo), G(s)

Figure 13-14 Overview of the compilation of the final handovercandidate list


EN/LZT 123 3314 R3A – 193 –

Sending the List

The resulting candidate list forms the basis for cell selection. Ifthe result is an empty list, then there simply is no option betterthan remaining in the present cell. Furthermore, no list is sent tothe central processor.

The first cell in the candidate list is the “best” candidate. Ideallythis is the cell to which the mobile should be connected. If thereare no channels available in the “best” cell, the functionshandling the handover in the BSC make an attempt to allocate achannel in the next cell in the candidate list.

Allocation Reply

The reply from the channel allocation to the locating individualcontains information about the result of the channel allocation.This result can be either a success or a failure. The failure can beeither due to congestion or to a signaling failure. The resultdetermines a number of actions, such as setting of penalties andenabling of certain timers.

6XFFHVV

A successful allocation means that an available radio resourcewas found in one of the cell candidates and that the transfer ofthe connection to the new channel was successful. At this pointa new locating individual (connected to the new radio channel)has already been activated. The old locating individual has butone more task to perform before it is deleted. If the handoverwas due to an urgency condition, a penalty must be calculated inorder not to cause immediate handback, this penalty must betransferred to the new locating individual.

Disconnection Criteria

The disconnection algorithms are not part of the locatingalgorithm, but (for completeness) the topic is discussed here.

The downlink disconnection criterion is managed by the mobile.It is controlled by the parameter 5/,1.7, which the mobilereceives from the base station on the broadcast channel.

The algorithm is of the type “leaky bucket”, and bases thedecisions on successfully decoded digital speech or datainformation frames. There is a “bucket” initially containing anumber given by 5/,1.7, which is also the volume of the


– 194 – EN/LZT 123 3314 R3A

“bucket” (i.e. the maximum number of units that it can contain).When a frame has been received but not successfully decoded,the “bucket” counter is decreased by one unit. When a frame issuccessfully decoded, it is increased by only two units (up tomaximum 5/,1.7). If the “bucket” counter runs down to zero,the mobile terminates the connection.

The “leaky bucket” counter is reinitiated at handover,assignment, subcell change, and intra-cell handover.

In the uplink, the algorithm is the same but managed by the BSCand basing the decisions on successfully decoded frames in theuplink (i.e. in the base station). A similar parameter controls thealgorithm.

As an additional uplink disconnection criterion, a threshold forthe timing advance is used.


EN/LZT 123 3314 R3A – 195 –

CHANNEL ADMINISTRATION

When a connection is to be established, or a channel in use mustbe changed for some reason, a new channel must be chosen andallocated. As soon as there is more than one type of channelpossible to choose, the order of preference must be defined.Furthermore, if there are several idle channels among the chosentype of channel, there must be some kind of rule to select themost suitable one.

Channel administration is a feature which selects and allocates asuitable channel when a channel is required. The feature isimplemented in the BSC.

The features “Overlaid/Underlaid Subcells” and “ImmediateAssignment on TCH” have brought forth a need for being ableto choose different strategies regarding which type of channel toallocate in a specific traffic situation. The feature channeladministration provides the ability to choose between differentchannel allocation strategies.

Several other features have an impact on the channeladministration including idle channel measurements, differentialchannel allocation, and frequency hopping.

The channel allocation algorithm in the feature channeladministration selects and allocates a suitable channel in eachtraffic situation that requires a channel. If several types ofchannels are possible to allocate in a specific traffic situation,the order in which the different types are preferred must bedefined.

The operator can choose among seven pre-defined channelallocation strategies, defined as seven different CHannelAllocation Profiles (CHAP). A CHAP is a list of all possibletraffic situations where each traffic situation is assigned differenttypes of channels in order of preference.

The operator can choose whether non-hopping TCHs on theBCCH frequency are allocated as a first choice, last choice, orwith no preference. A channel is then selected with regard tointerference and the number of frequencies on which thechannels hop.

Before the selection and allocation of a channel, a cell has beenselected by the locating algorithm or the idle mode cell selectionmechanism. There are three main situations in which a channelis allocated:


– 196 – EN/LZT 123 3314 R3A

• Immediate assignment

When a connection is to be established, a channel forsignaling must be allocated. Depending on the trafficsituation and the chosen CHAP, the channel could be eitheran SDCCH or a TCH.

• Assignment

After an immediate assignment to SDCCH, if a channel forspeech/data is needed, a TCH must be allocated.

• Handover

When a channel in use is to be changed, a new channel forspeech/data or signaling must be allocated.

In order to select a suitable type of channel, the following data isneeded:

• Traffic case -

the traffic case in which a channel is needed

• Preferred subcell -

the preferred subcell, overlaid or underlaid, according to theoverlaid/underlaid evaluations in locating

• Channel mode -

the channel mode, speech/data or signaling, for which thechannel is to be used

There are four different types of channels that can be chosen inthe selected cell:

• TCHs in overlaid subcell (TCH/OL)

• TCHs in underlaid subcell (TCH/UL)

• SDCCHs in overlaid subcell (SDCCH/OL)

• SDCCHs in underlaid subcell (SDCCH/UL)

An SDCCH can be used for signaling only (e.g. locationupdating, call set-up, or SMS). A TCH can be used for eitherspeech/data or signaling.

Multiple speech coders are supported (i.e. more than one speechcoder may be present in the system). Three speech coders can beprovided: full rate, half rate, and enhanced full rate. The half ratespeech coder offers an increase in capacity, since two half rateair interface channels occupy the same bandwidth as one fullrate air interface channel. The enhanced full rate speech coderoffers an improved speech quality.


EN/LZT 123 3314 R3A – 197 –

The Transcoder and Rate Adapter (TRA) pool provides efficientsupport to multiple speech coders. It is a resource poolcontaining a number of transcoder resources, e.g. full ratetranscoder units, enhanced full rate transcoder units, or half ratetranscoder units. The purpose of the introduction of centralizedTRA pools is to provide a system solution with efficientutilization of the installed transcoder hardware.

The support of multiple speech coders is implemented in theMS, BSC, and in the MSC/VLR.


– 198 – EN/LZT 123 3314 R3A

DYNAMIC MS (BTS) POWER CONTROL


Battery Power Consumption

When MS power control is used, MS power consumption isreduced, recharging is needed less frequently, and the maximumpossible speech time increases.

If mains power supply for the BTS is impeded, there is a batteryback-up, and the battery consumption is reduced. Furthermore,the maximum possible speech time increases, if BTS powercontrol is used.

Receiver Saturation

The high signal energy from MSs (BTSs) that are close to a BTS(an MS) might saturate the receiver. If the output power of theconcerned MSs (BTSs) is limited, the risk for this kind of radiofrequency blocking is reduced. The receiver(s) might still beblocked if the MS and BTS are very close to each other but thetotal amount of time when this can happen is reduced.

There is an initial mode of the dynamic MS power controlalgorithm to handle the BTS receiver saturation problem duringcall set-up.

Interference

When the strategy to decrease the output power of the MS (BTS)is adapted to all MSs (BTSs) in the network, the total amount ofradiated power is reduced. This implies that the uplink(downlink) co- and adjacent channel interference in the networkis reduced. It has been shown that with the power controlalgorithm, the number of connections with very low C/I ratiodecreases. Obviously the C/I ratio is increased for connectionswhere the MS and BTS are far apart. These connections will notbe subject to a reduced output power but they experience areduced interference level.

On the other hand, the C/I ratio is decreased for connectionswhere the MS and BTS are close together when they aresubjected to a reduced output power. The C/I reduction does not


EN/LZT 123 3314 R3A – 199 –

affect the speech quality, since these connections have a marginto the lowest tolerable C/I ratio.

Quality Impact

Quality is considered when calculating the power orders. Qualityis the measured signal quality, U[TXDO. Bad quality is taken intoaccount by increasing the output power.


General

The objective of the dynamic MS (BTS) power controlalgorithm is to adjust the output power of the MSs (BTSs) sothat the received signal strength in the BTS (MS) decreases ifpath loss increases. The power range, where regulation ispossible, is limited by the transmitters. Thus, the regulation isonly effective in a specific part of the cell, i.e. the regulationarea (Figure 13-15). The MS (BTS) transmits with minimum ormaximum output power outside this area.

Regulation area

Site

Figure 13-15 Schematic picture of the power regulation area ina cell

The regulation area shown in Figure 13-15 is indicated inFigure 13-16 as well. The upper graph in Figure 13-16 showsthe MS (BTS) output power versus the path loss between an MSand a BTS. An MS (a BTS) is only capable of transmitting atdistinct power levels. The lower graph shows how the receivedsignal strength varies with the path loss.


– 200 – EN/LZT 123 3314 R3A

SS1

SSDES(SSDESDL)

Received signal strength in the BTS (MS)

Path lossRegulation area

Path loss

Output power of the MS (BTS)Maximumallowedpower level

Minimumallowedpower level

Regulation area

Figure 13-16 Mobile (base station) output power and receivedsignal strength in the base station (mobile station). Quality isnot taken into account.

Close to the BTS (MS), the MS (BTS) transmits at its lowestpossible power level. Although the BTS (MS) receives a signalthat exceeds the desired value, the MS (BTS) cannot reduce thetransmitted power any further. Conversely, far away from theBTS (MS), the MS transmits at the maximum allowed powerlevel for the cell. The power cannot be increased even if thereceived signal strength in the BTS (MS) is weak.

During a call, the BTS (MS) measures uplink/downlink signalstrength and uplink/downlink signal quality. Thesemeasurements are sent to the BSC where they are used whencalculating a new MS (BTS) output power level.


EN/LZT 123 3314 R3A – 201 –

Algorithm

General

Dynamic MS (BTS) power control is performed for TCHs aswell as for SDCCHs.

All channels on the BCCH frequency are transmitted on fullpower, i.e. there is no power control of these channels.

The MS can change its output power every 13th TDMA frame.This equals EIGHT times every SACCH period. Each change isin units of 2 dB steps. This means that the maximum change is 8∗2 dB = 16 dB during one SACCH period (480 ms).

The BTS can change its output power on a time slot basis. Theresolution of the power steps is 2 dB. The maximumconfigurable change is 30 dB. The maximum change perSACCH period is also 30 dB.

Figure 13-17 shows which types of measurements are used inthe dynamic MS (BTS) power control algorithm.

Data description Source

signal strength

uplink (downlink) full set1

BTS (MS)

signal strength

uplink (downlink) subset1

BTS (MS)

quality

uplink (downlink) full set1

BTS (MS)

quality

uplink (downlink) subset1

BTS (MS)

power level used by MS (BTS) MS (BTS)

DTX used by MS (BTS) or not MS (BTS)

Figure 13-17 Data that is a base for the dynamic MS (BTS)power control evaluation5

5The BTS (MS) performs signal strength and signal quality measurements on the uplink (downlink). Measurements are made on the full set offrames (full set), as well as on the subset of frames where there is always traffic (subset). Which of the sets that will be used depends onwhether the MS (BTS) has used DTX or not during the measurement period.


– 202 – EN/LZT 123 3314 R3A

The dynamic MS (BTS) power control algorithm consists ofthree stages:

1. Measurement preparation

Missing measurements are estimated and a decision is madeabout which set of measurements (full set or subset1) to use.A decision is also made whether compensation for frequencyhopping should be used or not.

2. Filtering of measurements

Measurements are filtered in order to eliminate variations oftemporary nature. That is, to ensure that the decision base forthe next power order is stable.

3. Calculation of power order

The power order to the MS (BTS) is calculated so that thedesired signal strength and quality is received in the BTS(MS). A number of constraints are applied to the calculatedpower order.

In addition, the Dynamic MS power control algorithm canoperate in two modes (or regulation phases). The two modes are:

• Initial regulation

The algorithm operates in this mode when a new channel isassigned. The purpose with a special initial mode is toreduce a high MS power level as quickly as possible.

• Stationary regulation

This is the normal mode of the algorithm. In this mode, thealgorithm does not need to be as fast as in the initial mode.In return, a more stable behavior is obtained.

Initial Regulation (MS only)

At immediate assignment and at handover, the dynamic MSpower control algorithm is restarted. It is possible that thereceived signal strength in the BTS is quite high at this moment,especially when the MS is located close to the BTS. This highsignal strength may block the BTS or reduce its sensitivity. Thequality of other calls served by the same receiver in the BTSmight be affected. Therefore it is important that the MS reducesits output power as quickly as possible.

In the initial mode, only down regulation is performed. Thequality is not taken into account during the initial phase.


EN/LZT 123 3314 R3A – 203 –

Stationary Regulation (MS and BTS)

Figure 13-18 schematically shows how the signal strengthreceived in the BTS (MS) varies with the path loss between theMS and the BTS and how the behavior is affected by thecontrolling parameters.

SS1

SSDES(SSDESDL)

Received signal strength in the BTS (MS)

Path lossRegulation area

1

23

4

Figure 13-18 Received signal strength versus path loss (ordistance) stationary regulation. Quality is not taken into account

Each segment of the graph in Figure 13-18 is explained below.

� The path loss is large and the MS (BTS) transmits at itsmaximum power.

ô With power control not enabled, the received signal strengthincreases as the path loss decreases. The received powerincreases linearly (in dB units) as path loss decreases.

í With power control enabled, the MS (BTS) power isadjusted.

÷ The MS (BTS) transmits at its lowest possible outputpower.

When quality is taken into account the curve (Figure 13-18) issubjected to a parallel shift in the vertical direction (Figure 13-19).


– 204 – EN/LZT 123 3314 R3A

75 80 85 90 95 100 105 110 115 120 125-95

-90

-85

-80

-75

-70

-65

-60

rxle

v [d

Bm

]

Path loss [dB]

SSDES = -85 QDESUL = 20 LCOMPUL = 70 QCOMPUL = 30

CME 20 R5 algorithm

rxqual = 70 dtqu

rxqual = 45 dtqu

rxqual = 0 dtqu

Figure 13-19 MS power regulation at partial path losscompensation

Power orders with extra margin

In three different traffic cases, the ordered power level is alwaysincreased by a power margin at:

• TCH assignment

• Assignment failure or handover failure

• Intra-cell handover and subcell change


EN/LZT 123 3314 R3A – 205 –

DTX

WHAT CAN BE ACHIEVEDDiscontinuous transmission is a mechanism that allows the radiotransmitter to be switched off during speech pauses. In a normalconversation, this leads to a decrease in transmission time ofabout 50%.

DTX is available for speech as well as for non-transparent datatransmission. It is not used on a BCCH carrier. DTX is availableon a per cell basis.

The primary functions of DTX in the uplink and downlink are:

• uplink

– to save battery in the mobile station

– to reduce the interference in the system

• downlink

– to decrease BTS power consumption, especially duringperiods when the BTS is battery-operated due tomalfunction in the power supply

– to reduce the interference in the system

– to reduce the intermodulation products

(NOTE: If broadcasting a particular frequency, the harmonics at3x, 5x, 7x, etc. can cause intermodulation interference.)

When downlink DTX is used in conjunction with uplink DTX,the C/I ratio in the system improves. This improvement can beutilized for tighter cell planning especially when frequencyhopping is used, thus achieving higher capacity.

When something is sent, it is sent at the proper power level.Consequently, the received signal strength is not affected. Thisis the reason for the improvement in the C/I ratio in the systemwhen using DTX.

Measurements of signal strength and signal quality on theestablished connection are performed by the mobile and by thebase station. When using DTX, these measurements cannot beperformed as often as they are normally performed. Thisdisadvantage results in less accurate measurement reports toother algorithms, e.g. locating and power regulation.


– 206 – EN/LZT 123 3314 R3A

Another disadvantage of DTX is that the quality of plosives(sounds like “p”, “t” and “k”) may be somewhat decreasedbecause of the slowness in the Voice Activity Detector (VAD).


General

The VAD in the transmitter, BTS or MS, detects whether atraffic frame consists of speech, non-transparent data, orbackground noise. If a frame consists of only noise, thetransmitter sends one SIlence Descriptor (SID) frame, and thenthe transmission is stopped. After that, one new SID frame issent each SACCH period until speech or non-transparent data isdetected again. The measurement reports are sent as usual on theSACCH. A SID frame contains information about thebackground noise in the transmitting environment.

In the receiver, MS or BTS, a SID frame detector checks allincoming frames. The detector can separate SID frames fromspeech or non transparent data frames. When a SID frame isdetected, the comfort noise characteristics are updated andcomfort noise will be generated. The noise generation is stoppedwhen a speech frame is detected.

SID frames are sent for three reasons:

• To update the comfort noise characteristics on the receivingside

• To enable signal strength and signal quality measurements

• To avoid on/off distortion

The VAD must be operating at all times to assess whether theinput signal contains speech/non-transparent data or not.

The MS and/or the BTS sends information in the measurementreport/result (every 480 ms) telling whether it has used DTXsome time during previous SACCH interval or not.

Influenced Features

When DTX is used, the measurements of signal strength, signalquality and time alignment are less accurate. Therefore, featuresusing these measurements are affected, e.g. dynamic MS andBTS power control and locating. Features that are an integratedpart of the locating function are affected as well. These features


EN/LZT 123 3314 R3A – 207 –

are load sharing, assignment to another cell, hierarchical cellstructures, intra cell handover, overlaid/underlaid subcells, andcall disconnect.

If DTX is used in combination with cyclic frequency hopping,then hopping where (N) mod 13 = 0 should be avoided (N is thenumber of hopping frequencies). This is because all SACCHframes would then be sent on the same frequency.


– 208 – EN/LZT 123 3314 R3A

FREQUENCY HOPPING


Frequency Diversity

Frequency hopping can reduce the influence of signal strengthvariations caused by Rayleigh fading. In this document, thiseffect is referred to as frequency diversity.

Rayleigh fading is frequency dependent. This implies that thefading dips appear at different locations for differentfrequencies. Thus a mobile utilizing frequency hopping does notremain in a specific fading dip for a longer time than one singleburst. Thereby, signal strength variations are broken up intopieces of a duration short enough for the interleaving andchannel coding process to correct errors. Rayleigh fading dips(causing low signal strength) are thus leveled out and slowlymoving mobiles and cars sitting at a red light perceive a moreeven radio environment. Frequency hopping makes most of thefading dips appear more shallow (Figure 13-20).

distance

received signal strength

Figure 13-20 Schematic picture of multipath fading at twodifferent frequencies and at frequency hopping between the twofrequencies for a slowly moving mobile

The thin solid line and the thin dotted line in Figure 13-20 showthe received signal strength obtained at two frequencies. Thethick solid line shows the smoothing effect when the twofrequencies are combined by frequency hopping. The signal afterdecoding will in effect be an average between the signals at thetwo frequencies.


EN/LZT 123 3314 R3A – 209 –

Fast moving mobiles obtain the same type of improvement asfrequency diversity by their speed alone, since the radioenvironment can easily change between bursts. Frequencydiversity thus gives only a minor extra improvement for fastmoving mobiles.

Cyclic frequency hopping gives slightly better frequencydiversity performance than random frequency hopping.

Interference Averaging

Frequency hopping can break up persistent interference intoperiodic occasions of single burst interference. The cell planningmargin against situations of bad radio conditions can thus bedecreased, since the probability of encountering these conditionsdecreases.

Changing frequency at each burst offers a way to improve theinterference situation described previously. The co-channelinterference will change at every burst, which is beneficial forthe connection that otherwise might suffer from a severeinterference during the entire connection. Likewise, theinterference that one connection is causing another is spread outover a number of connections in single bursts. This effect iscalled interference averaging.

Occasionally there are frequency collisions causing stronginterference but of a very short duration (one burst). Again, thecoding and interleaving get a chance to deal with the situation.The more frequencies that are used in hopping, the more raresuch frequency collisions will be.

Frequency hopping reduces the effect of other types ofinterferences including: co-channel, adjacent channel,intermodulation products, etc.

The radio environment (in terms of interference) will becomemore even, and the cell planning margin can be decreased.Theperformance of the interference averaging is dependent of themode of hopping, cyclic, or random. The greatest improvementis obtained when the interferer and the interfered connectionsuse hopping sequences that are independent of each other (i.e.uncorrelated sequences). The lower the correlation, the higherthe averaging gain. If the interferer and the interfered connectionboth use cyclic hopping and in addition, the same frequencies,they may get in phase with one another. The two connectionsthen hop between the frequencies “hand in hand”, and the co-channel interference persists as if there was no frequency


– 210 – EN/LZT 123 3314 R3A

hopping at all. The effect is referred to as total correlation, andthe resulting improvement is very small.

In a fully loaded system, all possible interferers are transmittingsimultaneously and consequently interfering. Even in this case,random frequency hopping shows a performance gain. However,a requirement is that co-channel cells use different (independent)hopping sequences.

ConclusionsFrom a subscriber point of view, frequency hopping providesimproved speech quality. From an operator point of view, thebenefits are:

• a more dependable and predictable radio environment

• a possibility to give the subscribers a generally uniformspeech quality

• a possibility to decrease the cell planning margin, whichmight be used to employ a tighter frequency re-use yielding acapacity increase

The effects increase with the number of frequencies used forhopping, but the relative benefit by adding yet another frequencydiminishes.


General

Frequency hopping in the base stations can be implemented inone of two ways: baseband hopping or synthesizer hopping.Each has its advantages and disadvantages. The frequencyhopping feature can be switched on and off for each channelgroup separately.

The BCCH channel can never hop, since it is a broadcastchannel. The TCHs and SDCCHs, however, can hop.

Baseband Hopping

During baseband hopping, each transmitter operates on a fixedfrequency. During transmission, all bursts, irrespective of whichconnection they belong to, are routed from the TRX to thetransmitter of the proper frequency (Figure 13-21).


EN/LZT 123 3314 R3A – 211 –

TRX1

TRX2

TRX3

TRX4

transmitterf0 ... fntransmitterf0

transmitterf3

transmitterf2

transmitterf1

controller

controller

controller

controller

bus for routing of bursts

filtercombiner

Figure 13-21 Routing of bursts from the TRX to the transmitterat baseband hopping

The advantage of this mode is that narrow-band tunable filtercombiners can be used. These combiners have up to 16 inputs.This makes it possible to have many frequencies in a hoppingset without having to connect several combiners in cascade.

The disadvantage is that it is not possible to use a larger numberof frequencies than there are transmitters.

Synthesizer Hopping

Synthesizer hopping means that one transmitter handles allbursts that belong to a specific connection. The bursts are sentdirectly and not routed by the bus. The transmitter tunes to thecorrect frequency at transmission of each burst (Figure 13-22).

TRX1

TRX2

TRX3

TRX4

transmitterf0 ... fn

hybridcombiner





hybridcombiner

hybridcombiner

controller

controller

controller

controller

Figure 13-22 Sending bursts from the TRX to the transmitter atsynthesizer hopping

The advantage is that the number of frequencies that can be usedfor hopping is not dependent on the number of transmitters.

The disadvantage is that wide-band hybrid combiners must beused. This type of combiner has approximately 3 dB insertionloss, making more than two combiners in cascade impractical.


– 212 – EN/LZT 123 3314 R3A

Algorithm

Cyclic Frequency Hopping

In cyclic hopping, the frequencies are used in a consecutiveorder, e.g. the sequence of frequencies for cyclic hoppingbetween four frequencies may look like this:

... , I4, I1, I2, I3, I4, I1, I2, I3, I4, I1, I2, ...

There is only one cyclic sequence defined in the GSMspecifications. The sequence of frequencies goes from thelowest absolute frequency number (in the set of frequenciesspecified for that channel group) to the highest and over again.

Random Frequency Hopping

A random hopping sequence is implemented as a pseudo-random sequence. The sequence is stored in a reference table inthe mobiles as well as in the base stations. Sixty-threeindependent sequences are defined. Which of the 63 sequencesto be used is specified by a Hopping Sequence Number (+61)parameter. The algorithm for selecting the frequency for eachburst allows hopping on up to 64 frequencies. The actualfrequency to be used at each instant is obtained by a modulooperation with the available frequencies.

When random hopping is used, the frequencies are used(pseudo-) randomly. A hopping sequence for four frequencieslooks something like this:

... , I1, I4, I4, I3, I1, I2, I4, I1, I3, I3, I2, ...

The period for a random sequence is six minutes.


EN/LZT 123 3314 R3A – 213 –

INTRA-CELL HANDOVER


Intra-cell handover provides the ability to improve the speechquality during a conversation. This is done by changing thechannel when there is poor quality on the uplink or downlink,although the signal strength is high.

A limitation in the feature is the impossibility to separate thetemporary interference from time dispersion. In the timedispersion case, it is no use just changing channels within thesame cell, however, a cell change might solve the problem.

A large amount of unnecessary handovers between channels inthe same cell can occur, e.g. in the presence of time dispersion.In order to prevent this, there is a limit to how many consecutiveintra-cell handovers are allowed. If the number of intra-cellhandovers has exceeded this limit, a certain time has to elapsebefore another intra-cell handover is attempted.


General

A high received quality value corresponds to a high bit errorrate. The FQSS table (quality vs. signal strength function)defines the maximum acceptable received quality value forevery signal strength value. If, at a specific received signalstrength, the quality value exceeds this defined value, an intra-cell handover is requested.

Algorithm

Quality vs. Signal Strength Condition

Figure 13-23 illustrates the FQSS table mentioned above, wherethe curve represents the signal strength together with itsmaximum accepted received quality value. This curve can bedisplaced in the horizontal and vertical direction via parameters.


– 214 – EN/LZT 123 3314 R3A

RXLEV [rxlev units]

rxqual [dtqu1]

Intra-cell handovercriteria fulfilled

(63, 50)

, 50)

(31, 60)

10

20

30

4050

60

100

8090

70

30 4020 6050

Figure 13-23 The quality versus signal strength function(FQSS)

Figure 13-23 shows that the maximum acceptable receivedquality value decreases as the received signal strength increases.This is based on the assumption that the probability to find abetter channel within the same cell is higher at a high signalstrength, (i.e. an intra-cell handover attempt is made at a lowerU[TXDO�>GWTX�@ value than in the low signal strength case).

Intra-cell handover on a signaling channel

There is a possibility to permit intra-cell handover on signalingchannels.

6 dtqu = deci transformed quality units.


EN/LZT 123 3314 R3A – 215 –

ASSIGNMENT TO ANOTHER CELL


The assignment to another cell feature operates at call set-up,during which a traffic channel (TCH) is to be assigned. Aspreviously described, the feature makes it possible to select TCHin a cell other than the one that is currently serving theconnection. The result, in terms of the effect on the entirenetwork, depends on the radio conditions in the target cell. Thetarget cell is either ranked better than the serving cell in thelocating evaluations or worse.

• Better cell

When the locating evaluations start at an early stage in thecall set-up procedure, the locating evaluation may find cellswith better signal levels than the serving cell. To preventunnecessary handovers and also to improve the generalinterference environment, the assignment can be redirectedto such a cell.

• Worse cell

If a mobile cannot connect to the selected cell due tocongestion or if the Locating feature detects an urgencycondition during call set-up, assignment to cells with lesssignal strength than the serving cell can be recommended.The probability for a call to get a useable connection withthe network then increases.


When a call is to be established, the idle mode cell selectionalgorithm in the mobile selects a cell and an SDCCH (or a TCHin the case when immediate assignment on TCH is used) isallocated. On some occasions it might be better to allocate achannel in another cell than the one currently serving theconnection. On other occasions this might be the only possibilityto succeed with call set-up.


– 216 – EN/LZT 123 3314 R3A

DYNAMIC OVERLAID/UNDERLAID SUBCELLS


Increased Traffic Capacity

The dynamic overlaid/underlaid subcells feature provides a wayto increase the traffic capacity in a cellular network withoutbuilding new sites. The feature makes it possible to use twodifferent frequency re-use patterns: one pattern for overlaidsubcells and one for underlaid subcells. The availablefrequencies are divided between the overlaid and the underlaidsubcells. Each overlaid subcell serves a smaller area than thecorresponding underlaid subcell and the frequency re-usedistance for the overlaid subcells can therefore be made shorter.Consequently, the number of frequencies per cell can beincreased providing an increased traffic capacity in the cellularnetwork.

Figure 13-24 shows an example of a dynamic overlaid/underlaidsubcell structure. Two different frequencies are indicated: I

R in

the overlaid subcells and IX in the underlaid subcells.

fo

fu

fo

fu

fo

f u

fo

f o

fo

Figure 13-24 Dynamic overlaid and underlaid subcells


EN/LZT 123 3314 R3A – 217 –

A Local Increase of Capacity

Another situation where this feature can be useful is when anincrease in traffic capacity is required locally in an existingcellular network and there are no more frequencies available. Byadding an overlaid subcell, a local increase in capacity can beachieved without building a new site. As the overlaid subcell isrestricted in size, it is possible to re-use frequencies that wouldotherwise not be allowed to be used at that location according tothe existing frequency plan.

Trunking Considerations

Within a cellular network covered by overlaid subcells, it ispossible to use the available channels in both subcells. However,in those parts of the cellular network that are covered by theunderlaid subcells but not by the overlaid subcells, only thechannels in the underlaid subcells are accessible. Thus, atrunking loss occurs in the parts of the underlaid subcells notcovered by the overlaid subcells. In the area covered by bothsubcells, there is no trunking loss because all channels can beused.

The trunking loss can be minimized by first allocating theavailable channels in the overlaid subcells before allocating thechannels in the underlaid subcells. This is achieved by featurechannel administration. Thus, the channels in the underlaidsubcells will not be wasted but will be available for mobiles thatcannot use channels in the overlaid subcells.

Dimensioning Considerations

When dividing the available frequencies between the overlaidand the underlaid subcells, the size of the overlaid subcells inrelation to the size of the underlaid subcells should beconsidered. The geographical traffic distribution over thecellular network should also be considered when dimensioningthe overlaid/underlaid subcell structure.

It is advisable to have as many channels as possible in theoverlaid subcells because this increases traffic capacity.However, if there are idle channels left in the overlaid subcellswhen no idle channels in the underlaid subcells are available, atrunking loss occurs in the area covered by the underlaidsubcells only. The overlaid subcells should thus be dimensionedfor a higher level of congestion than the underlaid subcells.These considerations must be matched against subcell size and


– 218 – EN/LZT 123 3314 R3A

traffic distribution to obtain an optimal dimensioning of theoverlaid/underlaid subcells.


General

The overlaid and the underlaid subcells share BCCHs, whichreside in the underlaid subcells. SDCCHs can be defined in theoverlaid as well as in the underlaid subcells, but immediateassignment can only be made to the underlaid subcells.

There is a path loss threshold and a timing advance threshold foreach overlaid subcell as a criterion to maintain the restrictedserving area of the overlaid subcell. The downlink signalstrength and timing advance measurements from the serving cellare used in the criterion to evaluate whether the mobile isqualified for the overlaid subcell or for the underlaid subcellonly.

Algorithm

Subcell Change at Handover and Assignment to Another Cell

It is possible to make a handover from both the overlaid and theunderlaid subcell in the serving cell to either the overlaid or theunderlaid subcell in the target cell. Assignment to either anoverlaid or an underlaid subcell in another cell is also possible.Handover to another cell and assignment to another cell are onlyallowed to an overlaid subcell if the neighbor cell is co-sitedwith the serving cell.

Subcell Change and Intra-cell Handover

A change from overlaid to underlaid subcell can also occurwhen certain other conditions are fulfilled. If the criterion forintra-cell handover is fulfilled, the serving cell is an overlaidsubcell and the maximum number of allowed consecutive intra-cell handovers have been executed, a subcell change is permittedto the underlaid subcell. At this type of subcell change, a timer isset to prevent an immediate subcell change back to the overlaidsubcell.

Note that if the intra-cell handover criterion is fulfilled and thereare no idle channels in the overlaid subcell, an intra-cellhandover is always permitted to the underlaid subcell.


EN/LZT 123 3314 R3A – 219 –

TSC Considerations

The equalizer uses the Training Sequence Code (76&) to createa channel model. When configuring an overlaid subcell networkon top of an existing underlaid network, it is advisable toreassign the 76&s for the overlaid subcells according to theoverlaid frequency re-use pattern. Otherwise, co-channels in theoverlaid subcell network might obtain the same 76&�which canmake it difficult for the equalizer to distinguish between theseco-channels.


– 220 – EN/LZT 123 3314 R3A

HIERARCHICAL CELL STRUCTURES


The hierarchical cell structures feature provides the ability ofdividing the cell network into two or three layers. The higherlayers can be used for large cells and the lower layers for smallcells. One example is when large cells are added to a small(normal) cellular network to provide coverage at coverage gaps.The large cells then act as “umbrella” cells for the normal cells.Another example is when microcells are added to a normalcellular network to provide extra hot-spot capacity. The normalcells then act as “umbrella” cells for the micro cells. Thehierarchical cell structures feature makes it possible to passbetween cell layers in a controlled way thus facilitatingdimensioning and cell planning in cell structures where largeand small cells are mixed. Sufficient quality can be achieved,although the best cell is not always used as the serving cell.

The layer can also be seen as a pure priority designation with thelower layer as the higher priority. This implies that thehierarchical cell structures feature can have other uses than theone described above, e.g. it can be used if there is a need toprioritize certain kinds of cells, irrespective of their size. Oneexample is dual band networks where cells in the 1800 MHzband are combined with cells in the 900 MHz band. Single band900 MHz mobiles cannot access 1800 MHz cells. Thus, dualband mobiles should, if possible, select cells in the 1800 MHzband in order not to create unnecessary congestion in the 900MHz band. This can be achieved by assigning higher priority tothe 1800 MHz cells.

Here, the microcells (priority 1 cells) are called Layer 1 cells, thenormal cells (priority 2 cells) are called Layer 2 cells and theumbrella cells (priority 3 cells) are called Layer 3 cells(Figure 13-25).


EN/LZT 123 3314 R3A – 221 –

Layer 3 cells

Layer 2 cells

Layer 1 cells

Figure 13-25 A layered cell structure

The hierarchical cell structures feature takes ordinary locatingcriteria as well as priority of the layers into consideration.Locating knows to which layer each cell belongs. Thisinformation is taken into account when the handover candidatesare ranked.

The general idea for allocating mobiles using the hierarchicalcell structures feature is to try to direct them to cells within thelower layers. The purpose of the higher layer cells (Layers 2 and3) is to:

• WDNH�FDUH�RI�UDGLR�FRYHUDJH�KROHV - a handover to a higherlayer cell is performed if the signal strength falls below athreshold

• SURYLGH�VSDUH�FDSDFLW\ - a higher layer cell can be chosen ifthere is congestion in the lower layer cells, a higher layer celleven worse than the serving cell can be chosen if assignmentto a worse cell is allowed

• VHUYH�DV�UHVFXH�IRU�UDGLR�GLVWXUEDQFHV - a handover to ahigher layer cell can be performed at bad quality urgency


– 222 – EN/LZT 123 3314 R3A


General

A signal strength threshold exists for each cell in Layers 1 and 2.It is used to decide when to perform a handover to or from ahigher layer cell. When the signal strength in a lower layer cell(Layers 1 or 2) falls below the threshold, a handover to a higherlayer cell is performed, provided that no other cell in the same orin a lower layer is suitable. Normally no handover back to alower layer cell is made until a lower layer cell generates asignal strength above the corresponding threshold.

Algorithm

There is a signal strength threshold and a hysteresis for each cellin Layers 1 and 2. They are used to facilitate passing betweenthe three layers in a systematic way.

The threshold (modified by the hysteresis) is used when passingbetween the layers upwards as well as downwards.

Handling of Fast Moving Mobiles

As a way to prevent fast moving mobiles from performinghandover to lower layer cells, a temporary signal strengthpenalty is used. The first time a cell is reported as a neighbor inthe measurement result message, that cell is punished for acertain time interval.

This penalty only applies if the neighboring cell belongs to alower layer than the serving cell. Hence, only Layers 1 and 2cells can be temporarily punished.


EN/LZT 123 3314 R3A – 223 –

EXTENDED RANGE


Extended range is available for RBS 200.

The operator can configure a BTS for extended range. In a BTSconfigured for extended range, all channels are used forextended range.

By configuring a BTS for extended range, the maximumpossible cell radius is increased from 35 to approximately 72km. A consequence of using this feature is that the capacity insuch a cell is reduced.

An overlaid/underlaid subcell structure is not allowed for anextended range cell.


General

Extended range is obtained by allowing the BTS to receive thebursts later than during the time slot assigned to the MS. Thetime slot next to the one assigned to the MS is also reserved.Thus, the negative aspect of using extended range is that there isa loss in capacity.

Algorithm

The BTS is not restricted to handle a maximum TA of 63,although it is not allowed to order the MS a TA greater than 63.When a channel is configured for extended range, the burstsfrom the MS can be received later than during the assigned timeslot. The bursts slide into the next time slot when the mobilemoves further away from the base station than allowed by themaximum TA. The TA ordered to the mobile is still 63 but thetime difference between sending and receiving is now largerthan that given by the maximum TA.

In an extended range cell, the timing advance corresponding tothe real signal delay (measured in the BTS) is called VirtualTiming Advance (VTA). Thus VTA is the timing advance thatwould have been used if the air interface protocol had allowed aTA larger than 63. When VTA is larger than 63, the time slot


– 224 – EN/LZT 123 3314 R3A

next to the one assigned is also needed for reception. Atimmediate assignment the BSC extracts the VTA value from therandom access delay information it receives from the BTS.When the call is established, VTA is regularly sent from theBTS to the BSC.

The parameter 7$/,0 is used to specify the TA limit forurgency handover. TA is continuously monitored and if itexceeds 7$/,0, locating suggests that the cell be abandonedquickly. In such a case, a handover is permitted to a worse cellas well as to a better cell. 7$/,0 in a normal cell is restrictedto the values 0 to 63. For an extended range cell, 7$/,0 can belarger and is an upper limit for VTA instead of for TA.

In regions with bad coverage there is a risk that a mobile movesfar away from the serving cell site. In such situations the speechquality can severely deteriorate and it might be preferable tobreak the connection. The parameter 0$;7$ specifies atwhich TA a mobile will be disconnected. As for 7$/,0� a newvalue range is allowed for an extended range cell. For such acell, 0$;7$ is a limit for VTA instead of for TA.

If 7$/,0 and/or 0$;7$ is greater than what is allowed fornormal cells, a cell cannot be reconfigured from an extendedrange cell to a normal cell unless these parameter values arechanged to values allowed for a normal cell.

The allowed value range for 7$/,0 and 0$;7$ for anextended range cell is 0 - 219 (bit periods). During one bitperiod, (3.7 µs) the radio waves propagate approximately 1.1kilometer. Since VTA is given by the sum of the propagationdelay in both directions, VTA�equal to 219 corresponds to adistance between the site and the mobile of approximately 120kilometers.

RBS 200 supports a range of up to 72 km for an extended rangecell. The maximum value for VTA is 133 bit periods. Usinglarger values for 7$/,0 and 0$;7$ is effectively the sameas disconnecting the corresponding functionality for RBS 200.

To increase the traffic capacity in an extended range cell, a TRXcan be configured for combined control signaling. With thisTRX configuration, the same channel is used for both broadcastand dedicated signaling. The remaining physical channels are allavailable for traffic. An extended range cell with one TRX andconfigured for combined control signaling thus holds threetraffic channels. Only two channels are available for traffic if theTRX is not configured for combined control signaling.


EN/LZT 123 3314 R3A – 225 –

72 kmVTA = 133

Figure 13-26 Cell ranges for normal and extended range (RBS200) cells

Use “72 km” and VTA = 133” near the outer ring


– 226 – EN/LZT 123 3314 R3A

IMMEDIATE ASSIGNMENT ON TRAFFIC CHANNEL


The immediate assignment feature on TCH takes intoconsideration the traffic situation for which the channel isneeded. For speech/data connections, the operator can choosebetween three general strategies:

• Immediate assignment on TCH is not allowed

• Immediate assignment on TCH as last preference

A TCH may only be allocated at immediate assignmentwhen there are no idle SDCCHs available. The dimensioningof SDCCHs becomes less critical and the number ofSDCCHs can be reduced in favor of more TCHs. By thetrunking efficiency obtained, there is a capacity increase.

• Immediate assignment on TCH as a first preference

A TCH is allocated as a first preference at immediateassignment. An SDCCH may only be allocated at immediateassignment when there are no idle TCHs available. Thenumber of SDCCHs can be reduced in favor of more TCHs.However, the load on the TCHs increases.

For signaling connections, the strategy is always to allocate anSDCCH, irrespective of the selected strategy for speech/dataconnections. However, it is not always possible to follow thisstrategy if a “TCH first strategy” has been selected forspeech/data connections. In general, a TCH is never allocated atimmediate assignment if the BSC has enough information todetermine that the connection is of a type that requests anSDCCH as a carrier (e.g. location updating and SMS).

In the cases when sufficient information is not present, theoutcome (a TCH or an SDCCH allocated) depends on the detailsof the case at hand. To a certain degree, this can also becontrolled by the operator.

Short Technical Description

When a connection is to be established, a signaling channel forthe call set-up procedure must be allocated. The call set-upsignaling is performed by the procedures immediate assignmentand assignment. The immediate assignment feature on TCHallows the signaling to be carried by a TCH as well as by anSDCCH or:


EN/LZT 123 3314 R3A – 227 –

• Immediate assignment on SDCCH

An SDCCH is allocated at immediate assignment. If achannel for speech/data is needed, a TCH is allocated atassignment. If a channel for signaling is needed, theconnection remains on the allocated SDCCH.

• Immediate assignment on TCH

A TCH is allocated at immediate assignment. If a channelfor speech/data is needed, the connection changes channelmode at assignment, but remains on the allocated TCH. If achannel for signaling is needed, the connection also remainson the allocated TCH. However, the signaling traffic iscarried by the SACCH part (SMS) or the FACCH part(location updating, supplementary services, IMSI detach).

At immediate assignment, the MS sends a channel requestmessage to the BTS. The message contains the informationelement “Establishment cause”. Among other things, theestablishment cause gives an indication whether a channel forsignaling or for speech/data is needed. The decision whether toallocate an SDCCH or a TCH at immediate assignment is basedon this information.

Half rate channels are introduced and are usually first allocatedif the MS has indicated dual rate capability.


– 228 – EN/LZT 123 3314 R3A

DOUBLE BA LISTS


An MS in idle mode continuously measures the received signalstrength of the BCCH carriers in the serving cell and in neighborcells. The results of these measurements are used to select thebest cell to camp on.

In active mode, the MS measures the signal strength and biterror rate of the serving cell, as well as the signal strength of theBCCHs of the neighboring cells. The measurement results areused in locating and BTS power control.

The Double BA lists feature provides the MS with suitablemeasurement frequencies for both idle mode and active mode.The idle mode BA list is sent to the MS in system informationmessages on the BCCH, and the active mode BA list is sent onthe SACCH.

The lists may contain frequencies from the GSM900 and/orGSM 1800 bands.

The feature Double BA list is implemented in the BSC.

Better Measurement Accuracy

An MS in active mode has a certain time period to performsignal strength measurements. With a small number offrequencies to measure on, more samples can be taken for eachfrequency. This leads to higher statistical accuracy in thecalculated mean value. Better measurement reports improve thelocating result.

Faster Connection at Switch-on

If the MS has the option to save the idle mode BA list whenpowered off, it can search the latest known BCCH carriers atpower on. If no BA list is stored, the MS scans the entirefrequency band (124 frequencies for GSM 900, and 374frequencies for GSM 1800) and arranges the frequencies indescending order of signal strength. The MS then tunes to thestrongest of the RF channels and determines if the BCCH carrieris permitted. If it is permitted, the system information messagescan be decoded and the idle mode BA list for that cell can beused. If the system information messages cannot be decoded, the


EN/LZT 123 3314 R3A – 229 –

MS searches the carriers in descending signal strength order tofind a suitable BCCH carrier.


– 230 – EN/LZT 123 3314 R3A

IDLE CHANNEL MEASUREMENTS


Through the use of idle channel measurements the generalquality of radio connections in the network can be improved.For each channel allocation, the channel administration featurecan use the information from idle channel measurements toselect the most suitable channel.

Idle channel measurements assign each idle channel to one offive interference bands, which together cover the signal strengthmeasurement range. The interference band ranges determine theresolution of the measurements. The interference bands can bechanged from their default ranges on a per cell basis, so that theresolution can be adjusted for optimum feature performance, e.g.in microcells the tolerable interference level may be significantlyhigher than in rural areas, as the signal strength of the carrier ingeneral will be higher. Therefore, it can be beneficial to changethe interference bands so that a better resolution is achieved atthe interference levels that can be considered to be critical foreach cell.

Idle channel measurements can be activated and deactivated on aper cell basis. It is also possible, on a per cell basis, to have themeasurements active but not used for channel allocation. In thiscase, idle channel measurements are only used for statistics.These statistics show (per cell) the number of channels in aninterference band.

Note that as measurements are performed on a per channel basis,information about the interference on a specific frequency is notretrievable when frequency hopping is used.


General

Idle channel measurements are activated in a cell, i.e. uplinksignal strength measurements are continuously performed in theBTS on all idle channels. The measurements are made on TCHsand SDCCHs in the same way as the signal strengthmeasurements are performed on active channels. The channelsare then ranked in order of increasing interference. Depending


EN/LZT 123 3314 R3A – 231 –

on the measured signal strength, each channel is placed in one ofthe five corresponding interference bands.

At channel allocation, channel administration can use theranking of channels into interference bands as helpfulinformation when selecting a channel.

Algorithm

Basic Algorithm

When a channel is deblocked, it is initially placed in aninterference band that represents the lowest measuredinterference. The BTS continuously performs uplink signalstrength measurements as long as the channel is idle. Afterchannel deblocking, the BTS sends a report on measured uplinkinterference, at the latest, after two SACCH periods. After that,the BTS only reports to the BSC if the interference band for anidle channel has changed. The channel is then moved to theappropriate interference band in the BSC.

To determine a value for the interference on a channel, the BTStakes the average of measurements over a number of SACCHperiods. This is done once every SACCH period, and thenumber of SACCH periods to average over is determined by theparameter ,17$9(.

Interference Bands

Five interference bands are defined for each cell. Theinterference bands are disjunctive and together cover values inU[OHY units from 0 up to 63.

The interference bands can be changed through four parameters/,0,7� to /,0,7�� These parameters define the upper limit ofthe four lowest interference bands. The first interference bandalways starts from 0 and ranges up to /,0,7�. The secondinterference band ranges from /,0,7� + 1 up to /,0,7�, etc.The fifth interference band ranges from /,0,7� + 1 up to 63.Note that the parameter values must ascend as follows:

/,0,7� < /,0,7� < /,0,7� < /,0,7�


– 232 – EN/LZT 123 3314 R3A

interference band 5

interference band 4

interference band 3

interference band 2

interference band 1

interferencelevel

-47 dBm

-110 dBm

measuredrxlev

LIMIT4

LIMIT3

LIMIT2

LIMIT1

rxlev = 63

rxlev = 0

Figure 13-27 Definitions of interference bands

Channel Release

At normal release, the channel is placed in the same interferenceband that it was taken from at channel allocation. At abnormalrelease however, the channel is placed in the interference bandimmediately above the one from which it was taken. That is it isplaced in a band representing a higher interference, e.g. if achannel is allocated from interference band 2, it is placed backinto interference band 2 at a normal release, and intointerference band 3 at an abnormal release.

After a channel release, the BTS reports the measured uplinkinterference for that channel to the BSC, at the latest, after thefirst two SACCH periods just as is done after deblocking.

Dual Rate Channels

When evaluating the interference level for a full rate channelcomposed of two half rate channels, the average of theinterference bands for the half rate channels is used.

When a half rate channel is released and its related half ratechannel is idle, the channel that is released is placed in the sameinterference band as its related channel. This is done regardlessof whether the release is normal or abnormal.


EN/LZT 123 3314 R3A – 233 –

CELL LOAD SHARING


A mechanism that distributes the load between cells has theeffect that high load peaks in a cell are cut. Therefore, a higheraverage load can be allowed while the grade of service, as wellas the dimensioning, remains the same. The trunking efficiencyof a BSC is thus greater, resulting in an increase of trafficcapacity.

Cell load sharing increases the amount of handovers in the partof the network where the traffic load is unevenly distributed.


Algorithm

Overview

Cell load sharing is limited to traffic channels (TCH) in anymode (i.e. speech/data or signaling).

The cell load sharing algorithm consists of the followingactivities:

• The traffic load in all cells where load sharing is activated ismonitored, the load level determines further activities.

• If a cell has too high of a load, connections close to a cellborder are made to perform handover by recalculating theirranking value in locating.

• The handovers are carried out only if the receiving cells havea low enough load.

• Cell load sharing is activated in a BSC with the parameter/667$7(, and in individual cells with the parameter&/667$7(�.

7 /667$7( and &/667$7( are not entered into the AXE as a parameters, but represent state variables. They are changed by commands withoutan associated qualifier (parameter).


– 234 – EN/LZT 123 3314 R3A

Load Monitoring

The load in every cell is monitored in the BSC. The measure ofthe load is the amount of idle traffic channels. Two levels arerelevant for cell load sharing:

• If the amount of idle traffic channels decreases below&/6/(9(/ (the parameter value is given as the percentageof all idle traffic channels in the cell) in a cell, that cell triesto rid itself of some traffic by initiating load sharinghandover to neighboring cells.

• If the amount of idle traffic channels is above load&/6$&& (also given as a percentage) in a cell, that cell isprepared to accept incoming load sharing handovers fromother cells.

Ranking Recalculations

New ranking calculations are performed in locating for allconnections in a cell, when the amount of idle traffic channelsdecreases below &/6/(9(/. If a better neighbor cell is foundfor any of the connections as a result of this new cell ranking, aload sharing handover is requested for that connection.

Successive locating recalculations are performed with a linearramping down of the hysteresis. The ramping down is performedduring a defined time period or until the amount of idle trafficchannels increases above &/6/(9(/.

The purpose of ramping down the hysteresis is two-fold:

• The mobiles closest to the handover border are selected first.

• The mobiles selected for handover are few at a time, toomany load sharing handovers at the same time mightotherwise cause instabilities.

Certain additional conditions must be fulfilled for a load sharinghandover to take place:

• Locating conditions:

– load sharing handover is not allowed during theassignment

– load sharing handover is not allowed if there is anurgency condition


EN/LZT 123 3314 R3A – 235 –

• Conditions for the neighbor cell:

– the neighbor cell must be ranked worse than the servingcell

– the neighbor cell must belong to the same BSC

– the neighbor cell must belong to the same hierarchicallayer

– incoming load sharing handover must be allowed for thatcell

Dual rate channels have been introduced. The amount of idletraffic channels is counted as if only full rate channels exist.This means that if one of a pair of half rate channels in a dualrate channel is busy and the other is idle, the entire dual ratechannel is classified as busy.


– 236 – EN/LZT 123 3314 R3A

MULTIBAND OPERATION


The multiband operation feature allows for GSM 900 operatorsto increase their capacity and for GSM 1800 operators toincrease their coverage. Furthermore, it allows the operator toprovide service to both GSM 900 and GSM 1800 roamers.

An MSC can handle BSCs of different system types. This meansthat it is possible to have GSM 900 BSCs, GSM 1800 BSCs,and GSM 1900 BSCs connected to the same MSC.

Correspondingly, a BSC can handle cells of different systemtypes. This means that it is possible to have GSM 900 cells,GSM 1800 cells, and GSM 1900 cells connected to the sameBSC. Each cell must, however, only contain frequencies withinone frequency band.

The operator can choose whether to have a separate network foreach of the different frequency bands or join the GSM 900 cellsand the GSM 1800 cells into one combined network. The GSM1900 cells must always form a separate network.

A multiband mobile station can receive, transmit, and measureon GSM 900 frequencies as well as on GSM 1800 frequencies.In a multiband network, the system supports cell reselection,assignment and handover between GSM 900 cells and GSM1800 cells for multiband mobile stations.

Multiband mobile stations and all types of single band mobilestations can co-exist. This has two benefits:

• An operator that is changing the network from a single band(GSM 900 or GSM 1800) to a multiband network is stillable to serve the single band MSs on both frequency bands.

• A multiband subscriber will be able to use the multibandmobile station in a pure GSM 900 network as well as in apure GSM 1800 network.


EN/LZT 123 3314 R3A – 237 –


General

A multiband MS performs neighboring cell measurements onGSM 900 cells as well as on GSM 1800 cells. Both types ofcells are evaluated in the idle mode cell reselection algorithmand in locating, given that cell reselection, assignment, andhandover can be performed between the frequency bands.

A single band MS does not measure neighboring signal strengthon frequencies it cannot handle. This means that a single bandMS will never be allocated a channel from the wrong frequencyband.

As the frequency and power capabilities are transferred to apossible target BSC at inter-BSC handover, the multiband MScan be assigned a suitable channel in the new BSC in theappropriate frequency band.

Multiband operation does not put any restrictions on normaltraffic handling.

Power and Frequency Capability of a Mobile Station

All mobile stations inform the network about their powercapability as soon as possible after access. A multiband mobilestation informs the network about which frequency bands it canhandle and the power capabilities on each frequency band. Thisinformation is included in MS ClassMark 3 (CM3).

A message including CM3 received by the BSC is transferred tothe MSC.

The MSC transfers the CM3-information to the target BSC atinter-BSC handover.

Broadcast of Measurement Frequencies

An MS performs neighboring cell measurements in idle as wellas in active mode. The information about which neighbors are tobe measured is broadcast to all mobiles in the cell. A single bandmobile station ignores frequencies it cannot handle.

The frequencies of neighboring cells are given by a parameter.With this parameter, GSM 900 frequencies as well as GSM


– 238 – EN/LZT 123 3314 R3A

1800 frequencies can be defined as neighboring frequencies in amultiband network.

Measurement Reporting

A multiband MS measures the signal strength from GSM 900neighbors as well as from GSM 1800 neighbors. It is possible tocontrol how many GSM 900 cells and GSM 1800 cells, arereported in the measurement. This implies that the six strongestcells will not always be reported. If no cells are available in oneband, all six positions in the measurement report can be used forthe available frequency band.


EN/LZT 123 3314 R3A – 239 –

DIFFERENTIAL CHANNEL ALLOCATION


Differential channel allocation provides the ability todifferentiate between different types of subscribers (i.e. to givesubscriber segments different access to the network via anallocation control of the radio resources). For this purpose, eachsubscriber can be assigned a priority level in the subscriptiondata in HLR. In order to offer high access to the network inspecial cases as emergency calls or for subscribers with highpriority, a priority mechanism to reserve idle channels isimplemented. Other applications for differential channelallocation can be to construct areas where certain subscribershave full access to channels. Outside of these areas the samesubscribers have restricted access to the network or no access atall.

The operator can specify a unique priority level for emergencycalls, thus the ability to give these calls higher priority thanordinary calls is provided.


GeneralDifferential channel allocation can be activated per BSC basisand controlled per cell basis.

At activation of differential channel allocation, the priority levelfor emergency calls must be specified both in the BSC and theMSC/VLR. Note that the operator must ensure that the prioritylevel for emergency calls is set to the same level in the BSC andin the MSC/VLR.

The differentiation is performed by specifying a number ofchannels that cannot be allocated (inaccessible channels) forsubscribers with that specific priority level. This can becontrolled per overlaid or underlaid subcell as well as per logicalchannel type TCH or SDCCH.

Differentiation is performed during the first assignmentprocedure, before the call is established. It is also possible atintra-BSC handover. This means that a certain subscriber mayreceive a channel at call set up, but at a handover attempt themobile subscriber may be denied a channel because it has too“low” priority.


– 240 – EN/LZT 123 3314 R3A

ADAPTIVE CONFIGURATION OF SDCCH

INTRODUCTION

The Adaptive configuration of logical control channels feature isdesigned to optimize the traffic and signalling channel usage andmake Stand-alone Dedicated Control Channel (SDCCH)dimensioning less critical. The purpose of the feature is tominimize the risk of SDCCH congestion by automaticallyadapting the number of SDCCHs in a cell to the demand forsuch channels.

In this document, a dedicated signalling channel is referred to asan SDCCH subchannel, or simply an SDCCH. The SDCCHsubchannels are configured together on a Basic PhysicalChannel (BPC) either as eight SDCCH subchannels, or fourSDCCH subchannels combined with the Broadcast ControlCHannel (BCCH) and Common Control CHannel (CCCH).These different SDCCH configurations are referred to as anSDCCH/8 and an SDCCH/4 respectively.

The feature is implemented in the Base Station Controller(BSC).


Adaptive configuration of logical channels will dynamicallydimension the cell with more (or less) SDCCH/8s on demand.This prevents SDCCH congestion to occur as well as optimizingthe usage of SDCCHs and TCHs. As a result, the network willhave higher capacity and the revenue for the operator willincrease.

The dimensioning of the SDCCHs will also be less critical. Thecumbersome manual work for an operator to calculate theexpected signalling traffic based on traffic models, currenttraffic distribution, and statistics about handovers andcongestion rates is therefore minimized.

As a side effect, the number of intra-cell handovers mightincrease slightly. The reason is that when a timeslot is to bereconfigured, any ongoing traffic on that timeslot is handed overto other idle channels.


EN/LZT 123 3314 R3A – 241 –


General

By using the Adaptive configuration of logical channels featurethe operator needs only to dimension a basic SDCCHconfiguration. The basic SDCCH configuration in a cell shouldbe under-dimensioned rather than over-dimensioned regardingthe number of SDCCHs when this feature is used. Thedimensioning can be based on a low to average SDCCH load.Should an increased demand for signalling channels arise thefeature will dynamically replace one idle TCH in the cell with anSDCCH/8. This operation will be completed within a fewseconds and can be repeated several times. When the demandfor signalling channels returns to a lower level the procedure isreversed.

Adaptive configuration of logical channels is activated anddeactivated on a per cell basis using state variable ACSTATE.The parameters which control the operation of the feature can bechanged by command. The feature will only add and removeSDCCH/8s in the underlaid (UL) subcell if a subcell structure isdefined. An SDCCH/8 manually configured in the overlaid (OL)subcell is not affected by this feature.


– 242 – EN/LZT 123 3314 R3A

FREQUENCY ALLOCATION SUPPORT (FAS)

INTRODUCTION

Frequency Allocation Support (FAS) is a function based onuplink interference measurements. The interference can bemeasured for a number of frequencies in a number of cells. Theinterference data may be used for evaluating and/or findingpossible improvements to the current frequency plan. Inconnection to the interference data, FAS also supportspresentation of statistics on performance data for cells.

FAS is operated through the Operation and Support System(OSS).


FAS records the uplink interference level for specifiedfrequencies and cells. After a recording is completed the resultscan be presented to the operator in the OSS user interface.

The FAS outputs show information on the uplink interferencelevel for up to 150 frequencies at each cell. This enables theoperator to compare the interference levels of the frequenciesand possibly change a frequency in a cell to a less interfered one.

Thus the information from FAS can be used for frequencyplanning. Usages are:

• Finding better frequency allocations in a cell

• Helping introduction of a new transceiver (TRX) or cell in anetwork already in operation

• Monitoring the interference levels on frequencies allocated

FAS can also present a comparison of the results from tworecordings so that the effect of a frequency reallocation can beexamined. Furthermore, recording results can be exported forexternal processing and analysis.


EN/LZT 123 3314 R3A – 243 –


General

FAS is operated from OSS. OSS initiates the Base StationController (BSC) function Radio Interference Recordingaccording to the user’s settings. Traffic statistics are fetchedfrom the Performance Management database in OSS. Thismeans that the collection of traffic statistics in the system has tobe set up outside the FAS application.

The interference measurements are performed automatically bythe network according to what has been specified in therecording configuration.

Frequency Hopping does not affect the validity of any FASmeasurements.

A recording configuration is only accepted if all cells andfrequencies are of the same system type (GSM 900, GSM 1800or GSM 1900).

The activities in FAS can be divided into three parts:

1. The recording configuration, when the user schedules themeasurements.

2. The recording procedure, when the network performs themeasurements specified in the recording configuration.

3. The recording evaluation, when the user requests apresentation of recording data.

Recording Configuration

Before starting a measurement recording, a recordingconfiguration must be defined.

The parameters in the recording configuration are set jointly forall cells.

The recording configuration consists of:

• The cell set, defining in which cells to record.


– 244 – EN/LZT 123 3314 R3A

• One or more frequency sets specifying which frequencies tomeasure. These frequencies will be monitored by each TRXof the cells specified by the cell set. Instead of specifyingfrequency sets, the frequencies can be specified to be thoseallocated to each cell (in the cell set).

• A percentile value that specifies which RXLEV value toreport from the recorded interference distribution, e.g. the90th percentile value of a distribution is the value which 90per cent of the samples are lower than or equal to.

• The recording session defined by a time schedule. The timeschedule consists of a start date and the number of recordingperiods to run, a time mask and a week day mask.

• A yes/no indicator if intermediate results shall be reportedafter each recording period.

A recording configuration is rejected if the recording overlapsanother FAS recording in time and they have one or more cellsin common. It is also rejected if more than 10 simultaneousrecordings are ordered.

Recording Procedure

Recordings are controlled by OSS, which will initiate the RadioInterference Recording function in one or more BSCs, accordingto the time schedule specified by the user.

During a recording period, each TRX (within the cell set)measures the uplink interference repeatedly for the frequenciesspecified by the operator. The measurements are made with thesame resolution and range as normal RXLEV measurements.For each frequency a sample is taken once every 15 seconds.One sample represents the signal strength during one burst.Samples are taken on idle traffic channels and if necessaryduring the idle bursts on busy traffic channels. Each frequencywill be sampled on a randomly chosen (possible) channel.Samples will only be taken on the following logical channelcombinations:

1. TCH/F + FACCH/F + SACCH/TF

2. TCH/H(0.1) + FACCH/H(0.1) + SACCH/TH(0.1)

where the second case applies as long as there is no more thanone half-rate user on that channel. Maximum four bursts on thesame channel during one SACCH period (480 ms) are used forFAS measurements, irrespective of whether the channel is idleor not. During the 15 second cycle, the first 10 seconds are used


EN/LZT 123 3314 R3A – 245 –

for measurements on idle channels only. If the measurements arenot completed after that, busy channels will also be used theremaining 5 seconds. The measurements are not distributeduniformly in time but performed as fast as possible. This meansthat the specified frequencies may be measured during less timethan 15 seconds.

If traffic load statistics are available in the PerformanceManagement database, these are fetched for each FAS recordingperiod. From the traffic load statistics, validity counters arederived. The validity counters are stored together with theinterference data and may be used to determine the validity ofthe FAS measurements.

Recording Evaluation

After the final recording period, results will be retrieved fromeach Base Transceiver Station (BTS). If more than one TRX isallocated to a BTS, the retrieved results will be the average ofthe results for each TRX (at that BTS). The results will contain:

• The number of samples per measured frequency

• The median interference, in RXLEV, per measuredfrequency

• The Xth percentile value (in RXLEV) of the interferencedistribution per measured frequency, where X is preset in therecording configuration.

If intermediate results are requested, the results above will alsobe retrieved after each recording period.

If traffic load statistics are fetched, there will be two validitycounters per cell:

• The average traffic load during the recording periods

• The lowest quotient between the traffic load of any tworecording periods

If traffic load was low or irregular for some cells, FASmeasurements considered affected by that may be discarded (bythe user).

At the OSS user interface the results will be compiled andpresented as a report or graphically on a map. The operator canmodify the report to focus on different types of data. Forinstance, only interference data for the cells with new frequency


– 246 – EN/LZT 123 3314 R3A

allocations, and their current co-frequency and/or adjacentfrequency cells may be presented.

Upon request, FAS will also suggest a Base Station IdentityCode (BSIC) for a given BCCH frequency and cell.


EN/LZT 123 3314 R3A – 247 –

EFFICIENT PRIORITY HANDLING (EPH)

INTRODUCTION

The Efficient Priority Handling (EPH) feature enables theoperator to give high priority users good access to the network.This is achieved by means of removing connections of lowpriority users from a congested cell required by a high priorityconnection. Up to two attempts to hand over a low priorityconnection to another cell are made. If the handovers fail thelow priority connection is disconnected.

EPH is implemented in the Base Station Controller (BSC).


High priority users can be given very good access to the networkby allowing them to seize channels already occupied by lowpriority users. It is also possible to classify users as neithercapable of removing other connections, nor being vulnerable toremoval. The low priority connections are primarily handed overto other cells, but are disconnected if necessary.

EPH is applicable for both Stand-alone Dedicated ControlCHannels (SDCCHs) and Traffic Channels (TCHs), but not atimmediate assignment.

EPH is not applicable for transparent (fixed bandwidth) HighSpeed Circuit Switched Data (HSCSD) connections with morethan one timeslot, i.e. these connections can not initiatehandover or disconnection of low priority connections. Lowpriority HSCSD connections may however be subject to forcedhandover or disconnection.


General

For each connection there is one information bit, the Pre-emption Capability Indicator (PCI), that determines if theconnection may remove another connection. There is also onebit, the Pre-emption Vulnerability Indicator (PVI), thatdetermines if the connection may be removed by anotherconnection. The Mobile services Switching Center (MSC) sends


– 248 – EN/LZT 123 3314 R3A

the PCI and the PVI to the BSC in the ASSIGNMENTREQUEST and the HANDOVER REQUEST messages.

Transparent multislot connections are not allowed to removelow priority connections. If a non-transparent multislotconnection has PCI set, it is only the main channel that canremove a PVI connection (a connection with PVI set).

Following traffic cases do not initiate EPH:

• Handovers due to Cell Load Sharing

• Intra-cell handovers due to halfrate packing

• Subcell change due to blocking, or Subcell LoadDistribution, or if overlaid subcell is preferred

Releasing a channel

The first cell in the candidate list is tried for channel allocationat a handover or assignment attempt. If this cell is congested therest of the cells are tried one after each other. With EPH active,the search for a channel to release in favor of a PCI connectionis started immediately when the first cell is reported ascongested. If no channel can be found in the cell, the next cellon the candidate list is tried.

The cell is searched for PVI connections of any resource typethat is allowed by Channel Administration. The connections arenot chosen with respect to the priority between the resourcetypes. The first allowed channel, assigned to a PVI connection,that is found in the list containing all channels is reserved for thePCI connection.

If no suitable PVI connection can be found EPH reportscongestion and the next cell candidate is tried, if there is any lefton the candidate list.

If a suitable PVI connection is found, handover candidates forthis connection are requested from Locating and a handover ofthe PVI connection is initiated.

If no candidates can be found, or the handover attempt fails,EPH searches for another PVI connection starting with the lastreported connection from the channel list. If no more PVIconnections can be found the last reported connection isdisconnected. If a suitable second PVI connection is found, thesame procedure is performed as for the first connection. If thissecond handover attempt fails a third search for a PVI


EN/LZT 123 3314 R3A – 249 –

connection is performed. This is due to that it may take severalseconds for a failed handover to leave the channel for otherpurposes.

The last found, second or third, PVI connection is disconnected.The procedures are timer supervised, allowing two seconds forhandover and disconnection. If a handover attempt is notsuccessful within two seconds the attempt fails and theprocedure continues as is described in the preceding paragraph.There is a total time limit of seven seconds for assignments thatis always used, independently of EPH. If less than four secondsare left of the seven seconds, any handover attempt of the PVIconnection is interrupted and a disconnection is started.

Before a channel, previously belonging to a PVI connection, isreleased it is reserved for the PCI connection.

If the PVI connection is a HSCSD connection, handovercandidates are only requested for the main channel due to thefact that only the main channel is evaluated by Locating for aHSCSD connection. A handover of the whole multislotconnection is tried if at least one handover candidate is received.If the handover fails, a reduction of the bandwidth (number oftimeslots) is tried. This is not possible for transparentconnections. The bandwidth reduction corresponds todisconnecting one of the non-essential channels (all channels ina transparent connection and the main channel in a non-transparent connection are essential). If the bandwidth reductionfails the available multislot PVI connection is disconnected.


– 250 – EN/LZT 123 3314 R3A

NEIGHBORING CELL SUPPORT (NCS)

INTRODUCTION

The purpose of the Neighboring Cell Support (NCS) feature isto help the operator maintain well-working neighboring cell listsin the network. This is done by collecting data recorded fromMeasurement Reports and handover statistics. The data can bepresented in tables and charts for an analysis of whatneighboring cell relations should be defined in the network.

NCS is operated through the Operation and Support System(OSS).


NCS can be used to optimize the neighboring cell lists in thenetwork. The results presented by NCS can be analyzed to findwhat neighboring cell relations to add or remove in the network.

With NCS, the user can schedule recordings that will loginformation from Measurement Reports on both defined andundefined neighboring cells. During the recording, userspecified test frequencies are temporarily added to the activemode BA lists. This gives a chance also for undefinedneighboring cells to occur in Measurement Reports.

The recording results can be presented both on an overviewlevel and on a deeper level. On the overview level conciseinformation for all cells in a recording is presented. On the moredetailed level information for all reported neighboring cells areshown for one cell. Together with the recording data on thedeeper level, handover statistics for defined neighboring cellrelations can also be presented. The tables in which data arepresented can be sorted and/or filtered to further simplify theanalysis. In order to help with the evaluation of the results,recording results as well as the defined neighboring cellrelations can be shown by color-marking the cells in the OSSmap.

NCS also supports transfer of changes of the neighboring cellrelations to the OSS application Cellular NetworkAdministration (CNA). From CNA, the modifications can beimplemented.


EN/LZT 123 3314 R3A – 251 –


General

NCS is operated from OSS. OSS initiates the Base StationController (BSC) function Active BA List Recording accordingto the user’s settings. Handover statistics will be fetched fromthe Performance Management database in OSS. This means thatthe collection of handover statistics in the system has to be setup outside of the NCS application.

The activities in NCS can be divided into three parts:

1. the recording configuration, when the user schedules arecording

2. the recording procedure, when the system carries out theactivities specified in the recording configuration

3. the recording evaluation, when the user requests apresentation of recording data

Recording Configuration

In the recording configuration, the user specifies when, whereand what to record.

All parameters as well as the test frequencies in the recordingconfiguration are common for all cells. This means that in orderto set individual thresholds for different cells, several recordingshave to be initiated.

The recording configuration consists of:

• A cell set, defining where to record. A cell set may containcells from more than one BSC.

• A list of test frequencies, called a frequency set, specifyingwhat frequencies to add to the active mode BA lists.

• A time schedule that defines when recordings should bemade.

• A relative signal strength threshold.

• An absolute signal strength threshold.

• The length of recording segments.

• The maximum number of test frequencies that may be addedto the active mode BA lists at one time.


– 252 – EN/LZT 123 3314 R3A

If only one cell has been specified for a recording, thefrequencies can as an option be selected to be the BCCHs ofsecond order neighbors, i.e. the cell’s neighbors’ neighbors.NCS will then automatically generate a list of test frequencies.This option is useful for trouble-shooting, as this list is based ona clever guess of which BCCH carriers that are of interest. Ashorter list of test frequencies will result in a more efficientrecording.

Recording Procedure

Recordings are controlled by OSS, which will initiate the ActiveBA List Recording function in one or more BSCs, according tothe time schedule specified by the user. The time schedule maycontain intervals when the recording is active, i.e. logging data.Such an interval is called a recording period.

During a recording period, the test frequencies are added to theactive mode BA list according to the following procedure.

For each cell, the BA list is extended with as many as possibleof the test frequencies that are not already in the list. Themaximum number of frequencies that can be added at a time isspecified as a parameter in the recording configuration.However, the BA list can never contain more than 32frequencies.

When all test frequencies cannot be included at the same time,the recording is divided in segments. This means that after thelength of a recording segment the logging of data is interruptedand the test frequencies included in the first segment areexchanged for new ones and the recording begins again. Thisprocedure is repeated until the end of the recording period.When all test frequencies have been added to the list once, thefirst ones are included over again.

During a recording, some general data is logged:

• total recording time

• total number of Measurement Reports received

• average signal strength of serving cell (downlink)


EN/LZT 123 3314 R3A – 253 –

The following data is recorded for each combination offrequency and BSIC:

• number of times the combination occurred in MeasurementReports

• average signal strength of all occurrences

• number of times it was reported above the signal strength ofserving cell plus the relative signal strength threshold

• number of times it was reported above the absolute threshold

• number of times it was the only one in the MeasurementReport

• number of times it was reported as the strongest, secondstrongest, etc. up to sixth strongest frequency/BSIC

• average signal strength when reported as strongest, secondstrongest, etc. up to sixth strongest frequency/BSIC

In addition, the following is recorded for each test frequency:

• total recording time with this frequency in the active modeBA list (This may differ from total recording time, becauseof the segmentation of the recording. )

• total number of Measurement Reports with this frequency inthe active mode BA list

At the end of a recording period, the active mode BA list in eachcell is restored to its original content.

Results are reported from the BSC(s) to OSS after eachrecording period, which means that for recordings of more thanone recording period it is possible to look at intermediate results.The results are accumulated and the intermediate results areoverwritten when the next recording period is finished and newresults are ready.

If handover statistics are available in the PerformanceManagement database, handover data for each recording periodin NCS are fetched for each cell in the recording.

A recording can be stopped at any time. With the stop order, itcan be specified whether the results obtained so far shall besaved, or if they shall be discarded.

The cells in a recording can be included in new NCS recordings,as soon as the previous recording result is ready.


– 254 – EN/LZT 123 3314 R3A

Recording Evaluation

*HQHUDO

The user can select to look at any stored recording result. Theresults can be presented in reports, charts and graphically in theOSS map.

Within NCS there are three different reports:

1. An overview report presenting only concise BA listrecording data for all cells included in the recording.

2. A cell report presenting the most important categories ofdata from the BA list recordings for one cell. Handoverstatistics for defined neighboring cells are also shown.

3. A detailed cell report presenting all data from the recordingfor one cell.

2YHUYLHZ�5HSRUW

In the overview report, the following is shown for each cell:

• The highest percentage of Measurement Reports, for anyfrequency/BSIC combination of undefined neighbors, wherethe frequency/BSIC was reported as above the relative signalstrength threshold.

• The lowest percentage of Measurement Reports, for anyfrequency/BSIC combination of defined neighbors, wherethe frequency/BSIC was reported as above the relative signalstrength threshold.

From the overview report, the user can decide for which cellsthe results are interesting to look at in more detail. This is usefulas one recording may be done for up to 1000 cells.

&HOO�5HSRUW

The cell report is divided into two parts, one for the definedneighboring cell relations and one for undefined neighboringcell relations (frequency and BSIC reports that do notcorrespond to a defined neighboring cell). The following isshown for each frequency and BSIC that was reported during therecording:

• Cell name behind the frequency/BSIC combination. Forundefined neighbors, the cell name is the most likely onejudging from distance to surrounding cells of that frequencyand BSIC.


EN/LZT 123 3314 R3A – 255 –

• Number of reports above the relative threshold

• Number of reports as number 1 (i.e. strongest)

• Average signal strength when reported as number 1

For the part of the report with defined neighbors the followinghandover statistics are shown: handover attempts, handoversuccessful, handover reversions.

'HWDLOHG�&HOO�5HSRUW

In the detailed cell report all data from the recording ispresented. However, handover statistics are not shown.

2WKHU�2SWLRQV

In both the cell report and the detailed cell report, it is possibleto choose if the counters for number of reports shall be shown inpercent of all Measurement Reports, or in absolute numbers.

All tables in reports can be sorted in ascending or descendingorder on any column, and they can also be filtered on anycolumn, e.g. in the cell report, the table can be filtered on thenumber of reports ranked as number one, so that onlyfrequency/BSIC combinations reported as strongest neighbor inmore than, say, 1.0 % of the Measurement Reports are shown.

The data in the cell report can also be presented in a bar chart.

Numerical data in the overview report and the cell reports canalso be presented by color marking cells on a map.

Looking at the data presented in reports, it may be found thatmodifications of the neighboring cell definitions should bemade.

Orders of such changes can be prepared in NCS and exported tothe Cellular Network Administration application in OSS, fromwhere the actual implementation of the change can be made.

Radio Cell planning principles

Technology

channel planning

planning process

document number

training document

system description

system balance

system tuning

system growth