Sie sind auf Seite 1von 86

GSM Frequency

Hopping and Variable


Interference Planning
RF Engineering Guideline
EG: GSMVIP

401-380-365
Issue 1.4
January 2001

Lucent Technologies - Proprietary


This document contains proprietary information
of Lucent Technologies and is not to be disclosed or used
except in accordance with applicable agreements

Copyright 2000 Lucent Technologies


Unpublished and Not for Publication
All Rights Reserved
Copyright ©2001 by Lucent Technologies. All Rights Reserved.

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

For permission to reproduce or distribute, please contact:


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

Notice

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

1. ABOUT THIS GUIDE 1

2. INTRODUCTION TO FREQUENCY HOPPING 3

2.1. Frequency hopping overview 3

2.2. Why use frequency hopping 5

Multipath fading 5
Interference 5

2.3. Hopping sequences 6

Cyclic hopping 6
Random hopping 6

2.4. Hopping at the base station 6

Baseband hopping 6
Synthesiser hopping 7

2.5. GSM network implementation 8

Sequence generation 8
Common control channels 8
Reception level measurements 9
Quality measurements 9
Frequency redefinition procedure 10
General Packet Radio Service (GPRS) network 10
Mobile stations 10

2.6. Key benefits 11

Issue 1.4 – January 2001 Lucent Technologies – iii


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Frequency diversity 11
Impact on network planning 16
Interference diversity 16
Associated techniques 18
Impact on network planning 23

3. INTRODUCTION TO VARIABLE INTERFERENCE PLANNING 25

3.1. VIPone 25

VIPone properties 28
VIPone examples 28

3.2. VIPtwo 29

VIPtwo properties 29
VIPtwo examples 30

3.3. VIPone and VIPtwo compared 31

Combined plans 31

4. CONFIGURING FREQUENCY HOPPING 33

4.1. Base station hardware 33

Base model 33
Antenna coupling equipment 34

4.2. Software release support 35

4.3. Configuration 35

FHS configuration rules 35


Other limitations 40

4.4. Feature activation and system parameters 40

BTS hopping mode 40


BSS feature enabling 41
OMC parameter configuration 41
Feature activation 42

4.5. Fault management 43

iv Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Baseband hopping 43

4.6. DTX 44

Uplink DTX 44
Downlink DTX 44

4.7. Dynamic power control 45

5. VARIABLE INTERFERENCE PLANNING DEPLOYMENT 47

5.1. Introduction 47

When to use VIP 47


Implementation strategy 48

5.2. Choosing the right plan 48

More than three transceivers per cell 48


Three or fewer transceivers per cell 49
Large spectrum allocation 49
Microcells 49
Planning for future capacity 49

5.3. Planning the frequencies and the HSN 49

VIPone 49
VIPtwo 50
VIPone/VIPtwo 50
Microcells 50
BCCH planning 51

5.4. Collecting performance data 52

Collection equipment 52
Performance data types 52

5.5. Deployment results 56

Activating frequency hopping 56

5.6. Optimising performance 60

Quality-based handovers 60
Quality-based power control 60

Issue 1.4 - January 2001 Lucent Technologies – v


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Hopping over two frequencies 60


Discontinuous Transmission measurement accuracy 60
Other scenarios 61

6. WORKED EXAMPLES 63

6.1 Scenario 1 63

Existing configuration 63
Objectives 64
VIP plan choice 64
Planning the frequencies 64
Mapping the frequency plan to OMC settings 65

6.2 Scenario 2 66

Existing configuration 66
Objectives 66
VIP plan choice 66
Planning the frequencies 66
Mapping the frequency plan to OMC settings 67

6.3 Scenario 3 69

Existing configuration 69
Objectives 69
VIP plan choice 69
Planning the frequencies 69
Mapping the frequency plan to OMC settings 70

6.4 Scenario 4 71

Existing configuration 71
Objectives 71
VIP plan choice 71
Planning the frequencies 71
Mapping the frequency plan to OMC settings 72

6.5 Scenario 5 73

Existing configuration 73
Objectives 73

vi Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

VIP plan choice 73


Planning the frequencies 73
Mapping the frequency plan to OMC settings 74

7. LIST OF ACRONYMS 75

Issue 1.4 - January 2001 Lucent Technologies – vii


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

This page is intentionally left blank

viii Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

About this Guide

1
1. About this Guide

This guide provides a detailed description of the Frequency Hopping and Variable Interference
Planning (VIP) solutions offered by Lucent Technologies for GSM 900 and 1800 networks. It
contains the following chapters:

• Chapter 2 Introduction to Frequency Hopping


An overview of frequency hopping concepts and techniques, and their benefits and network
impacts.

• Chapter 3 Introduction to Variable Interference Planning


An overview of VIP concepts and techniques, and their benefits and network impacts.

• Chapter 4 Configuring Frequency Hopping


Describes how to configure and activate frequency hopping in a network, from the
equipment point of view.

• Chapter 5 Variable Interference Planning Deployment


Describes when and how to implement VIP into a network.

• Chapter 6 Worked Examples


Examples of different scenarios and suggested implementations.

• Chapter 7 List of Acronyms


Definitions of the acronyms used in this guide.

Issue 1.4 – January 2001 Lucent Technologies – 1


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

This page is intentionally left blank

2 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Introduction to Frequency
Hopping

2
2. Introduction to Frequency Hopping

This chapter describes the main concepts of frequency hopping and its implementation in GSM.

2.1. Frequency hopping overview

In frequency hopping systems, each call hops between a defined set of frequencies. Poor signal
quality on any specific frequency therefore affects only a small portion of the transmission. This
makes it much easier to recreate any lost transmission bits and preserve overall call quality.

Frequency hopping is the principal component of the Variable Interference Planning solutions
offered by Lucent Technologies, and is supported in both GSM 900 and GSM 1800 networks.

GSM networks use “slow” frequency hopping; a hop occurs before each time slot is transmitted
(every 4.615 millisecond, or 217 hops per second). This distinguishes it from fast frequency
hopping systems, which use several hops per symbol. This slow frequency hopping is illustrated
in Figure 1.

Issue 1.4 – January 2001 Lucent Technologies – 3


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

frequency

time

Figure 1: Slow frequency hopping in the time frequency domain

Frequency hopping exploits two underlying GSM error correction techniques:

• Channel coding

• Interleaving

These coding and interleaving techniques are shown in the following illustration.

Class 1 bits P+T Class 2 bits


182 bits 3+4 78 bits

Rate 1/2 convolutional coding

456 bits
Segmentation
and
interleaving

Normal burst - 1 Normal burst - 2 ... Normal burst - 8


57 bits 57 bits 57 bits

Tail Information SF TS SF Information Tail Tail Information SF


... 3 bits 57 bits 1 bit
3 bits 57 bits 1 bit 26 bits 1 bit 57 bits 3 bits

Normal GSM burst Normal GSM burst

Figure 2: GSM coding and interleaving

4 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Channel coding takes the digital message flow (speech or data) and divides the bit stream into
blocks. Control bits used to detect and correct transmission errors are applied to the start and
end of each block. Each block of message bits and control bits is known as a “code word”.
Code words for speech are 456 bits long.

Interleaving divides each code word into chunks of 57 bits at a time and mixes (interleaves)
them with chunks from adjoining code words. Splitting the bit stream in this way prevents errors
occurring across entire code words. This improves channel coding correction rates, as it is
much easier to correct isolated bit errors than bursts of errors.

2.2. Why use frequency hopping

Frequency hopping mitigates two problems with transmission quality over the air interface:

• Multipath fading

• Interference

Multipath fading

Usually a radio signal is received as scattered signals travelling over separate paths. When the
signals combine, they produce an interference pattern of fading. For a given position the fading
depends on the transmission frequency. This multipath fading particularly impacts slow moving
mobiles, as they may stay in one position and fade long enough to suffer information loss
(interleaving can only spread a code word over a limited number of time slots).

With frequency hopping, because the frequencies change, so do the fading patterns associated
with them. Transmissions on a frequency that is subject to multipath fading, will move out of the
fade at the next hop (“frequency diversity”).

Frequency diversity, combined with interleaving and channel coding, improves transmission
quality - in particular for slow moving mobiles.

Interference

Any specific call may suffer interference because of calls on neighbouring cells transmitting on
or close, to the specific call frequency. This interference has a continuous impact on
transmissions because it exists for the duration of the interfering calls.

Frequency hopping mitigates this effect by spreading, or averaging, the interference across
multiple calls (“interference diversity”). This prevents a situation where one call has
unacceptable levels of interference and others have very good levels. When coupled with
channel coding and interleaving, frequency hopping increases the probability that all calls will
have acceptable quality, rather than some having very good quality and others having
unacceptable quality.

Issue 1.4 – January 2001 Lucent Technologies – 5


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Interference diversity has another advantage. It ensures that consecutive bursts of information
are received under different interference conditions, reducing the risk of large sequential
information loss.

2.3. Hopping sequences

In frequency hopping systems, the hopping sequence between the frequencies assigned to a
particular transmission can be either cyclic or random.

Cyclic hopping

Frequencies are used in fixed rotation.

For example: f1, f2, f3, f4, f1, f2, f3, f4, f1, f2, f3, f4, f1, f2, f3, f4, f1, f2, f3,...

Random hopping

Frequencies are used in a pseudo-random sequence.

For example: f2, f4, f1, f3, f4, f2, f3, f2, f4, f1, f1, f4, f3, f4, f2, f1, f3, f2, f2, ...

When using the same set of frequencies with random hopping, the probability of two calls using
the same frequency in the same time slot is 1/N, where N is the number of hopping frequencies.

For example: Mobile Station 1: f1, f4, f4, f2, f1, f3, …; Mobile Station 2: f2, f1, f4, f3, f2, f1, …

2.4. Hopping at the base station

Frequency hopping can be generated in two ways:

• Baseband hopping

• Synthesiser hopping

Baseband hopping

In baseband hopping, each transceiver within a base station operates on fixed frequencies.

A transceiver provides the functionality of eight channels, according to the GSM air interface.
Transceivers perform both baseband signal processing (channel coding, interleaving,
encryption, and Time Division Multiple Access (TDMA) burst information) and RF signal
processing (generation of RF signal and modulation of TDMA bursts).

In Lucent transceivers (also known as TRXs or RTs) the DRCC (Digital Radio Codec and
Control) unit does the baseband processing and the RFU (Radio Frequency Unit) does the RF
processing.

6 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

With baseband hopping, the digitised, or baseband, speech signal generated at the DRCC is
switched between the RFUs of the transceiver before transmission.

Each frame of eight timeslots is input to a different RFU and so to a different frequency. In this
way the transceivers do not need to retune to different frequencies, but each channel effectively
hops over the available frequencies.

The primary limitation of baseband frequency hopping is that the number of hopping
frequencies is limited to the number of RTs (Radio Terminals) in the cell.

D
DRRC
CCC R
R FF U
U f1

D
DRRC
CCC R
R FF U
U f2

Figure 3: Baseband hopping

Synthesiser hopping

With synthesiser hopping, each RFU within a transceiver retunes to a different frequency
(following a defined hopping sequence) before transmitting a frame. Therefore, unlike baseband
hopping, the output of each baseband processing section is always connected to the same
RFU. This allows each transceiver to hop over as many frequencies as desired, regardless of
the number of transceivers in the cell.

f1
D
DRRC
CCC R
R FF U
U f2

f2
D
DRRC
CCC R
R FF U
U f1

Figure 4: Synthesiser hopping

Issue 1.4 – January 2001 Lucent Technologies – 7


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

However, traditional filter combiners (which are frequency specific) cannot be used with
synthesiser hopping because they are too slow in changing frequency. Hybrid combiners (which
can operate across a frequency range and are therefore also known as wide-band combiners)
must be used instead. Because hybrid combiners have much higher insertion losses than filter
combiners, the maximum number of radios per cell is reduced.

2.5. GSM network implementation

This section describes how frequency hopping is implemented in GSM networks.

Sequence generation

Each call has its time slots transmitted in sequence across a defined set of hopping
frequencies. The sequence is derived from an algorithm (see GSM Recommendation 05.02).
Frequency hopping occurs between time slots; a mobile station transmits or receives on a fixed
frequency during one time slot, then changes frequency before the time slot on the next TDMA
frame.

The total number of available hopping sequences is 64 multiplied by the number of hopping
frequencies (64xN). Hopping sequences are described per channel, by two network
parameters:

• HSN (Hopping Sequence Number): defines a number that is fed into the frequency hopping
algorithm to generate the hopping sequence. Values can be 0 to 63. Value 0 defines cyclic
hopping; all other values generate a pseudo random sequence

• MAIO (Mobile Allocation Index Offset): defines the starting frequency, or offset, the
transmission will start on within a hopping sequence. The value can be 0 to N-1 where N is
the number of allocated frequencies

Two channels with the same HSN but a different MAIO are the shift version each other.
Therefore, they will never use the same frequency at the same time.

Two channels using the same frequency list and the same time slot, but with a different and non
0 HSN, will interfere in 1/Nth of bursts, as if the sequences were chosen randomly.

Channels in the same cell using the same hopping frequency set should have the same HSN,
and different MAIO, to avoid co-channel interference within the cell.

If random hopping is used, each channel in distant cells using the same frequency set should
have a different HSN; this optimises the benefits of interference diversity.

Common control channels

In order to ease initial synchronisation acquisition, the following common control channels must
use a fixed frequency:

8 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

• FCCH (Frequency Correction Channel)

• SCH (Synchronisation Channel)

• BCCH (Broadcast Control Channel)

• PAGCH (Paging Access Grant Channel)

• RACH (Random Access Channel)

Common channel extension sets must use the same fixed frequency as the primary group. This
avoids the need to transmit their frequency organisation description on the BCCH.

Note: Traffic channels on the rest of the time slots in BCCH transceivers can hop. Only the
common channels cannot hop.

Reception level measurements

If dynamic power control is in use, and frequency hopping occurs on the BCCH frequency
among other frequencies, reception level measurement accuracy is an issue in traffic channels
that use this combination.

Power control cannot be applied on the BCCH frequency, which must transmit at constant
power in the downlink. This means that power control applies to a subset of bursts only. Bursts
that use the BCCH frequency are sent at fixed transmission power. If reception level
measurements in the downlink were to be averaged on all frequencies, including the BCCH, the
measurements would not be accurate for the power control algorithm.

To alleviate this problem, the power control indicator tells the mobile station to ignore BCCH
frequency slots in reception level estimations. The indicator is sent to the mobile station at
connection, if the following conditions are met:

• The channel hops on at least two frequencies

• One of those frequencies is the BCCH frequency

• Dynamic power control is in use on the downlink transmission

Quality measurements

Mapping Received Signal Quality (RXQUAL) measurements to subjective speech quality, varies
with the propagation environment. This is because it is a measure of the raw bit error rate,
estimated by backward coding the decoded bit sequence and comparing it to the received bit
sequence. Hence it does not consider the varying efficiency of channel coding, interleaving and
bit error correction under different environmental conditions.

So under different conditions the same RXQUAL values can result in different actual speech
quality, and calls with different RXQUAL values can have the same speech quality.

Issue 1.4 – January 2001 Lucent Technologies – 9


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

The interference diversity property of frequency hopping means that interference conditions
vary from time slot to time slot. This means that with frequency hopping, even in the same
propagation environment, calls with the same RXQUAL can have different speech quality and
vice-versa.

Accordingly, when frequency hopping is used, RXQUAL is not a reliable measurement of


connection quality. Therefore to assess the quality of the network with frequency hopping, the
operator should use other quality indicators, such as FER (Frame Erasure Rate) or subjective
voice quality indicators. These indicators are only available in drive test equipment.

Frequency redefinition procedure

This procedure is used dedicated mode, to minimise disruption to calls when channel
frequencies and hopping sequence allocations change in the network. When this happens, the
network sends a FREQUENCY_REDEFINITION message to the mobile stations that are
currently in call. This contains the new parameters and a start time indicator. Parameters that
can be updated are the cell channel description, mobile allocation, and MAIO.

At the indicated time slot, the base station and assigned mobile stations update their allocated
frequencies and hopping sequences to match the new parameters. So, this time slot is the first
to use the new parameters. No other functions are normally disturbed by the change. However,
some calls may be lost in the following circumstances:

If the Mobile Switching Centre (MSC) requests a handover channel, and the request is
acknowledged with the actual channel information. Then, if a redefinition procedure
subsequently starts for that channel and the mobile station is handed over to the channel at the
same time, the call is lost.

General Packet Radio Service (GPRS) network

GPRS is supported in Network Release 9.1. However, frequency hopping on dual service
channels Packet Data Transfer Channel (PDTCH/TCH) is not supported. Therefore, for a
network which has frequency hopping, it is recommended that the dual service channels should
be allocated to the BCCH frequency. Since the BCCH frequency usually does not hop,
allocating dual service channels on this carrier will not degrade frequency hopping performance.

Frequency hopping will be supported on dual service channels in future releases.

Mobile stations

Currently there are some unresolved problems with certain types of older mobile station models:

• Some do not support the frequency redefinition procedure

• Some cannot hop on Standalone Dedicated Control Channel (SDCCH) channels, or have
problems when using frequency hopping in conjunction with Discontinuous Transmission
(DTX) on the downlink, or with dynamic power control

10 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

2.6. Key benefits

This section describes in more detail the primary benefits of frequency hopping:

• Frequency diversity

• Interference diversity

Frequency diversity

Multipath fading is speed and frequency dependent. The high speed of some mobiles is enough
to allow GSM error correction to overcome its effects. For slower moving users, the correction
mechanisms are insufficient on their own. However, by using frequency hopping, the same
performance levels can be obtained for slow moving users.

Figure 5 compares the required carrier to noise (C/No) ratio as a function of vehicle speed for a
bit error rate (BER) of 0.5% (considered acceptable for speech) in the 900 MHz band, first at a
fixed frequency allocation, and then using ideal frequency hopping. Ideal frequency hopping
occurs when hopping takes place on uncorrelated frequencies. That is, their fades are
independent of each other.

14
C/No [dB] for BER=0.5%

12

10 Without FH

8 FH

4
0 50 100 150 200
v [km/h]

Figure 5: Required C/No against vehicle speed (BER 0.5%) – 900 MHz

Without frequency hopping the performance of the system depends on the vehicle speed. The
faster the mobile, the better the error correction mechanisms work and the lower the minimum
signal to noise ratio must be to achieve a certain BER.

With ideal frequency hopping (infinite number of frequencies and infinite separation between
them), optimum transmission quality is obtained at almost all vehicle speeds.

Note: Slight degradation occurs with frequency hopping at very high vehicle speeds. This is
caused by a significant change in the multipath profile at the time slot level that cannot be
tracked by the equaliser.

Issue 1.4 – January 2001 Lucent Technologies – 11


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Similar improvements are gained in co-channel or adjacent-channel interference. Figure 6


compares the required carrier to interference (C/I) ratio in terms of current vehicle speed for
fixed frequency, and ideal frequency hopping operation. The dependence is even more marked
than for noise interference, as in this situation the power of the interference signal also
fluctuates with the speed.

18
C/I [dB] for BER=0.5%

16
14
Without FH
12
FH
10
8
6
0 50 100 150 200
v [km/h]

Figure 6: Required C/I against vehicle speed (BER 0.5%) – 900 MHz

Optimising frequency diversity

Frequency diversity optimisation is governed by two factors:

• Number of frequencies

• Frequency spacing

Number of frequencies

Ideal frequency diversity requires that a different frequency is used for each time slot within an
interleaved code word.

If this is not the case, at least two of the time slots over which a code word is spread are
transmitted at the same frequency. The fading effect is strongly correlated for them at low
velocity, thus reducing the gain.

With cyclic frequency hopping, to achieve ideal frequency hopping, the hopping period must be
at least as long as the interleaving depth (eight time slots for speech). This ensures a different
frequency in each time slot. A longer period does not provide additional frequency diversity
gains.

Figure 7 uses bit error curves to illustrate the likely performance losses at low vehicle speed (5
km/h). Note that hopping over just four frequencies comes as close as 1dB to the gain of
hopping over eight frequencies.

12 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

0.1 Without FH
2 f reqs
BER

4 f reqs
0.01 8 f reqs

0.001
0 2 4 6 8 10 12 14
C/No [dB]

Figure 7: Effect of number of frequencies on BER (v=5 km/h)

With random hopping, the probability of using the same radio frequency channel within the
interleaving depth is depth/N, where N is the number of frequencies in the hopping sequence.
This means that the fading decorrelation within one interleaving block is never optimal,
regardless of the number of hopping frequencies.
-2
The following table shows the C/No required for a BER of 10 using both cyclic and random
frequency hopping over different numbers of frequencies. Note that the frequency diversity gain
with eight frequencies is 1 dB to 2 dB lower for random hopping than for cyclic.

Similar results would be expected with interference.

Issue 1.4 – January 2001 Lucent Technologies – 13


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Cyclic hopping Random hopping

C/No for Gross C/No for C/No for Gross C/No for
Class 11 FER2=2% Class 1 FER=2%

No. of Level Gain Level Gain Level Gain Level Gain


frequencies
[dB] [dB] [dB] [dB] [dB] [dB] [dB] [dB]

1 9.5 0.0 11.5 0.0 9.5 0.0 11.5 0.0

2 7.0 2.5 8.5 3.0 7.5 2.0 9.5 2.0

3 6.0 3.5 7.5 4.0 6.5 3.0 8.5 3.0

4 5.0 4.5 6.5 5.0 6.0 3.5 8.0 3.5

8 4.0 5.5 5.5 6.0 5.5 4.0 7.5 4.0

12 4.0 5.5 5.5 6.0 5.0 4.5 7.0 4.5

Table 1: Frequency diversity gains

The results in this table cannot be compared directly with the previous figures because different
propagation conditions apply in different environments, particularly in typical urban (TU) areas.

Note: Frequency hopping gains may be smaller than predicted, due to the diminished severity
of multipath propagation when compared to flat fading. In normal environments the different
paths arrive at different times, thus cancelling out some of the fading impact.

Frequency spacing

Frequency spacing must be sufficient to ensure that uncorrelated fading affects different
frequencies.

The coherence bandwidth can be defined as the frequency separation required for propagation
paths, and hence fading, to be considered totally independent. In most outdoor environments,
coherence bandwidths of less than 1MHz can be expected, so 1MHz (5 GSM channels) is
recommended as the minimum frequency spacing for outdoor systems. However, in TU
environments, channel separation of 400kHz to 600kHz (2 to 3 GSM channels) is enough.

1 Bits produced by the GSM encoder are ranked in importance as Class 1 and Class 2. Class 1 bits are
protected by redundancy codes.

2 Frame Error Rate (FER) is the fraction of entire speech frames erased by the speech decoder
because of irrecoverable bit errors.

14 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Indoor environments are generally characterised by large coherence bandwidths. Typically, an


indoor frequency hopping system gives lower frequency diversity gains than an outdoor system
with the same hopping bandwidth. However, as indoor users are generally slow moving, there is
still potential for frequency diversity gains from frequency hopping.

Simulations show that although the gain achieved is smaller, it is still significant: Assuming 5
MHz bandwidth and an FER of 2%, the gain in C/No is between 1.7 dB and 3.3 dB (compared to
5 dB in typical urban areas).

Antenna diversity

Antenna diversity is another technique used to combat multipath fading. Like frequency
hopping, it achieves gains in conjunction with channel encoding and interleaving, but since it
uses space rather than frequency diversity, the gain is independent of vehicle speed.

Combining frequency hopping with antenna diversity produces significantly increased gains.
However, the total gain does not equal the sum of the individual gains. The following table
illustrates the likely gain in C/No (in dB) for an FER of 2% when using antenna diversity with
frequency hopping:

C/No for FER of 2% C/I for FER of 2%

No FH Ideal FH No FH Ideal FH

Level Gain Level Gain Level Gain Level Gain


[dB] [dB] [dB] [dB] [dB] [dB] [dB] [dB]
No diversity 12.5 0.0 5.5 7.0 15.5 0.0 7.3 8.2
Ideal diversity 5.8 6.7 1.8 10.7 8.0 7.5 3.2 12.3

Table 2: Effects of antenna diversity

Note: In tests, antenna diversity gains are high in TU environments but drop in other test
conditions such as rural areas and hilly terrain.

Issue 1.4 – January 2001 Lucent Technologies – 15


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Impact on network planning

Existing radio network planning is generally based on poor transmission conditions at slow
vehicle speed. Frequency hopping compensates for this degradation in transmission quality,
making it largely independent of vehicle speed. Potentially, the smaller C/No and C/I values for
medium speed vehicles can be used when planning for areas of significant pedestrian use.

However, because the BCCH carrier cannot hop, the reduction in C/No that arises from
frequency diversity does not translate into reduced sensitivity values for BCCH planning.

Similarly, the improvement in C/I values does not translate into a tighter reuse pattern for BCCH
carriers. However, planning gains are obtained for TCH carriers. Here, different C/I
requirements can be set in the frequency planning process (for example, by using TU50 values
rather than TU3) allowing frequency diversity to be used to increase capacity.

The capacity increase depends primarily on the number of frequencies in the hopping
sequences (as discussed earlier, this affects the required C/I value). In addition, the separation
between frequencies assigned to a cell must be appropriate for the propagation environment.

To maximise the benefits of frequency diversity, where possible, the traffic channels on the
BCCH frequency should be included in the hopping sequences. However, no control channel
(BCCH and CCCH) should be in the hopping sequence.

Interference diversity

To-date, interference diversity has been primarily associated with Spread Spectrum systems.
Frequency hopping now enables GSM networks to exploit the benefits of interference diversity.

Example

This section illustrates the principle of interference diversity.

Figures 8 and 9 show a GSM system with two sectors that use the same set of four
frequencies. At a given time there are four mobile stations communicating in each of the cells.

In cell A, the mobile on f1 is suffering high interference levels, because the interfering mobile in
cell B is very near to the border (Figure 8). Speech quality is poor as a result.

The other mobiles in cell A are subject to lower interference levels (Figure 9). The actual
interference level and resulting speech quality varies across the mobiles, but, unlike the mobile
on f1, all yield acceptable speech quality.

16 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Figure 8: Example configuration without frequency hopping

Figure 9: Example configuration without frequency hopping

Issue 1.4 – January 2001 Lucent Technologies – 17


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Figure 10 shows the effect of switching on random frequency hopping between the four
assigned frequencies. The mobile in cell A that previously had high interference levels, now has
varying interference levels, because the interference from the mobiles in cell B varies with each
time slot. The rest of the mobiles in cell A, which previously had better quality, are in a similar
interference situation. This is interference diversity.

Figure 10: Effect of switching on random frequency hopping

Because the GSM channel coding and interleaving algorithm can correct interference errors in
time slots, the result is that all four mobiles in cell A now have acceptable speech quality.

In summary, the previous peaks and troughs in quality within the system are averaged to
produce acceptable quality across the whole system.

Note: If cyclic frequency hopping was used in this scenario, there would be no interference
diversity effect since the interfering mobile would always be the same.

Associated techniques

This section describes three techniques that can be used with frequency hopping to maximise
the benefits of interference diversity:

• Discontinuous transmission (DTX)

• Dynamic power control

• Fractional loading

18 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

DTX

Telephone traffic has alternating periods of silence and activity. The typical activity factor for
telephone conversations (the fraction of time a given user is actually speaking) is around 40%.
Data transmissions over switched circuits generally have an even lower activity factor.

In certain GSM transmission modes (in particular speech and non-transparent data) DTX
exploits this fact by inhibiting transmission of the radio signal when there is no information to
send (voice or data).

In the case of speech, the optimum goal is to encode speech at a bit rate of 13 kbps when the
user is talking, and at around 500 bps during silences (sufficient to generate background noise
so that the listener does not think the connection is broken).

Low encoding rates during silences result in decreased radio transmissions with a
corresponding reduction in channel interference levels, and improvement in quality.

Using DTX alone, this improvement in quality levels cannot be translated into increased
capacity since system planning must be done on a worst case basis. DTX is characterised by
an on/off nature. Peaks in interference levels are the same whether or not DTX is used, and the
rate of switching (between periods of activity and silence) is not high enough for channel coding
and interleaving to average out the variations.

However, when DTX is used with frequency hopping, the peaks in interference levels are
levelled out. The quality increase produced by the lower interference levels can now be
translated into tighter reuse and hence increased capacity.

Dynamic power control

Dynamic power control (or simply “power control”) regulates transmission power levels
dynamically during a connection. The mobile station and the base station can independently
reduce their power level when the received signal strength on the other end exceeds
requirements.

This conserves battery power in the mobile stations. But also, and importantly for frequency
hopping, by reducing overall power levels it reduces channel interference.

The following figure illustrates the typical situation in the downlink without power control. It
shows the C/I ratio perceived by a mobile station as a function of the distance to the base
station normalised to the distance between interfering base stations. To ensure acceptable
quality at the cell borders, significant power is wasted when the mobile station is near the base
station.

Issue 1.4 – January 2001 Lucent Technologies – 19


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

70

60

50

W asted power
40

30
C/I

20
Cell edge
10
Target C/I
0
0 0.2 0.4 0.6
-10
d/D

Figure 11: C/I ratio as a function of normalised distance (without power control)

Interference in this situation depends on the location of the interferer mobile station (the mobile
station to which the interfering base station transmits). This is illustrated in Figure 12, where the
C/I ratio is plotted, again with and without power control, for different interferer locations.

When the mobile station is at the cell border, base stations generally transmit at maximum
power3, with or without power control (C in C/I remains the same). In systems with power
control, a mobile station at the cell border only perceives the same C/I as without power control
when the interferer mobile station is also at the cell border. In the other cases, interference is
lower and the perceived C/I is higher.

Mobile stations near the base stations receive a lower signal strength (C) than without power
control. In some cases, this will result in a lower C/I ratio (when the interferer mobile station is
not near its base station and the reduction of power is less). However, as these mobile stations
had very good quality before, the degradation is not noticeable.

3 Power control parameters should generally be set so that mobiles at the cell edge transmit at full
power. This is to prevent unwanted interference effects that would take place if a mobile on the cell edge,
which starts its transmission at full power, had to regulate. In the time the mobile would take to reduce its
power, it would be causing high interference levels on mobiles that are already transmitting at the
required power.

20 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

80 without PC

60
PC, average
40

C/I
PC, interf. at cell
20 border
0 PC, interf. near
base-station
0 0.2 0.4 0.6 0.8
-20
d/D

Figure 12: C/I ratio as a function of normalised distance (with power control)

Something very similar happens in the uplink.

In summary, power control improves global quality (fewer calls suffer from bad C/I values)
which can be translated into a capacity increase. The gain, however, is not enough to allow a
jump from a 4/12 reuse factor to a 3/9 reuse factor, but its effect might be noticeable with
automatic planning tools that take power control into consideration.

When used with frequency hopping, power control generates more variation between the
interference signals, improving the performance of the averaging properties of frequency
hopping. This is shown on page 23.

Fractional loading

Networks are typically planned for full load on the busy hour. The aim is to assign just sufficient
resources to handle busy hour traffic, and no more (so that the minimum possible number of
frequencies are needed).

Fractional loading changes this planning model by assigning more bandwidth (frequencies) to
each base station than is strictly necessary, to handle busy hour traffic.

The fractional load of a system is then defined as the average percentage of frequency usage;
that is, the traffic/number of traffic channels that the assigned frequencies can hold.

When frequency hopping is used with a fractional load, even when the network is operating at
maximum traffic level, some frequencies will suffer from no interference at all. This is because
of channel coding and interleaving error correction algorithms that allow time slots on these
frequencies to be used to correct errors in time slots that have interference. The result is that
the threshold C/I value (the C/I value for the given FER or BER required for marginal quality) is
reduced, allowing tighter frequency reuse.

The effect of fractional loading on a frequency hopping system is illustrated in the following
figure. The figure shows a system hopping over four frequencies, but with only one call per cell.

Issue 1.4 – January 2001 Lucent Technologies – 21


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Figure 13: Fractional loading and frequency hopping

Note: In frequency hopping systems, the transceiver carrying the BCCH does not normally
operate in hopping mode (because the BCCH frequency must transmit continuously on the
downlink).

Fractional loading can be implemented by either of the following methods:

• Implementing an admission control procedure

• Installing fewer transceivers than allocated frequencies and using synthesiser frequency
hopping

Admission control procedures

If the number of transceivers were to equal the number of allocated frequencies, the network
would respond to overload conditions by reducing the quality of all calls, rather than blocking
calls. This could result in more dropped calls, the effect of which is worse for subscribers than a
blocked call.

Admission control procedures could potentially minimise dropped calls by allowing more
effective handling of local traffic peaks. As a large number of channels are temporarily available
in a sector (provided that the load in surrounding co-channel sectors is low), the admission
control procedures could use these.

However, suitable algorithms have not yet been found, and therefore fractional loading should
be used only in conjunction with synthesiser frequency hopping, by installing fewer transceivers
than allocated frequencies.

22 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Fewer transceivers than frequencies

In this case, the fractional load is often calculated as the number of transceivers divided by the
number of assigned frequencies. This is not the actual fractional load of the frequencies, as it
does not take into account the blocking of the system. However, since it is simple to calculate
and widely used, this definition of fractional load is used in this document, unless otherwise
stated.

Also, finer granularity of the levels of fractional load can be achieved by disabling some time
slots in the transceivers.

Impact on network planning

In high traffic areas such as large cities, the capacity of a cellular system is limited by the
system’s own interference caused by frequency reuse.

Most systems aim to satisfy as many customers as possible, therefore the system is planned
on the basis that only a given small proportion of calls at the cell edge (usually around 10%)
may suffer bad quality due to interference. With this “worst case” method, the capacity of a
system increases if the statistical spread of the C/I around its mean value is as small as
possible. This is illustrated in the following figure:

1 1
0.9 0.9
0.8 0.8
0.7 0.7 STANDARD
DEVIATION
0.6 0.6
12 dB 12 dB
0.5 0.5
7 dB 7 dB
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0 0
0 5 10 15 20 25 30 35 40 45 50 0 5 10 15 20 25 30 35 40 45 50
C/I [dB] C/I [dB]

Figure 14: Example of C/I distributions

The diagram on the left shows the C/I distribution for systems with an equal average C/I value
but different deviation.

The diagram on the right shows how a smaller deviation allows a lower average C/I value for
the same planning objectives. The interference diversity property of frequency hopping has
exactly this effect, it averages quality across the network and decreases the deviation. By
reducing the average C/I value in this way, network operators can plan for tighter frequency
reuse.

Issue 1.4 – January 2001 Lucent Technologies – 23


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Two factors optimise the averaging effect of interference diversity:

• High numbers of interference sources for frequency hopping to switch over

• Low correlation between the interference they cause (that is, interference variation)

The number of hopping frequencies governs the first factor. The higher the number the better.

The second factor depends partly on the locations of the interference sources. In the example in
Figure 8, the interferers are different mobile stations assigned to the same cell. In the same
scenario, but in the downlink, although the interfering communication is different in every time
slot, the interference source is always the same (the same base station; different frequencies).
Hence the correlation is high and the averaging effect is small. This is illustrated in the following
figure:

Figure 15: No interference diversity in the downlink

Variable interference planning solutions from Lucent Technologies are designed to counteract
this problem. These techniques maximise levels of variable interference in the network,
particularly in the downlink, in order to exploit fully the benefits of frequency hopping.

Variable interference planning techniques are described in the following chapters.

24 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Introduction to Variable
Interference Planning

3
3. Introduction to Variable Interference
Planning

Lucent offers two Variable Interference Planning (VIP) solutions for frequency hopping. This
chapter describes these two variable interference planning solutions:

• VIPone – based on variable reuse patterns

• VIPtwo – based on fractional loading

3.1. VIPone

VIPone is based on variable reuse patterns. Variable reuse patterns implement different reuse
patterns within the same cellular network. Typically, a loose reuse pattern such as 4/12, is used
for the transceiver that holds the BCCH control channel. A progressively tighter reuse is applied
to the second and third TCH transceivers, and so on.

One way to implement variable reuse patterns is to divide the allocated spectrum into sub-
bands, each band containing a different number of separately planned carriers. One or more
frequencies from each sub-band is allocated to each sector. For example, a 12 reuse for the
BCCH transceiver, and a 9 and 3 reuse for the second and third TCH transceiver respectively.
The result gives a total average reuse of 8 ((12+9+3)/3 = 8)4. This reuse pattern is illustrated in
the following figure:

4 This is an example. It is not achievable in a real network

Issue 1.4 – January 2001 Lucent Technologies – 25


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

BCCH TCH1 TCH2


12 9 3

... ...

Figure 16: Variable reuse pattern 12/9/3

Variable reuse patterns can also be accomplished by using automatic frequency planning tools
such as Generalised Radio Network Design (GRAND), by planning each of the transceivers in a
cell for progressively lower C/I threshold levels. In this way network irregularities are catered for,
a feature that cannot be implemented using regular reuse patterns.

VIPone uses frequency hopping in conjunction with variable reuse patterns in order to:

• Produce the necessary interference variation in the downlink

• Improve the existing interference variation in the uplink

Downlink

As described in the previous chapter, with regular reuse patterns there is no interference
variation in the downlink because the interfering source is always the same.

With variable reuse patterns the interfering base station is different for each time slot. Each
base station also belongs to a different “tier” of interferers; each tier corresponding to a different
reuse pattern. This produces interference diversity in the downlink, so that the averaging affect
of frequency hopping can work. The following figure illustrates downlink diversity.

26 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Figure 17: Interference diversity with variable reuse - downlink

Uplink

In the uplink, the effect of variable reuse patterns is that the interfering mobile stations are
assigned to different base stations belonging to different tiers. This results in a higher
decorrelation of the interference signals, and again, a better averaging effect with frequency
hopping than in a network with a regular reuse pattern

Figure 18: Interference diversity with variable reuse - uplink

Issue 1.4 – January 2001 Lucent Technologies – 27


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

VIPone properties

Since the number of hopping frequencies is always equal to the number of hopping
transceivers, VIPone can be implemented using baseband frequency hopping.

Additionally, because interference diversity is already achieved simply by the difference in


reuse, VIPone can be used with both cyclic and random hopping. The choice will generally
depend on the number of hopping frequencies. For small numbers of frequencies (such as two)
cyclic hopping should be used because it achieves better spectrum use.

Field tests in live networks show that an average frequency reuse factor of as low as 7.5 is
possible without impacting network quality. By applying power control and DTX in the downlink,
the average reuse can be decreased below 7.

Variable reuse patterns can also be used to handle unevenly distributed traffic, (that is, different
numbers of transceivers per cell). This is illustrated in the next section.

Another potential benefit of VIPone is to free up frequencies for the initial deployment of a
microcell layer.

VIPone examples

Scenario 1: Unevenly distributed traffic

In the previous 12/9/3 example, the operator might not initially need a third transceiver in all
cells. This means the effective reuse on the third sub-band will be higher than 3, and there will
be less interference in the network. But as capacity need increases, the third transceiver can be
installed in more cells, providing a progressively tighter average reuse without the need to
recalculate the frequency plan.

Scenario 2: VIPone plan in a real network

Another illustration is a 12/8/6/4 frequency plan, requiring 30 carriers (already in use in


commercial GSM networks). This allows an operator to assign up to 4 transceivers per cell,
roughly double the capacity of a standard 4/12 reuse pattern.

Scenario 3: Freeing up frequencies for the microcell layer

In another network, an initial reuse factor of 16.9 requiring 40 carriers, was tightened to a
14/10/6/2 configuration (average reuse of 12.87 and 32 carriers) and even to a 12/10/4/2
configuration (reuse 11.26 and 28 carriers). There was no change in the number of dropped
calls. Some degradation of perceived speech quality occurred in the second case, but this was
identified as a result of interference in the tighter BCCH band.

Note: All reuses quoted were achieved in capacity-limited networks, with an existing cell layout
optimised for capacity (almost homogeneous antenna height, orientation, and location).

28 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

3.2. VIPtwo

The capacity of a GSM network is generally limited by one of the following:

• Number of traffic channels (hard blocking)

• Interference from neighbour cells (soft blocking)

From a hard blocking viewpoint, small reuse factors give better performance than higher factors
(“trunking efficiency”). However, small reuse factors are limited by soft blocking (interference)
and cannot accept more than a given amount of traffic. This means that they would need to be
planned with a certain degree of fractional loading. As a result, the maximum capacity will lie
somewhere between a high and a low reuse factor.

Traditionally, networks have been planned to be limited by hard blocking. That is, the frequency
reuse has been set high so that only very few calls suffer from bad interference conditions. The
maximum capacity of the system is defined by the hard blocking limit set by the restricted
number of frequencies.

However, particularly with interference diversity, this is not the optimum way.

To illustrate, various reuse schemes with frequency hopping in the traffic carriers have been
simulated (COST 231). Their maximum capacity has been identified as the minimum of the
capacity that hard and soft blocking allow. The soft blocking limit was set so that less than 10%
of the calls were subject to an average C/I lower than 9 dB (as specified in GSM
Recommendation 05.05).

For an operator with 36 TCH frequencies (9.8 MHz), ideal power control, and DTX with a voice
activity factor of 50%, the maximum capacity per site was obtained for a sectorised base station
and a frequency reuse factor of 1/3 with a real fractional load of 30% (no. of transceivers/ no.
frequencies ≈ 38% - see page 21).

These results were obtained using a regular site lay out and homogeneous propagation
conditions. In practice, “off-grid” placements, irregular propagation conditions, and uneven
traffic loads will produce additional interference variation, allowing a higher fractional load.
Fractional loads of up to 50% (with DTX and power control) have been used in real networks
without a noticeable decrease in quality.

VIPtwo is based on these ideas. It consists of using very tight reuse patterns, typically 1/3 or 1/1,
and fractional loading to introduce the required interference variation.

VIPtwo properties

VIPtwo uses fractional loading and therefore requires synthesiser frequency hopping.

As previously mentioned, 1/3 reuses allow fractional loads of up to 50%. Field trials show that
fractional loads of 15%-20% are possible with a 1/1 reuse.

Issue 1.4 – January 2001 Lucent Technologies – 29


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Like VIPone, it can cater for unevenly distributed traffic, simply by setting different fractional loads
to different base stations. It can also be used to free up frequencies to be used in a
microcellular layer.

VIPtwo also provides the following benefits:

• It eases the planning effort, since the whole pool of frequencies is assigned to each site or
each cell, and only the control frequencies detailed careful planning

• The network can be planned with VIPtwo from the beginning, even if the number of
transceivers required per cell is initially low. Further transceivers can be added as
necessary without modifying the frequency plan. Quality will not be compromised as one of
the advantages of frequency hopping is its ability to smoothly trade-off quality and capacity
depending on the traffic load

VIPtwo examples

Scenario 1: 1/3 reuse

An operator with 7.5 MHz (37 frequencies) could achieve 3-3-3 configurations by using a typical
4/12 reuse factor, supporting 14.9 Erlangs/cell (2% blocking). If VIPtwo were used, 12
frequencies would be assigned to the transceivers containing the control channels, using a 4/12
reuse. The rest of the frequencies could be planned with a 1/3 reuse. This means 8 hopping
frequencies per sector and a spare frequency for optimisation. 4-4-4 configurations can then be
achieved with 1 control transceiver and 3 traffic transceivers per sector. The result is 30 traffic
channels and a traffic level per cell of 21.9 Erlangs: a 47% capacity increase. The fractional
load is 3/8= 37.5%.

Scenario 2: Greater capacity

Increasing the fractional load to 50% increases the number of traffic transceivers to four. This
means 37 traffic channels (2 dedicated control channels are now used) and a traffic level per
cell of 28.3 Erlangs: a 90% capacity increase.

Scenario 3: 1/1 reuse

With lower fractional load, 1/1 reuses are also possible. In the example, this would involve
hopping over 24 frequencies. Four traffic transceivers would therefore be allowed in all cells,
and up to five in selected cells.

Scenario 4: Network irregularities


The configuration in scenario 1, with a 1/3 reuse at a 2-sector site (typically used for highway
coverage) means that each sector can be assigned 8 hopping frequencies, taken from the total
pool of 24. This means that some of the frequencies on the pool will not be used. The
interference averaging capabilities of frequency hopping allows the system to exploit this
situation, by improving perceived quality levels.

30 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

3.3. VIPone and VIPtwo compared

The choice between VIPone and VIPtwo will be governed by spectrum allocation and the radio
equipment in use.

For operators with a large spectrum allocation and a high number of transceivers per cell, VIPone
is the natural choice. It allows the use of lower loss filter combiners at the base stations, thereby
preserving the coverage footprint while taking advantage of the system gain provided by
frequency hopping. VIPone is typically used by operators either with filter combiners in
widespread use, or with a more generous spectrum allocation and a need for high configuration
cells.

In contrast, for operators with a small amount of spectrum, or with a base station infrastructure
already equipped with hybrid combiners, VIPtwo is potentially a more flexible approach, because
of the ease of frequency planning.

Combined plans

It is also possible to combine VIPone and VIPtwo, by using fractional loading in conjunction with
variable reuses. Typically this might be used by operators with high spectrum allocation and
wideband antenna coupling equipment. This combined plan allows the operator to take into
account future growth. Reuse strategy can be set tight from the beginning, to cater for future
traffic increases. Transceivers can be added to the sites as needed, without changing the
frequency plan. Initially, the low fractional load (transceivers/assigned frequencies) ensures high
quality, which will then be traded off for capacity, as the need arises.

Issue 1.4 – January 2001 Lucent Technologies – 31


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

This page is intentionally left blank

32 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Configuring Frequency
Hopping

4
4. Configuring Frequency Hopping

This chapter details the hardware and software configurations required to support frequency
hopping in a GSM network, and the parameters required to activate it. The chapter also
discusses DTX and dynamic power control deployment.

4.1. Base station hardware

Base model

The base model contains the following components:

• RBS900 family (900 band) supports baseband hopping only.

• BTS-2000 family (900 and 1800 bands) support both synthesiser and baseband hopping:

- RFUs. These vary depending on model type, as denoted by the second letter of the
equipment code. B denotes support for baseband hopping; S denotes support for both
baseband and synthesiser hopping. Filter configurations (with TXFU09/TXFU18 filter
combiners) support baseband frequency hopping only.

- SRFUs support both synthesiser and baseband hopping.

• BTS-2000/2C supports synthesiser hopping only, and only in its second transceiver.

• FLEXENT family (900 and 1800 band) supports both synthesiser hopping and baseband
hopping.

Issue 1.4 – January 2001 Lucent Technologies – 33


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Antenna coupling equipment

• BTS 2000 family:

- Filter configurations (with TXFU09/TXFU18 filter combiners) support baseband


frequency hopping only.

- Hybrid and diplexer configurations (which use TXHU09/THDU18 hybrid combiners,


TXDU09/TXDU18 diplexers, or both) support baseband and synthesiser hopping.

- Due to hybrid combiner losses (each hybrid layer introduces a 3 dB loss into the overall
combining loss), hybrid configurations of up to 4 transceivers are available with two
antennas, and 6 transceivers with three antennas.

• FLEXENT family:

- Filter configurations (with TX4F09/TX4F18 filter combiners) support baseband


frequency hopping only.

- Hybrid configurations (which use TX2H09/TX2H18/TX4H09/TX4H18 and TX4H19


hybrid combiners) support baseband and synthesiser hopping.
- Due to the hybrid combiner losses (each hybrid layer introduces a 3 dB loss into the
overall combining loss), hybrid configurations of up to 4 transceivers are available with
two antennas, and 6 transceivers with two (with the optional hybrid combiner). When
the optional hybrid combiner is used, there is an addition 3.5dB loss to the TRX that is
connected to it.

34 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

4.2. Software release support

GSM release 7.0, 8.0 and 9.0 supports baseband and synthesiser hopping.

4.3. Configuration

Frequency hopping configuration in Lucent equipment is based on the concept of a Frequency


Hopping System (FHS). An FHS consists of:

• A set of hopping frequencies (from the pool of frequencies that are available at the cell)

• An HSN (Hopping Sequence Number)

The HSN is used to generate the hopping sequence in which the set of allocated hopping
frequencies is used. Allowable values are 0 to 63. Value 0 generates a cyclic hopping
sequence; all other values generate a pseudo-random sequence.

Each channel (defined as a transceiver and time slot pair) must have an associated FHS that
determines the frequencies the channel hops on and the hopping sequence.

Additionally, the MAIO (Mobile Allocation Index Offset) is automatically generated by the
system to prevent Um interface collision (channels using the same frequency at the same time)
between channels belonging to the same cell. The value can be 0 to N-1 where N is the number
of hopping frequencies.

FHS configuration rules

To avoid adjacent channel interference within a cell, the frequencies in an FHS should generally
obey a minimum co-site spacing rule, there should therefore be a separation of 2 or 3 GSM
carriers between them. The configuration must also comply with the following rules:

• The maximum number of frequencies in an FHS is 8

This means that a channel can hop on a maximum of 8 frequencies. In a future release,
hopping on up to 18 frequencies will be allowed in the case of synthesiser hopping.

• The maximum number of FHSs using the same frequency is 2 in BTS-2000 and 1 in RBS-
900

For the RBS-900 this means that either the transceiver holding the BCCH must be left to
non-hopping, or time slots 0 (and 2, 4, and 6 if additional CCCH channels are present) of all
transceivers must be non-hopping.

• The maximum number of FHS in a BTS is 8

• The maximum number of FHS in a BSS is 48

Issue 1.4 – January 2001 Lucent Technologies – 35


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Example 1: BTS with 4 RTs, 1 additional CCCH, baseband hopping

BCCH frequency non hopping:

CHN0 CHN1 CHN2 CHN3 CHN4 CHN5 CHN6 CHN7

RT0 BCCH -- CCCH -- -- -- -- --

RT1 FH1 FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT2 FH1 FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT3 FH1 FH1 FH1 FH1 FH1 FH1 FH1 FH1

FH0 no FH0

FH1 RT1,RT2, RT3

BCCH frequency hopping (RBS-900):

CHN0 CHN1 CHN2 CHN3 CHN4 CHN5 CHN6 CHN7

RT0 BCCH FH1 CCCH FH1 FH1 FH1 FH1 FH1

RT1 -- FH1 -- FH1 FH1 FH1 FH1 FH1

RT2 -- FH1 -- FH1 FH1 FH1 FH1 FH1

RT3 -- FH1 -- FH1 FH1 FH1 FH1 FH1

FH0 no FH0

FH1 RT0, RT1, RT2, RT3

BCCH frequency hopping (BTS-2000/FLEXENT BTS):

CHN0 CHN1 CHN2 CHN3 CHN4 CHN5 CHN6 CHN7

RT0 BCCH FH1 CCCH FH1 FH1 FH1 FH1 FH1

RT1 FH0 FH1 FH0 FH1 FH1 FH1 FH1 FH1

RT2 FH0 FH1 FH0 FH1 FH1 FH1 FH1 FH1

RT3 FH0 FH1 FH0 FH1 FH1 FH1 FH1 FH1

FH0 RT1, RT2, RT3

FH1 RT0, RT1, RT2, RT3

36 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

• The maximum number of physical channels using the same FHS is 42

In baseband hopping this restriction limits the maximum number of hopping frequencies
even further.

Example 2: BTS with 6 RTs, baseband hopping

BCCH frequency non hopping:

CHN0 CHN1 CHN2 CHN3 CHN4 CHN5 CHN6 CHN7

RT0 BCCH -- -- -- -- -- -- --

RT1 FH1 FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT2 FH1 FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT3 FH1 FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT4 FH1 FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT5 FH1 FH1 FH1 FH1 FH1 FH1 FH1 FH1

FH0 no FH0

FH1 RT1, RT2, RT3, RT4, RT5

BCCH frequency hopping (RBS-900):

CHN0 CHN1 CHN2 CHN3 CHN4 CHN5 CHN6 CHN7

RT0 BCCH FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT1 -- FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT2 -- FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT3 -- FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT4 -- FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT5 -- FH1 FH1 FH1 FH1 FH1 FH1 FH1

FH0 no FH0

FH1 RT0, RT1, RT2, RT3, RT4, RT5

Issue 1.4 – January 2001 Lucent Technologies – 37


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

BCCH frequency hopping (BTS-2000/FLEXENT BTS):

CHN0 CHN1 CHN2 CHN3 CHN4 CHN5 CHN6 CHN7

RT0 BCCH FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT1 FH0 FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT2 FH0 FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT3 FH0 FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT4 FH0 FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT5 FH0 FH1 FH1 FH1 FH1 FH1 FH1 FH1

FH0 RT1, RT2, RT3, RT4, RT5

FH1 RT0, RT1, RT2, RT3, RT4, RT5

Example 3: BTS with 7 RTs, baseband hopping


BCCH frequency non hopping:

CHN0 CHN1 CHN2 CHN3 CHN4 CHN5 CHN6 CHN7

RT0 BCCH -- -- -- -- -- -- --

RT1 FH1 FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT2 FH1 FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT3 FH1 FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT4 FH2 FH2 FH2 FH2 FH2 FH2 FH2 FH2

RT5 FH2 FH2 FH2 FH2 FH2 FH2 FH2 FH2

RT6 FH2 FH2 FH2 FH2 FH2 FH2 FH2 FH2

FH0 no FH0

FH1 RT1, RT2, RT3

FH2 RT4, RT5, RT6

38 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

BCCH frequency hopping (RBS-900):

CHN0 CHN1 CHN2 CHN3 CHN4 CHN5 CHN6 CHN7

RT0 BCCH FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT1 -- FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT2 -- FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT3 -- FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT4 FH2 FH2 FH2 FH2 FH2 FH2 FH2 FH2

RT5 FH2 FH2 FH2 FH2 FH2 FH2 FH2 FH2

RT6 FH2 FH2 FH2 FH2 FH2 FH2 FH2 FH2

FH0 no FH0

FH1 RT1, RT2, RT3

FH2 RT4, RT5, RT6

BCCH frequency hopping (BTS-2000/FLEXENT BTS):

CHN0 CHN1 CHN2 CHN3 CHN4 CHN5 CHN6 CHN7

RT0 BCCH FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT1 FH0 FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT2 FH0 FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT3 FH0 FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT4 FH2 FH2 FH2 FH2 FH2 FH2 FH2 FH2

RT5 FH2 FH2 FH2 FH2 FH2 FH2 FH2 FH2

RT6 FH2 FH2 FH2 FH2 FH2 FH2 FH2 FH2

FH0 no FH0

FH1 RT1, RT2, RT3

FH2 RT4, RT5, RT6

Issue 1.4 – January 2001 Lucent Technologies – 39


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

• In Network Release 9, frequency hopping is not supported on PDCH/TCH.

Example 4: BTS with 4 RTs, 1 additional CCCH, 6 PDCH/TCH, baseband hopping

BCCH frequency non hopping:

CHN0 CHN1 CHN2 CHN3 CHN4 CHN5 CHN6 CHN7

RT0 BCCH PDCH CCCH PDCH PDCH PDCH PDCH PDCH


/TCH /TCH /TCH /TCH /TCH /TCH

RT1 FH1 FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT2 FH1 FH1 FH1 FH1 FH1 FH1 FH1 FH1

RT3 FH1 FH1 FH1 FH1 FH1 FH1 FH1 FH1

FH0 no FH0

FH1 RT1,RT2, RT3

Other limitations
• Intra-cell handover is disabled when frequency hopping is active on a channel. In the case
of mixed configurations (hopping and non-hopping channels) intra-cell handovers will take
place between the non-hopping channels and from non-hopping channels to hopping
channels, but not from hopping channels
• For dual band operation, frequency hopping is allowed only between frequencies belonging
to the same band

4.4. Feature activation and system parameters

Once the hopping configuration is defined, frequency hopping must be configured and activated
in the requisite network elements:

BTS hopping mode

In the RBS-900, the hopping mode (which is baseband only) is implicitly defined by the BTS-
HW configuration.

In the BTS-2000 family, which allows both types of hopping, the hopping mode (baseband or
synthesiser) is set via the RBT-2000 (Radio Base Station Tester) software in the IMW-20005.
The default value is baseband.

5 A notebook PC with dedicated software for BTS-2000 and BCF-2000 administration

40 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

In the FLEXENT family, the hopping mode can be set via the RBT-2000 software in the IMW-
2000. However, the FLEXENT BTS can auto-detect the hardware configuration of the coupling
equipment and set the appropriate hopping mode. Therefore no manual setting is required.

BSS feature enabling

Frequency hopping is a purchased option, and a factory access code is required to enable
frequency hopping in a BSS. The feature is then enabled in the BSS Local Configuration Data.

OMC parameter configuration

The following Operations and Maintenance Centre (OMC) objects must be created or modified
(internal parameter names are used, followed in brackets by the OMC GUI and AUI parameter
names respectively).

BTS: Each frequency in the FHS must be defined in the cellAllocation (CellAllocation,
CELLALLOC) attribute of the BTS object.

RT: For baseband hopping, an RT (Radio Terminal) object must be created for each frequency
to be used in an FHS. The frequency is defined by the attribute initialFrequency
(InitialFrequency, INITFREQ).

For synthesiser hopping there is no relationship between the initialFrequency and the
frequencies belonging to the FHS. However the initialFrequency of the back-up RT should be
set carefully to ensure the initial frequency of BCCH RT is transferred to the Backup RT if the
BCCH RT fails.

FH: A Frequency Hopping (FH) object must be created for each required hopping sequence.
The following attributes must be defined for each FH object:

• allocatedFrequencies (AllocatedFrequencies, ALLOCFREQ): frequencies belonging to the


hopping sequence must be defined here (entries must match the cellAllocation attribute of
the BTS object).

• HSN (SequenceNumber, HOPSEQNO): defines the HSN to be used by the hopping


sequence generator. Specify 0 for cyclic hopping; or a number in the range 1 through 63 for
random hopping.

CHN: Each channel (CHN) object must be defined as hopping or non-hopping via the
freqHoppRelationship (FHRelationship, FREQHOPREL) attribute. This specifies an associated
FH object (hopping channels) or is left empty (for non-hopping channels).

A CHN object with channelType (Channel Type) of CCCH cannot be defined as a hopping
channel.

An MAIO will be generated internally for the channel (according to GSM Recommendation
05.02). A maximum of 42 channels can be associated with the same FHS.

Issue 1.4 – January 2001 Lucent Technologies – 41


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Automatic parameter update in the OMC


All the necessary changes required to install frequency hopping in a selected BTS can be
automated using the site independent OMC script inst-fhs.r78. The script installs frequency
hopping systems in the selected BTS depending on:

• Number of RTs connected to the BTS


• Hopping type (baseband or synthesiser)
• Presence of CCCH(s)

Additionally, the Automatic Network Modification for Frequency Hopping feature (omc-cm093)
of OMC-2000 release 4.5 has automated the process of frequency replanning with frequency
hopping. This feature allows the frequency planner to provide the OMC operator with frequency
and frequency hopping plans in electronic format.

The OMC reads the plan, validates the data, and generates a set of AUI scripts that will update
the existing OMC data to match the new plan. The scripts may be executed immediately or
scheduled for later execution.

The procedure is as follows:

The OMC operator requests a frequency plan report in raw format, via the Configuration Report
Generator implemented in the OMC. This report can be exported to an off-line PC where the
frequency planner can modify the frequency plan. The OMC operator can then import the data
back to the OMC. Using the Receive Plan Option on the Expert AUI window, an AUI script is
generated containing the modifications required to change the OMC data to the data specified
in the file.

For more information, refer to the OMC-2000 System Operator’s Guide - OMC Release 4.5.

Feature activation

When the FHSs are activated, the reconfiguration process involves two steps:

1. The reallocation procedure provides the BTS with the necessary information, and instructs
it to reconfigure its hopping behaviour at a specified start time.

2. The frequency redefinition procedure triggers the call handling function to start the
frequency redefinition in the mobile stations (see Frequency redefinition procedure on page
10).

Note: Activation and deactivation of the FH takes time because as previously stated, it involves
a frequency redefinition procedure that takes up to 3.5 minutes per hopping system. However,
this procedure has little impact on established calls (see Frequency redefinition procedure on
page 10), that is, it causes no down time of RTs or base stations. Only modification of the RTs
Initial Frequency produces RT downtime, and that is regardless of whether or not the system
hops.

42 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

4.5. Fault management

Frequency hopping is automatically disabled in the following situations:

Baseband hopping

If the number of available hopping frequencies used by the FHS falls below a given threshold
due to severe RT faults, this threshold represents the percentage of frequencies in the cell that
must still be operational before FH is switched to non-hopping. This percentage is defined in the
LMB Enable/Disable CH options of the BSS Controller Equipment (BCE) or Base Station
Controller Frame (BCF) local configuration area.

As the BSS does not redefine the list of allocated frequencies when an RT fails, the threshold
should be set to 100% (the default) to avoid bad quality connections due to frequency loss.

Note: The Frequency redefinition procedure (Page 10) triggered with the deactivation of
frequency hopping, prevents all the calls in the base station from being dropped when an RT
fails. Only calls served by the affected RT are dropped (as happens in a non-hopping system).
The quality of the other calls will degrade for as long as the Frequency redefinition procedure
takes place. The level of degradation depends on the number of hopping frequencies: the
greater the number, the lower the degradation. This behaviour is however typical of any
baseband hopping system, irrespective of the vendor, and is due to the way baseband
frequency hopping is generated.

Issue 1.4 – January 2001 Lucent Technologies – 43


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

4.6. DTX

Uplink DTX

Uplink discontinuous transmission is set on a per-BTS basis. To do this, set the Uplink DTX
parameter in the BTS Detail View of the OMC GUI (AUI parameter DTX of the managed object
class BTS).

The parameter can have three values:

• May be used (0)

• Shall be used (1)

• Shall not be used (2)

Downlink DTX

Downlink DTX can be set independently for speech and for non-transparent. To do this the
corresponding parameters must be set in the BSS, the MSC, and the InterWorking Function
(IWF).

Speech

This feature is enabled or disabled on a per BTS basis, via the OMC. To do this, set the
Downlink DTX Speech parameter in the BTS Detail View of the OMC (the AUI DownlinkDtx
attribute of the managed object class BTS contains the boolean downlinkDtxSpeech). The
default setting is disabled (false).

In a Lucent MSC, the switch option Downlink DTX Mode in the WBOPM (Wireless Base Office
Parameters Miscellaneous) view, can be enabled and disabled in the corresponding windows of
the Recent Change and Verify (RC/V) program. This applies to all the BSS supported by one
MSC. The default setting is disabled.

DTX is permitted for the connection if DTX is requested by the MSC and enabled by the OMC.

Data

This feature is enabled or disabled on a per BTS basis via the OMC. To do this, set the
Downlink DTX Data parameter in the BTS Detail View of the OMC (the AUI DownlinkDtx
attribute of the managed object class BTS contains the boolean downlinkDtxData). The default
setting is disabled (false).

If this parameter is enabled, the BTS acts according to the DTX commands issued by the IWF
in the received RLP frames.

To enable DTX in the IWF, the IWF option DTX Mode is set by changing the value in the IWF-2
menu. The default mode is disabled.

44 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

4.7. Dynamic power control

Power control for communications through a given BTS, can be deployed independently in the
downlink and the uplink via the parameters EN_MS_PC (uplink) and EN_BS_PC (downlink) of
the POWER object associated with the BTS.

Before applying this control the following parameters of the POWER object must be set to their
proper values:

• Maximum transmit power values:

MS_TXPRWR_MAX: defines the maximum TX power an MS is permitted to use on a


dedicated control channel or a traffic channel within the serving cell.

• Averaging measurement parameters:

A_LEV_PC: defines the averaging window size for receive power level measurements.

A_QUAL_PC: defines the averaging window size for quality measurements.

W_QUAL_PC: defines the weighting factor to account for effects of DTX on power control
signal quality measurements during averaging process.

• Threshold levels:

L_RXLEV_UL_P, U_RXLEV_UL_P, L_RXQUAL_UL_P, U_RXQUAL_UL_P: defines the


uplink lower (L) and upper (U) RX_LEV and RX_QUAL thresholds

L_RXLEV_DL_P, U_RXLEV_DL_P, L_RXQUAL_DL_P, U_RXQUAL_DL_P: defines the


downlink lower (L) and upper (U) RX_LEV and RX_QUAL thresholds

• Power step sizes:

POW_INCR_STEP_SIZE, POW_RED_STEP_SIZE: defines the step sizes used when


increasing or decreasing the MS and BTS transmit power

• Timer values:

P_CON_ACK: defines the power control acknowledge time

P_CON_INTERVAL: defines the minimum interval between successive modifications of the


radio frequency power level

Issue 1.4 – January 2001 Lucent Technologies – 45


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

This page is intentionally left blank

46 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Variable Interference Planning


Deployment

5
5. Variable Interference Planning
Deployment

This section describes the Variable Interference Planning (VIP) deployment rationale.

5.1. Introduction

When to use VIP

There are two main reasons why an operator might implement VIP and frequency hopping:

• To improve quality in an area with interference problems

• To increase capacity in an already saturated area (in terms either of a need for more
transceivers to meet traffic loads, or a need to free up frequencies in the existing spectrum
for use in other layers)

Frequency hopping should not be used to try to improve poor quality in networks where the
underlying cause is poor coverage, network planning or tuning. In such cases, frequency
hopping can cause further deterioration in performance.

Issue 1.4 – January 2001 Lucent Technologies – 47


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Implementation strategy

The interference averaging effect of frequency hopping generally diminishes the number of
interference problems in the network. However, those problems that remain will be more difficult
to resolve. For this reason, we strongly recommend that the process of increasing capacity with
frequency hopping and VIP should be implemented in stages as follows:

1. Switch on frequency hopping.

2. Tighten the frequency reuse or the fractional load step by step, as and when new capacity
is needed.

3. When the capacity gains from frequency hopping have been exhausted (but not before),
implement DTX and power control.

Each step should be deployed in a small trial area first, with extensive data collection made at
each stage in order to assess accurately the impact of the new plan on the network. No two
networks behave the same when frequency hopping is switched on, so it is important that
detailed results data is collected for each network.

The conclusions drawn from the initial deployment can then be used in the overall deployment
to minimise the initial impact and the subsequent optimisation work. In particular they can be
used to optimise the frequency plan and the radio link control parameter setting.

5.2. Choosing the right plan

This section describes how to identify the appropriate VIP plan for the implementation area.

More than three transceivers per cell

Areas with typical configurations of more than three transceivers per cell can use either VIPone or
VIPtwo. The choice will depend on the type of antenna coupling equipment already in place, the
investment that the operator is prepared to make, and the operator’s requirements for flexibility
and future growth. Specific factors that might influence the final choice include:

• VIPtwo has the big advantage of eliminating the need for frequency planning of the traffic
carriers in a network. So a VIPtwo plan is very flexible when it comes to introducing new base
stations.

• The main disadvantage of VIPtwo is that it requires hybrid or diplexer antenna combiner
equipment, which might not be in place.

• For network areas with existing filter-type combining equipment, any decision to swap-out
existing equipment should be carefully considered. This is because the increase in insertion
losses can affect the performance of the network, particularly for in-building coverage.

48 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

• If for whatever reason the use of hybrid combiners is not considered feasible, then VIPone is
the appropriate plan.

Three or fewer transceivers per cell

Areas with typical configurations of three or fewer transceivers per cell can use VIPtwo only.

Large spectrum allocation

This is for areas where synthesiser frequency hopping is possible. However, a high number of
available frequencies means that 1/3 or 1/1 reuse patterns are not possible, a mixed
VIPone/VIPtwo plan is therefore recommended.

Microcells

In microcellular environments (where configurations are normally low) VIPtwo is the best option
for the non-BCCH transceivers. VIPtwo allows frequency reuse from the macro layer, with no
need to take into account interference from other micro cells, if the fractional load is low
enough. This means that capacity can be added to the micro layer with minimal impact on the
existing frequency plan.

Planning for future capacity

Implementing DTX and dynamic power control in the downlink can produce further capacity
gains. However, remember that they should be introduced in stages; not at the same time.

5.3. Planning the frequencies and the HSN

VIPone

In areas with an average number of transceivers per cell of more than three, a quality increase
in terms of interference can be expected just by switching on random frequency hopping over
the existing assigned frequencies.

The more irregular the existing frequency plan, the higher the levels of variable interference,
and the greater the improvement. However, the gains may not be noticeable in networks with
existing high quality levels.

For cells with only two transceivers, it is best to either enable cyclic hopping or leave the cell as
non-hopping. These cells will still benefit from the interference diversity caused by surrounding
interfering cells randomly hopping over the same frequency set.

If VIPone has been chosen primarily for capacity gains, the first stage in the design process is to
calculate the average reuse factor required to handle the proposed capacity increase. Even if
additional capacity is needed in only a few cells, the calculation must be done as if all cells in
the area under consideration were to be upgraded.

Issue 1.4 – January 2001 Lucent Technologies – 49


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

For example, in a network with 3-3-3 configurations, only certain sites will be upgraded to 4-4-4.
Calculations will be done as if all the sites would be upgraded to 4-4-4.

Once the average reuse factor is determined, a variable reuse plan should be devised that
spreads the reuse factors around the average value to as great an extent as possible (taking
into account the number of transceivers within the plan).

For example: for three transceivers and 24 frequencies (average reuse of 24/3=8) a 12/8/4 plan
would work better than a 12/6/6.

Reuse capability depends greatly on the reuses allowed by the network infrastructure.
Homogeneous networks (grid site locations, regular antenna orientation and height) can support
a reuse value of 12, while others may require values as high as 15 or even 18.

If the number of hopping frequencies is 2, the HSN should be set to 0. Otherwise, it should be
set to an integer in the range 1 to 63, ensuring that all values are evenly distributed across the
area.

VIPtwo

VIPtwo always requires frequency re-planning, whether it is implemented to achieve quality


improvement or capacity gains.

The first step is to decide on the reuse factor (1/1 or 1/3). If the number of available frequencies
is low and a 1/3 reuse would mean hopping over fewer than six frequencies, a 1/1 reuse should
be used.

In other cases the operator can choose between the two options, taking into account the
maximum number of hopping frequencies available.

The HSN should be set to a value between 1 and 63, ensuring that it is different for base
stations using the same set of hopping frequencies. In the case of a 1/1 reuse it is important to
set the HSN of base stations to different values. Cells that belong to the same base station
should use the same HSN but a different MAIO, or co-channel interference will be possible.

VIPone/VIPtwo

In mixed plans, the number of hopping frequencies should be set to the maximum of eight, and
the VIPone plan should be designed assuming eight transceivers per cell.

As in the case of VIPone, the HSN should be set to an integer in the range 1 to 63, ensuring that
all values are evenly distributed across the area.

Microcells

Each microcell that requires additional capacity must be allocated a set of eight frequencies
from the traffic transceivers in the macro layer. A propagation prediction tool can be used to
select the frequencies with the lowest probability of interference within the area covered by
each given microcell. Random frequency hopping can then be activated.

50 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

The fractional load will ensure that interference from the macro and micro layer does not have
an adverse impact on transmission quality.

The HSN should be set to an integer in the range 1 to 63, ensuring that all values are evenly
distributed across the microcell area.

BCCH planning

Unless the capacity of the network is already stretched to its limits, the BCCH transceiver is
best left to non-hopping. In any case it is generally better if it is planned separately, using
frequencies specifically set aside for the BCCH.

This approach has the following benefits:

• The high levels of interference generated by the BCCH transceiver downlink are limited to a
specific band. (As the BCCH transceiver must transmit continuously, even when there is no
information to transfer, dynamic power control and DTX cannot be applied to it).

• Gains from implementing dynamic power control and DTX elsewhere in the network are
maximised.

• Control channel behaviour is separated from the traffic load. This is required to ensure
successful cell selection, handover, locating, access, and paging activities. Base
Transceiver Station Identity Code (BSIC) decoding on the SCH is especially important for
handover performance (poor handover performance causes more dropped calls).

• Capacity in existing cells can be increased without having to replan the BCCH
Generally the BCCH transceiver will only be set to hopping in the case of a VIPone plan, where
the number of hopping frequencies would otherwise be less than three.

Issue 1.4 – January 2001 Lucent Technologies – 51


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

5.4. Collecting performance data

To assess the benefits of deploying a VIP plan into the network, performance data must be
collected before and after the deployment.

First collect performance data for the current network configuration and frequency. This has two
purposes:

• To provide a data source for optimisation and tuning purposes.

• To provide a performance benchmark for comparison of data collected under the new plan.

Collection equipment

To collect the optimal range of performance data, the following equipment is required:

• GSM drive test equipment:


− Test handsets
− Data collection kit, preferably with reverse path measurement capability
− Scanner
− Post processing/analysis tool
− Voice quality measurement equipment

• Performance management tool (such as the OMC-PMS)

• Abis link monitor and protocol analyser

• Coverage prediction and frequency planning tool

Performance data types

This section details the various types of data that ideally should be collected for performance
measurement purposes. Each data type can be categorised as one of the following:

• Global information

• Drive test information

Note: When frequency hopping is switched off, ideally the performance data should differentiate
between BCCH and non-BCCH transceivers, in cases where the BCCH is non-hopping. This
may not be possible with global information, but drive test information should allow it.

52 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Global information

This type of performance data is usually obtained via the OMC-PMS. The ideal collection
method is to collect the data on a per cell basis both for all cells within the deployment trial area
and for cells surrounding the trial area.

Global information includes both traffic-related and quality-related data:

Traffic-related data

As traffic load is a major factor in frequency hopping performance, traffic data should be
collected before and after frequency hopping is implemented. This enables accurate analysis
and comparison of subsequent quality measurement results.

Lucent recommends that the following traffic data is collected, at a minimum for the busy hour,
and ideally also on a daily and historical basis:

• Busy hour (the hour segment with the largest TCH traffic value)

• TCH seizure attempts

• TCH seizures

• TCH seizure blocks

• % TCH blocking

• TCH traffic in Erlangs

• Mean TCH holding time

• SDCCH seizures

• SDCCH seizure blocks

• % SDCCH blocking

• SDCCH traffic in Erlangs

• Mean SDCCH holding time

Quality-related data

Quality data is used to compare performance results before and after frequency hopping is
implemented. Accurate analysis of the before and after performance data requires the following
conditions for the data collection:

• Values should be per Erlang wherever possible.

Issue 1.4 – January 2001 Lucent Technologies – 53


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

• Traffic conditions before and after the implementation should be sufficiently similar to
ensure no significant variation in interference and GOS (Grade of Service) levels. The traffic
data described in the previous section should be used to ensure equivalent traffic.

Note: These conditions apply to both hopping and non-hopping cells, and to the cells
surrounding the deployment area.

Lucent recommends that the following quality data is collected, at a minimum for the busy hour,
and ideally also on a daily and historical basis:

Dropped calls

• TCH seizures dropped for radio reasons

• % dropped TCH

• TCH dropped calls/Erlang

• SDCCH seizures dropped for radio reasons

• % dropped SDCCH

• SDCCH dropped calls/Erlang

Handovers

• Total number of handover attempts

• Intracell handover attempts

• Intracell handover failures

• % intracell handover failures

• Intercell handover attempts

• Intercell handover failures

• % intercell handover failures

• Uplink quality handovers

• % uplink quality handovers

• Uplink level handovers

• % uplink level handovers

• Downlink quality handovers

54 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

• % downlink quality handovers

• Downlink level handovers

• % downlink level handovers

RXQUAL statistics

If possible, RXQUAL statistics should be obtained (this will require an Abis protocol analyser).
Measurements should be performed at least over the busy hour.

Ideally, all Abis links for base stations in both the deployment area and surrounding areas should
be monitored. However, the equipment may restrict the number of links that can be monitored.

Drive test information

Drive tests should be performed over the most significant routes, including the main traffic
routes and, if possible, routes with known or potential conflict problems. In-building walk tests
are also useful when assessing the impact of frequency hopping on in-building quality.

Ideally, the drive tests should be performed during the busy hour (both before and after
implementation).

If possible they should also be repeated a number of times to ensure no external events
influence the results.

The following data should be collected:

• BSIC, BCCH frequency and TCH frequency of the serving cell.

• BSIC, BCCH frequency and RXLEV of neighbouring cells.

• Downlink RXLEV and RXQUAL measurements.

• Downlink co-channel and adjacent channel C/I (if this measurement is not available in the
drive test equipment, a scanner can be used). In the case of frequency hopping, the C/I
should be obtained for each hopping frequency.

• FER on the downlink and, if possible, on the uplink.

• Voice quality on the downlink and uplink (in the uplink take care not to introduce external
sources of quality degeneration).

• Handover, power control, and dropped call events and their causes (this requires call
tracing capabilities in the Abis monitor).

Issue 1.4 – January 2001 Lucent Technologies – 55


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

5.5. Deployment results

This section describes the results that can be expected at each VIP implementation stage:

1. Activating frequency hopping

2. Tightening frequency reuse

3. Implementing DTX

4. Implementing dynamic power control

Activating frequency hopping

The expected results of activating frequency hopping are:

• Dropped calls and failed handovers will decrease. In a medium loaded network, frequency
hopping may reduce the number of dropped calls by about 20%.

• RXQUAL statistics will show an increase in the reported RXQUAL values (see RXQUAL
behaviour below). The increase is generally about one unit. Under normal circumstances
(no frequency hopping) this would imply serious degradation of transmission quality, but it is
not the case with frequency hopping.

• As a consequence of the increase in RXQUAL values, the percentage (and possibly


absolute numbers) of quality based handovers will increase (see Figure 20).

• The number of intra-cell handovers will be very small (mixed hopping and non-hopping
configurations) or 0 (only hopping configurations).

• FER and voice quality (as measured in the drive tests) will improve.

• FER/voice quality versus RXQUAL/carrier to interference ratio will show improvement but
higher deviation.

These improvements are expected to be higher in the uplink than in the downlink. The downlink
will be the capacity-limiting link, with better quality in the uplink than the downlink.

Localised areas with previously bad interference problems but good coverage, should show
significant quality improvement, particularly if the channel used belongs to a hopping
transceiver.

56 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Figure 20: Impact of frequency hopping on handover causes

Figure 21 shows a typical drive test output when using frequency hopping. It shows the
RXQUAL, FER and SQI (Speech Quality Indicator - the measure of the speech quality TEMS
equipment offer) measured by a TEMS piece of equipment before and after handing over
between a channel that does not use frequency hopping and a channel that uses frequency
hopping.

Figure 21: Output from TEMS when handing over between a non-hopping and a hopping
channel

Issue 1.4 – January 2001 Lucent Technologies – 57


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

It can be seen that in the channel with no frequency hopping (shown left of the first marker in
the figure) bad RXQUAL values translate into an increasing FER (in red) and a decreasing SQI
(in grey). In the channel with frequency hopping (shown right of the first marker in the figure)
even worse RXQUAL values than before translate into no FER and a very slight degradation of
the SQI.

Important: Some trial implementations of frequency hopping have reported performance


degradation in cells that hop over only two frequencies. Such cells should be monitored for this
effect. To do this, aggregate the performance results according to the number of hopping
transceivers in the cell and compare the performance with the results obtained when the cells
were non-hopping.

RXQUAL behaviour

The increase in reported RXQUAL values is caused by the following reason:

The RXQUAL parameter does not increase linearly with the error rate of unprotected bits.
Instead it increases with its logarithm (RXQUAL increases by one unit if the BER is doubled or
decreases by two units if the BER is divided by four).

The following table shows RXQUAL values obtained in a cell after frequency hopping was
activated over four transceivers:

RXQUALFH RXQUALTRXi Average Potential speech


i=1,n RXQUALTRXi quality
5 0, 2, 7, 1 2.50 good
5 5, 5, 5, 5 5.00 fair
4 0, 6, 0, 0 1.50 excellent
2 0, 0, 4, 1 1.25 excellent
1 3, 0, 0, 0 0.75 excellent
1 0, 0, 0, 2 0.50 excellent
0 0, 1, 0, 0 0.25 excellent

Table 3: Example RXQUAL values with frequency hopping

With frequency hopping active, the BER (for unprotected bits) for the different hopping
sequence frequencies are averaged and then mapped into an RXQUALFH value for the hopping
channel. This means the RXQUALFH value is not calculated as the arithmetical average of the
RXQUALTRXi values of the individual TRX transceivers in non-hopping mode (as illustrated in the
table above).

This logarithmic behaviour means that RXQUALFH ≥ average(RXQUALTRXi).

58 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

As the values in the previous table show, with frequency hopping there is no direct mapping or
correlation of actual speech quality to RXQUAL.

Tightening the reuse

The expected result of tightening the average reuse is that the performance of the system, in
terms of dropped calls, remains constant up to a certain point. At that point, which represents
the capacity limit of the system and the current configuration, performance begins to deteriorate
rapidly.

However, the performance deterioration may be due to interference just in the BCCH band. This
should be investigated first. If the deterioration is identified as being due to BCCH interference,
then careful optimisation of the allocation in this band may produce further capacity gains.

DTX implementation
When DTX is switched on, the number of dropped calls may increase. This is because, by the
nature of DTX, some channel slots may not be used for transmission. Measurements on these
slots will obviously report a low reception level, and corresponding bad quality.
To avoid this problem, GSM Recommendations specify the following requirements:

• At least 12 bursts (a Slow Access Control Channel [SACCH] superframe) must be sent
within each reporting period. These bursts mirror the systematic use of the SACCH (four
bursts constitute a coding block) plus eight bursts on the TCH itself. For speech, these
bursts contain silence description frames (SIDs).

• The BTS and the mobile station must report two distinct sets of measurements concerning
the connection:
− “full” measurements for all slots that may be used for transmission in the reporting
period.
− “sub” measurements for the mandatory sent bursts and blocks only.

• Both the BTS and the mobile station must report for each measurement period, whether or
not discontinuous transmission was used. This allows the processes using the
measurements (power control and handover) to discard the “full” measurements in cases
when discontinuous transmission was used.
Results based on “sub” measurements are less accurate due to the reduced number of input
values for the averaging process (reception level is averaged on 12 bursts instead of more than
100 bursts). This affects quality measurements in particular. Because they are based on
estimated error probabilities before channel decoding, they are more sensitive to the statistical
unreliability introduced by subset measuring.
Specifically in the case of frequency hopping, this unreliability causes an increase in reported
RXQUAL values with a corresponding increase in dropped call rates.

Issue 1.4 – January 2001 Lucent Technologies – 59


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Dynamic power control implementation

The RXQUAL behaviour described means that an increase in RXQUAL reported handovers
(intercell and intracell) can be expected.

5.6. Optimising performance

Quality-based handovers

The increase in reported RXQUAL values leads to an unwanted increase in the percentage of
quality-based handovers. The easiest way to avoid this effect is to increase the handover
quality thresholds by approximately the same amount as the increase in the average RXQUAL
value.

If the percentage of quality based handovers remains high, the RXQUAL averaging window
should be increased, since the effect is probably due to statistical nature of the measurements.

Quality-based power control

A similar solution can be used to counteract the effect of increased RXQUAL based power
control commands following power control implementation. That is, increase the power control
quality thresholds by approximately the same amount as the increase in the average RXQUAL
value.

Hopping over two frequencies

Performance may deteriorate in cells that hop over only two frequencies. If this happens,
frequency hopping should be switched off in the affected cells. If the number of such cells is
small, there should be an improvement in their performance compared with a non-hopping
network, even though they do not hop.

Discontinuous Transmission measurement accuracy

A weighting algorithm has been devised in Lucent equipment that overcomes the potential
measurement inaccuracies introduced by DiscontinuourDTX. Full measurements are given a
higher weight than the “sub measurements” (which are more likely to be inaccurate) in the
average RXQUAL value calculation used in the power control and handover processes.

Setting the averaging parameters in this manner will improve performance in systems that use
DTX.

60 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Other scenarios

The following situations may also require investigation and optimisation:

• The global number of dropped calls and failed handovers either does not reduce, or even
increases. There could be three reasons for this:
− Poor coverage conditions

It has been reported that frequency hopping can aggravate problems arising from poor
coverage (as yet the reasons are unclear).

This situation is indicated by unusually high percentages of mandatory handovers (good


coverage networks should show a majority of power budget handovers) both with and
without frequency hopping.
− Very poor quality in the network before frequency hopping was implemented

In this situation, the averaging effect of frequency hopping will degrade quality further.
The few good quality mobiles will decrease their quality in an attempt to improve the
bad quality mobiles. However the bad quality mobiles will remain bad.

This situation is indicated by existing high numbers of dropped calls and failed
handovers before frequency hopping is implemented.
− Strong interferers exist in the network

Depending on location, some base stations can produce much higher interference
levels than the others in the network. For example, this often happens with base
stations at a higher than average height.

With frequency hopping, this interference is spread across all channels. The best
indicator of an offending base station is a permanently high measured level of
interference when it is scanned in drive tests. To avoid this effect, such sites should be
treated separately in terms of frequency planning and, in extreme situations, taken out
of the deployment area.

• The number of dropped calls and failed handovers in a particular cell either does not
reduce, or even increases. The possible reasons are:
− Poor coverage conditions

This situation is indicated by unusually high percentages of mandatory handovers in the


cell, both with and without frequency hopping.
− Locally high interference conditions

The cell may suffer localised interference from a very strong interferer. The propagation
prediction tool can be used to pinpoint the possible interferers. Then, drive tests can be
used to scan the BCCH frequency of these base stations to determine whether received
levels from them are high enough to cause interference.

Issue 1.4 – January 2001 Lucent Technologies – 61


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

If a strong interferer is found, the frequency plan should be modified to prevent the two
cells (the interferer and the cell suffering the interference) being used as co-channels.
The same average reuse or fractional load must be maintained. These changes may
involve rearranging of frequencies, which can be done either manually or with a
frequency planning tool.

Note: This situation often arises with BCCHs that are included in the hopping
sequences - because they transmit continuously, frequency hopping cannot take
advantage of traffic variations.

In VIPtwo implementations, if frequency rearrangement is not possible, the frequencies suffering


high interference should be taken out of the hopping sequences in the affected cell, even if this
reduces the number of hopping frequencies available.

62 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Worked Examples

6
6. Worked Examples

This chapter provides examples of VIP implementations. Each example describes the current
network configuration, the objectives of the implementation, and the planning and design
requirements.

6.1 Scenario 1

Existing configuration

• Operator working in the 1800 band

• Wide-band combiner equipment

• 48 frequencies allocated

• Network still growing with an irregular network layout that is mainly coverage ridden

• To maintain good quality in the existing network, current BCCH planning requires a reuse of
7/21

• Configurations are mostly 2-2-2

• Microcells in use for “cold-spot” coverage with plans to develop them into a continuous
microcell layer

Issue 1.4 – January 2001 Lucent Technologies – 63


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Objectives

• To improve quality in the network, especially in localised areas where propagation


conditions cause high interference levels.

• To generate a frequency plan that is able to cope with the rapid pace of change and growth
in the network.

VIP plan choice

• Since the existing combiner equipment is already wide-band, VIPtwo is the easiest and most
flexible solution to implement.

• Because the network is still growing, it is recommended that the macrocell base station
BCCHs are planned on a separate sub-band. This ensures that future capacity expansion in
existing base stations, or addition of new microcells, will not require modifications to the
BCCH frequency plan.

Planning the frequencies

• The frequency band will be divided into three subsets of 21, 18, and 9 frequencies. The first
sub-band will be used for macro BCCH planning, the second for macro TCH, and the third
for micro BCCH.

• The BCCH in the macro layer will be planned using a 7/21 reuse that is already known to
give adequate performance in the current network conditions.

• The BCCH in the micro layer, once continuous coverage is achieved, will be planned using
a 9 reuse, which is known to be adequate in a micro-cellular environment.

• The additional TCH transceiver in the existing 2-2-2 configurations will be planned using a
1/3 reuse. That is, assigning 18/3=6 frequencies per sector and switching on synthesiser
frequency hopping. (Reuse of 1/1 is not possible in current releases because it would imply
hopping over 18 frequencies).

• If additional capacity is subsequently required in the macro layer, it will be necessary to


upgrade base stations to 3-3-3 configurations. The frequency plan will not need to be
changed since a 1/3 reuse can easily accept fractional loads of 2/6=33% without noticeable
impact on the quality of the network

• If even more capacity is required in the macro-layer, the frequency plan will need to be
changed because a fractional load of 50% is too close to the maximum limit beyond which
network quality may degrade. However:

Ideally, by this time network growth in terms of base stations will have stabilised and the
network layout will have been rationalised into a more homogeneous layout (for example,
grid locations, similar antenna height and orientation)

64 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

If this is the case, the BCCH will allow a much tighter reuse: probably around 15. The spectrum
allocation can then be split into three bands of 15, 24, and 9. The TCH transceivers will be
assigned 24/3=8 hopping frequencies, which can accept a load of 3/8=37.5%. 4-4-4
configurations would therefore be possible

If the eventual network layout is insufficiently homogeneous, this must be corrected; no


additional capacity gains will be possible until this is done

Mapping the frequency plan to OMC settings

As stated earlier, each cell will be assigned one BCCH frequency fBCCH and 6 hopping
frequencies, fFH1, …, fFH6.

To do this:

1. Add frequencies fFH1, …, fFH6 to the cellAllocation attribute of the BTS object that relates to
the cell.

2. Set the initialFrequency attribute of the RT (Radio Terminal) object that will hold the BCCH
channel to fBCCH.

3. As the eight CHN (Channel) objects belonging to this RT are non-hopping, the
freqHoppRelationship (FH ID) attribute is left empty.

4. Set the initial frequency of the second RT (and the third RT if using 3-3-3 configurations) to
one of the hopping frequencies.

5. Create an FH (Frequency Hopping) object with attribute allocatedFrequencies equal to fFH1,


…, fFH6.
Set the attribute sequenceNumber to an integer in the range 1 through 63. Make sure this
value is different from the values used by FH objects of surrounding cells that have been
assigned the same set of hopping frequencies.
The ID attribute can be set to any value in the range 0 through 7.

6. Set the eight CHN objects belonging to the second RT (and the third RT if using 3-3-3-
configurations) to hopping. To do this, set the freqHoppRelationship (FH ID) attribute to
match the ID of the FH object created in the previous step.

Note: If the frequency plan is later rearranged and additional capacity introduced to support 4-
4-4 configurations, the OMC process will be the same, except that eight hopping frequencies
rather than six are added to the cell allocation, and a new FH object is created with those 8
frequencies.

Issue 1.4 – January 2001 Lucent Technologies – 65


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

6.2 Scenario 2

Existing configuration

• Operator working in the 900 band

• Existing combiner equipment is all filter type

• 40 frequencies allocated

• Network already stabilised

• Average of 2.48 transceivers per cell; actual configurations vary between one and four
transceivers per cell

• Average reuse of 16.13. It is not possible to add another carrier in the area with
conventional frequency planning

• Microcellular layer will be added for capacity increase

Objectives

• To free enough frequencies to be able to plan the BCCH of the microcell layer, without any
additional investment.

VIP plan choice

• Since the existing combiner equipment is filter type and the operator is not willing to invest
in swapping combiners, VIPone is the choice for the initial solution.

• Since the number of transceivers per cell is relatively low, the BCCH transceiver will be
included in hopping sequences.

Planning the frequencies

• It is envisaged that eight frequencies will be sufficient for the microcell layer. The new
macrocell plan will therefore use only 32 frequencies.

• Since the maximum configuration is 4-4-4, the planning assumes that all configurations are
4-4-4.

• Average reuse: 32/4=8.

• Frequency plan obtained by spreading around 8: 14/10/6/2.

• Actual average reuse: 32/2.48=12.9.

• Cells with only one transceiver will be left to non-hopping.

66 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

• Cells with only two transceivers will be hopping, but using cyclic hopping sequences.

• Cells with more than two transceivers will need two FHS. One FHS will be hopping on the
control channel TS of non-BCCH transceivers. The other FHS will be hopping on all traffic
channels of all transceivers.

Mapping the frequency plan to OMC settings

Each cell will be assigned n frequencies f1, …, fn where n is the number of transceivers in that
cell (in this scenario a value between 1 and 4).

In this scenario, f1 belongs to the sub-band of 14 frequencies, f2 belongs to the sub-band of 10


frequencies, f3 belongs to the sub-band of 6 frequencies and f4 to the sub-band of 2 frequencies.

If n>1, the following steps are required:

1. Add frequencies f1, …, fn to the cellAllocation attribute of the BTS object that relates to the
cell.

2. Set the initial frequency of each of the four RTs to fI where i is the number of the RT.

If the base station is of type RBS-900 or if n=2:

3. Set all CHN objects that have a CHN object ID attribute of 0 (or 2, 4, or 6 if CCCH channels
are present in these time slots of the BCCH transceiver) to non-hopping. To do this, leave
the freqHoppRelationship (FH ID) attribute blank. These channels belong to air interface
time slot 0 (or 2, 4, or 6 if CCCH channels are present in these time slots of the BCCH
transceiver).

4. Create an FH object with attribute allocatedFrequencies set to f1, … fn.

The attribute sequenceNumber should be set to 0 if n=2.


Otherwise set it to an integer in the range 1 through 63, ensuring that all the values are
used evenly across the area.

Set the ID attribute to a value in the range 0 through 7.

5. The CHN objects not included in step 3 should be defined as hopping. To do this, set the
freqHoppRelationship (FH ID) attribute to the ID of the FH object created in the previous
step.

Issue 1.4 – January 2001 Lucent Technologies – 67


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

If base station is of type BTS-2000 (RBS-918) and n>2:

3. Create an FH object with attribute allocatedFrequencies set to f2, … fn. (Create the first
FHS)

The attribute sequenceNumber should be set to 0 if n=3. Otherwise set it to an integer in


the range 1 through 63, ensuring that all the values are used evenly across the area.

Set the ID attribute to a value in the range 0 through 7.

4. Set all CHN objects in the non-BCCH transceivers that have a CHN object ID attribute of 0
(or 2, 4, or 6 if CCCH channels are present in these time slots of the BCCH transceivers) to
hopping. To do this, set the freqHoppRelationship (FH ID) attribute to the ID of the FH
object created in the previous step.

5. Create an FH object with attribute allocatedFrequencies set to f1, … fn. (create the second
FHS)

The attribute sequenceNumber should be set to 0 if n=2. Otherwise set it to an integer in


the range 1 through 63, ensuring that all the values are used evenly across the area.

Set the ID attribute to a value in the range 0 through 7, but different to the one of the FH
object created in step 3.

6. The CHN objects not included in step 3 should be defined as hopping. To do this, set the
freqHoppRelationship (FH ID) attribute to the ID of the FH object created in the previous
step.

If n=1, then the only RT object will be assigned an initial frequency belonging to the sub-band of
14 frequencies.

In the case of base stations of type RBS-900, because all channels in time slot 0 (or 2, 4, 6) are
non-hopping, they can potentially suffer from unacceptable interference levels. This is because
they are using frequencies with a very tight reuse, particularly those using frequencies from the
sub-bands of 6 and 2 frequencies. Performance on these channels should be closely monitored
and if quality is unacceptable, they should be shut down. This will imply a small loss in the
macro layer capacity. However, the increase in capacity provided by the frequencies that have
been freed for the micro-layer will more than compensate for this.

68 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

6.3 Scenario 3

Existing configuration

• Operator working in the 900 band

• 50 frequencies allocated

• Underlay microcell layer using BTS-2000/2C (CUBE) base stations. Continuous coverage
and 1 transceiver per micro

• Frequencies divided in three subsets: 18 for the macro BCCH, 19 for the macro TCHs and
9 for the micro BCCH

Objectives

• To add one transceiver to all cells in the micro layer

VIP plan choice

• As CUBEs support synthesiser hopping, VIPtwo is the most appropriate plan

Planning the frequencies

• In order to minimise disruption to the existing frequency plan, the frequencies belonging to
the macro TCH sub-band will be reused

• A set of eight frequencies will be chosen for each microcell. Frequencies should be selected
from the TCH transceivers of the macrocells that cause the least interference to the
microcell

This information can be obtained by using a scanner and doing a drive-test of the area (if it
is small enough), or with a propagation prediction tool. The GRAND tool allows the
probability of interference matrices to be calculated between the macrocells and the
microcells. For each microcell, the suggested frequencies will belong to the macrocells with
the lowest probability of interfering with that microcell

Issue 1.4 – January 2001 Lucent Technologies – 69


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Mapping the frequency plan to OMC settings

Each microcell will be assigned one BCCH frequency fBCCH and eight hopping frequencies, fFH1,
…, fFH8.

To do this:

1. Add frequencies fFH1, …, fFH8 to the cellAllocation attribute of the BTS object that relates to
the micro-cell.

2. Set the initialFrequency attribute of the RT that will hold the BCCH channel to fBCCH.

3. Set the eight CHN objects belonging to this RT as non-hopping (leave the
freqHoppRelationship (FH ID) attribute empty).

4. Set the initial frequency of the second RT to any of the hopping frequencies.

5. Create a FH object and set attribute allocatedFrequencies equal to fFH1, …, fFH8.

Set attribute sequenceNumber to an integer in the range 1 through 63. Make sure as far as
possible that all values are used and that they are spread evenly across the whole
microcell layer.

Set the ID attribute to any value in the range 0 through 7.

6. Set each of the eight CHN objects belonging to the second RT as hopping. To do this, set
the freqHoppRelationship (FH ID) attribute to match the ID of the FH object created in the
previous step.

70 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

6.4 Scenario 4

Existing configuration

• Operator working in the 900 band

• Network already stabilised. With maximised cell splitting, regular layout, and reasonably low
antennas

• 20 new frequencies acquired in the 1800 band

• Collocated 900 and 1800 base stations

• Wide-band combiners available in the 1800 band equipment

Objectives

• To maximise capacity in the small 1800 band

VIP plan choice

• Since the existing combiner equipment is already wide-band, VIPtwo is the most appropriate
solution

Planning the frequencies

• The regular network layout and low antenna heights mean that the network can support
4/12 reuses on the BCCH. Hence the 1800 band spectrum allocation will be divided in two
bands, one of 12 frequencies to plan the BCCH; and one of 8 frequencies for the extra
TCHs

• A 1/1 reuse is most appropriate given the small number of frequencies available for the
TCHs

• 2-2-2 configurations imply a fractional load of 12.5% (1/8). This is below the 15-20%
threshold for 1/1 patterns

• Widespread 3-3-3 configurations are unlikely to be possible, even with DTX and dynamic
power control switched on (2/8 gives a fractional load of 25%). However, subsequent extra
capacity can be added to selective locations without significant impact on quality, thanks to
the averaging properties of frequency hopping. One alternative would be to add a third
transceiver in all locations, but to activate only some of the channels (up to a maximum of 3
or 4)

Issue 1.4 – January 2001 Lucent Technologies – 71


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Mapping the frequency plan to OMC settings

Each cell in the 1800 band will be assigned one BCCH frequency fBCCH and eight hopping
frequencies, fFH1, …, fFH8.

To do this:

1. Add frequencies fFH1, …, fFH8 to the cellAllocation attribute of the BTS object that relates to
the 1800 cell.

2. Set the initialFrequency attribute of the RT that will hold the BCCH channel to fBCCH.

3. Set each of the eight CHN objects belonging to this RT as non-hopping (leave the
freqHoppRelationship (FH ID) attribute empty).

4. Set the initial frequency of the second RT to any hopping frequency.

5. Create an FH object with attribute allocatedFrequencies equal to fFH1, …, fFH8.


Set attribute sequenceNumber to an integer in the range 1 through 63. Ensure so far as
possible that all values are used and that they are spread evenly across the whole 1800
layer. Set the ID attribute to any value in the range 0 through 7.
6. Set each of the eight CHN objects belonging to the second RT as hopping. To do this, set
the freqHoppRelationship (FH ID) attribute to match the ID of the FH object created in the
previous step.

72 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

6.5 Scenario 5

Existing configuration

• Operator working in the 1800 band

• Combiner equipment is wide-band

• 40 frequencies allocated

• Mature network that is already stabilised

• Average of 3.3 transceiver per cell

• Antenna height is quite low, so average reuse has been set to 12

Objectives

• To increase capacity in the network to allow four transceivers per cell

VIP plan choice

• Since the existing combiner equipment is wide-band, VIPtwo is the most appropriate solution

• However, leaving 12 frequencies for the BCCH means a 1/3 reuse would require hopping
over more than 9 frequencies, which is not possible with Lucent equipment. Accordingly, a
mixed VIPone/VIPtwo solution is chosen

Planning the frequencies

• 12 frequencies are set aside for the BCCH, since it has already been proven that the
network can support that reuse. This leaves 28 remaining frequencies

• The plan will assign 6 hopping frequencies, a high enough number to benefit from
frequency hopping

• Average reuse: 28/6=4.6

• Frequency plan obtained by spreading around 4.6: 9/6/6/3/3/1

• Actual reuse: 28/4=7

Issue 1.4 – January 2001 Lucent Technologies – 73


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Mapping the frequency plan to OMC settings

Each cell will be assigned one BCCH frequency fBCCH and 6 hopping frequencies, fFH1, …, fFH6.

To do this:

1. Add frequencies fFH1, …, fFH6 to the cellAllocation attribute of the BTS object that relates to
the cell.

2. Set the initialFrequency attribute of the RT that will hold the BCCH channel to fBCCH.

3. Set the eight CHN objects belonging to this RT as non-hopping (leave the
freqHoppRelationship (FH ID) attribute empty).

4. Set the initial frequency of the second, third, and fourth RT to a hopping frequency
belonging to one of the looser reuses.

5. Create a FH object and set the allocatedFrequencies attribute equal to fFH1, …, fFH6.

Set attribute sequenceNumber to an integer value between 1 and 63. Ensure so far as
possible that all values are used and that they are spread evenly.

Set the ID attribute to any value in the range 0 through 7.

6. Set each of the eight CHN objects belonging to the second, third, and fourth RT as hopping.
To do this, set the freqHoppRelationship (FH ID) attribute to match the ID of the FH object
created in the previous step.

74 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

Acronyms

7
7. List of Acronyms

The following acronyms are used in this document:

AUI ASCII User Interface

BCCH Broadcast Control Channel

BCE BSS Controller Equipment

BCF Base Station Controller Frame

BER Bit Error Rate

BSS Base Station Subsystem

BSIC Base Transceiver Station Identity Code

BTS Base Transceiver Station

CCCH Common Control Channel

DRCC Digital Radio Codec and Control

DTX Discontinuous Transmission

FCCH Frequency Correction Channel

FER Frame Erasure Rate

GUI Graphical User Interface

GPRS General Packet Radio Service

GSM Global System for Mobile Communications

Issue 1.4 – January 2001 Lucent Technologies – 75


Proprietary
See Notice on first page
GSM Frequency Hopping and VIP Engineering Guideline

HSN Hopping Sequence Number

MAIO Mobile Allocation Index Offset

MSC Mobile Switching Centre

OMC Operations and Maintenance Centre

PAGCH Packet Access Grant Channel

PDCH Packet Data Channel

PDTCH Packet Data Traffic Channel

PWRC Power Control Indicator

RACH Random Access Channel

RFU Radio Frequency Unit

RXQUAL Received Signal Quality

SACCH Slow Associated Control Channel

SCH Synchronisation Channel

SDCCH Standalone Dedicated Control Channel

SID Silence Information Descriptor

TCH Traffic Channel

TDMA Time Division Multiple Access

TRX Transceiver

VIP Variable Interface Planning

76 Lucent Technologies – Issue 1.4 – January 2001


Proprietary
See Notice on first page
Comments Form

Comments Form

Document number: 401-380-365 Issue number: 1.4 Date: January 2001

Title: GSM Frequency Hopping and VIP Engineering Guideline

Lucent Technologies welcomes feedback on this customer Information Product. Your


comments are of great value in assisting us to improve our Information Products.

1. Please rate (tick the box) the effectiveness of this Information Product in the following areas:
Excellent Good Fair Poor Not Applicable
Ease of use
Clarity
Completeness
Accuracy
Organisation
Appearance
Examples
Illustrations
Overall Satisfaction

2. Please place a tick against any improvement that could be made to this Information Product:
Improve the overview/introduction Make it more concise/brief

Improve the table of contents Add more step-by-step procedures/tutorials

Improve the organisation Add more troubleshooting information

Include more figures Make it less technical

Add more examples Add more detail

3. Please provide details for the suggested improvement in the box below:

Lucent Technologies –
Proprietary
See Notice on first page
Comments Form

4. What did you like most about this Information Product:

5. Any additional comments:

6. If we may contact you concerning your comments, please complete the following:

Date:

Name:

Company/Organisation:

Address:

Job Title:

Telephone No:

When you have completed this form, please fax it to:

The Technical Manager, CTIP-UK, Fax Number (+44) 1666 824515

Lucent Technologies –
Proprietary
See Notice on first page

Das könnte Ihnen auch gefallen