Physical Layer measurements in 3GPP LTE

Aalborg University Diploma thesis

Physical Layer measurements in 3GPP LTE

Author: Rasmus Birkelund Nielsen Mauritio B. G. M. Nielsen

Supervisors: Kim Højgaard-Hansen

February 3, 2012

ii

Institut for Elektroniske Systemer Elektronik & IT Fredrik Bajers Vej 7 9220 Aalborg Ø Telefon 99 40 86 00 http://es.aau.dk

Synopsis: Titel: Physical Layer measurements in 3GPP LTE Denne rapport beskriver throughput Tema: Kommunikationssystemer Projektperiode: 3. november 2011 – 3. februar 2012 Forfattere: Mauritio Birk Georg Musil Nielsen Rasmus Birkelund Nielsen Vejleder: Kim Højgaard-Hansen

Oplagsantal: 4 Sideantal: 97

målinger på udrullet 3GPP LTE netværk. Disse målinger skal benyttes til at undersøge hvorledes dæmpning og intermodulations forvrængning påvirker andre UEs i nærheden. For at undersøge disse aspekter er der foretaget en række live målinger på et i forvejen udrullet LTE netværk, hvor der i den første måling foretages en simpel undersøgelse ved at dæmpe downlink signalet vha. attenuatorer. Den anden måling blev foretaget ved at sammenkoble signalet modtaget på UE, med at støj-signal fra en SMIQ 06b Vector Signal Generator. Det kunne ses fra resultaterne at der tydeligvis var en effekt ved at udsætte en UE for disse støj niveau. Desværre, er de foretagede målinger ikke nok i sig selv, og der bør foretages yderligere målinger. Forslag til hvilke aspekter der bør fokuseres på, er angivet i perspektiveringen.

Bilagsantal: 1 praktikrapport Afsluttet: 3. februar 2012

Rapportens indhold er frit tilgængeligt, men offentliggørelse (med kildeangivelse) må kun ske efter aftale med forfatterne.

iv

Department of Electronic Systems Electrical Engineering Fredrik Bajers Vej 7 9220 Aalborg Ø Telefon 99 40 86 00 http://es.aau.dk

Synopsis: Title: Physical Layer measurements in 3GPP LTE This report describes throughput meaSubject: Communication Systems Project period: November 3rd, 2011 – February 3rd, 2012 Authors: Mauritio Birk Georg Musil Nielsen Rasmus Birkelund Nielsen Supervisors: Kim Højgaard-Hansen

Copies: 4 Page count: 97 Appendix: 1 internship report

surements performed on a deployed 3GPP LTE network. The measurements are to be utilised to examine how attenuation and intermodulation distortion affects other UEs in the vicinity. Inorder to examine these aspects, a number of live measurements were performed on a deployed LTE network, where in the first measurement, a simple examination is made by attenuating the downlink signla with attenuators. The second measurement was performed by combining the signal received on the UE with a noise signal generated from a SMIQ 06b Vector Signal Generator. It could be seen from the results that there was clearly an effect, by exposing the UE for these noise levels. Unfortunately, the performed measurements are not enough, in the sense that there should be proformed additional measurements. Some proposals as to which aspects should be investigated further, and are given in the Perspective chapter.

Completion of project: February 3rd, 2012

The contents of the report is freely available however, publication (with reference) may only happen per agreement with the author(s).

vi

Chapter 1

Preface This project has been made by Rasmus Birkelund Nielsen and Mauritio Birk Georg Musil Nielsen, as part of the diploma thesis in Electronic Engineering at Aalborg University. The overall theme of the project is based on “Communication Systems”, and was conducted over the period from November 3rd , 2011 to February 3rd , 2012. The supervisor for this project is Kim Højgaard-Hansen, ph.d student at Networking & Security, associated with School of Information and Communication Technology, at Aalborg University. Futhermore, the group would like to give a special credit the employees at Agilent Technologies, Aalborg for their help and guidance throughout the internship and on this project.

Rasmus Birkelund Nielsen

Mauritio Birk Georg Musil Nielsen

vii

viii

Contents 1 Preface

vii

2 Introduction

1

3 Project Goal

3

I

4

Analysis

4 Long Term Evolution System Overview 4.1 Network architecture . . . . . . . . . . . . . . . . . . . . . . . 4.2 Protocol architecture . . . . . . . . . . . . . . . . . . . . . . . 5 LTE Physical Layer 5.1 Introduction To The Physical Layer . . . . . . . . 5.2 Architectural Overview . . . . . . . . . . . . . . . 5.2.1 Frame And Slot Structure . . . . . . . . . . 5.3 Modulation Scheme and Coding . . . . . . . . . . . 5.3.1 Adaptive Modulation and Coding (AMC) . 5.3.2 Downlink: Orthogonal Frequency Division Access (OFDMA) . . . . . . . . . . . . . . 5.3.3 Uplink: Single-Carrier Frequency Division Access (SC-FDMA) . . . . . . . . . . . . . 5.4 Summary . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Multiple . . . . . . Multiple . . . . . . . . . . . .

6 Concepts Of Interference 6.1 Noise . . . . . . . . . . . . . . . 6.1.1 Johnson-Nyquist noise . 6.1.2 Gaussian noise . . . . . 6.1.3 Signal-to-Noise ratio . . 6.2 Interference . . . . . . . . . . . 6.2.1 Co-Channel Interference 6.2.2 Intersymbol interference 6.3 Intermodulation . . . . . . . . .

. . . . . . . .

ix

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

5 5 6 10 10 10 10 16 16 17 18 20 21 21 21 22 22 23 23 23 23

6.4 6.5

II

In Denmark . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Measurements

25 26

27

7 Introduction 7.1 Throughput measurement . . . . . . . . . 7.2 Intermodulation Distortion measurement . 7.2.1 Directional coupler . . . . . . . . . 7.3 Case: Downlink blocks Uplink . . . . . . . 7.4 Key Performance Indicator (KPI) . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

28 28 28 28 29 30

8 Initial test 33 8.1 Setup and test procedure . . . . . . . . . . . . . . . . . . . . 33 8.2 Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 8.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 9 Intermodulation Distortion test 9.1 Introduction . . . . . . . . . . . 9.2 Setup and test procedure . . . 9.3 Observation . . . . . . . . . . . 9.3.1 13-01-2012 . . . . . . . . 9.3.2 14-01-2012 . . . . . . . . 9.3.3 15-01-2012 . . . . . . . . 9.4 Summary . . . . . . . . . . . .

III

Assesment

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

40 40 40 41 41 46 50 54

56

10 Final conclusion

57

11 Perspective

59

IV

63

Appendices

A MIMO

64

B Duplexing and Multiplxing

66

C Channel Access Methods 68 C.1 Basic Channel Access Methods . . . . . . . . . . . . . . . . . 68 C.2 OFDM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

x

D Modulation schemes 70 D.1 BPSK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 D.2 QPSK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 D.3 QAM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 E Attenuation measurement report E.1 Purpose . . . . . . . . . . . . . . E.2 Requirements and equipement . . E.3 Test setup . . . . . . . . . . . . . E.3.1 Connectivity setup . . . . E.3.2 Location . . . . . . . . . . E.4 Performing the measurement . . E.5 Results 08-12-2011 . . . . . . . . E.5.1 Summary . . . . . . . . . E.6 Results 11-12-2011 . . . . . . . . E.6.1 Summary . . . . . . . . . E.7 Conclusion . . . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

72 72 72 73 73 73 73 74 76 77 79 80

F Blocker measurement report F.1 Purpose . . . . . . . . . . . . . F.2 Requirements and equipement . F.3 Test setup . . . . . . . . . . . . F.3.1 Connectivity setup . . . F.3.2 Location . . . . . . . . . F.4 Performing the measurement . F.5 Results from the 13-01-2012 . . F.5.1 Conclusion . . . . . . . F.6 Results from the 14-01-2012 . . F.6.1 Conclusion . . . . . . . F.7 Results from the 15-01-2012 . . F.7.1 Conclusion . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

81 81 81 82 82 82 83 84 88 89 91 92 96

xi

. . . . . . . . . . . .

xii

List of Abbreviations 3GPP

3rd Generation Partnership Project

ACK

Acknowledgement

AMC

Adaptive Modulation and Coding

ARQ

Automatic Repeat Request

AS

Access Stratum

AWGN Additive white Gaussian noise BER

Bit Error Ratio

BLER

Block Error Ratio

BPSK Binary Phase-Shift Keying CCI

Co-Channel Interference

CDM

Code-Division Multiplexing

CDMA Code-Division Multiple Access CP

Cycle Prefix

CQI

Channel Quality Indicator

eNB

evolved-Node B

EPC

Evolved Packet Core

EPS

Evolved Packet System

FDD

Frequency-division Duplexing

FDM

Frequency-Division Multiplexing

FDMA Frequency-Division Multiple Access FEC

Forward Error Correction

GP

Guard Period

GSM

Global System for Mobile Communications

HARQ Hybrid Automatic Repeat Request ISI

Inter-Symbol Interference

KPI

Key Performance Indicator

LTE

Long Term Evolution

MAC

Medium Access Control

MME

Mobility Management Entity

MCS

Modulation and Coding Scheme

MIMO Multiple Input-Multiple Output NACK Negative Acknowledgement NAS

Non-Access Stratum

OFDM Orthogonal Frequency Division Multiplexing OFDMA Orthogonal Frequency Division Multiple Access P-GW Packet Data Network Gateway PAPR Peak-to-Average Power Ratio PDCP Packet Data Convergence Protocol pdf

probability density function

PDN

Packet Data Network

PDU

Packet Data Unit

PHY

Physical layer

PSK

Phase Shift Keying

QAM

Quadrature Amplitude Modulation

QPSK Quadrature Phase-Skift Keying RAN

Radio Access Network

RB

Ressource Block

RLC

Radio Link Control

RRC

Radio Resource Control

RSRP

Reference Signal Receive Power

RSRQ Reference Signal Receive Quality RSSI

Received Signal Strength Indicator

RV

Redundancy Version xiv

S-GW Serving Gateway SC-FDMA Single-Carrier Frequency Division Multiple Access SAE

System Architecture Evolved

SDM

Space-Division Multiplexing

SDMA Space-Division Multiple Access SDU

Serving Data Unit

SINR

Signal-to-Interference plus Noise Ratio

SNR

Signal-to-Noise Ratio

TDD

Time-division Duplexing

TDM

Time-Division Multiplexing

TDMA Time-Division Multiple Access UE

User Equipment

xv

xvi

List of Figures 4.1 4.2

4.3

4.4 5.1 5.2 5.3

The network architecture in Long Term Evolution (LTE), showing how the system is split and interconnected. . . . . . The protocol stack in the user plane. It consists of 3 layers. Layer 1 is the Physical layer, Layer 2 consisting of 3 sublayers; Medium Access Control (MAC), Radio Link Control (RLC), Packet Data Convergence Protocol (PDCP), and Layer 3 as the Radio Resource Control (RRC) layer. . . . . . . . . . . . The protocol stack for the control plane. It functions exactly as in the user plane, however serves mainly as a carrier for control messaging from the RRC which may contain NonAccess Stratum (NAS) messaging, rather than user data. . . Overview of the ideal system. The main focus of the measurements will be on the physical layer downlink side. . .

6

7

8 9

5.8

Frame structure for type 1 for FDD mode. (Rumney 2009) . . Frame structure for type 2 for TDD mode. (Rumney 2009) . Orthogonal Frequency Division Multiplexing (OFDM) symbol structure for normal cyclic prefix case. (Rumney 2009) . . OFDM symbol versus cyclic prefix + OFDM symbol . . . . . Ressource grid for 1 uplink slot (a) and 1 downlink slot (b) . Subcarrier allocation in OFDM and OFDMA. By assigning different OFDM sub-channels, Frequency-Division Multiple Access (FDMA) is achieved. . . . . . . . . . . . . . . . . . . . Transmission of a series of QPSK symbols in both OFDMA and SC-FDMA . . . . . . . . . . . . . . . . . . . . . . . . . . Simplified signal generation of SC-FDMA and OFDMA . . .

6.1 6.2

Intermodulation distortion from third order product . . . . . intermodulation interference . . . . . . . . . . . . . . . . . . .

24 25

7.1 7.2

The basic construction of a -20dB directional coupler . . . . Block diagram of how the user case is assumed. . . . . . . . .

29 29

5.4 5.5 5.6 5.7

xvii

11 11 12 12 13 18 19 20

7.3

UE reporting CQI to the eNB, which afterwards sends a request of which modulation and coding the UE should use next. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

8.1

Test setup for measurement the impact of different tion while downloading in the 1800 MHz band. . . 8.2 Throughput from Meas-1. Throughput is measured 8.3 Measured SINR from Meas-1. . . . . . . . . . . . . 8.4 Reported CQI index from Meas-1. . . . . . . . . . 8.5 Measured RSSI in Meas-1. . . . . . . . . . . . . . . 8.6 Throughput from Meas-2. It is measured in kbit/s. 8.7 SINR from Meas-2. . . . . . . . . . . . . . . . . . . 8.8 Reported CQI index, from Meas-2. . . . . . . . . . 8.9 Resource block allocation in Meas-2 . . . . . . . . 8.10 Resource block allocation in Meas-1 . . . . . . . . 8.11 Measured RSRQ from Meas-2. . . . . . . . . . . . 8.12 Measured RSRQ from Meas-1. . . . . . . . . . . .

attenua. . . . . . in kbit/s. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

9.1

Test setup for measurement the intermodulation on 1800 MHz band . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Throughput in 1869.7 MHz . . . . . . . . . . . . . . . . . . . Throughput in 1870.2 MHz . . . . . . . . . . . . . . . . . . . RSRQ in 1869.7 MHz . . . . . . . . . . . . . . . . . . . . . . RSRQ in 1870.2 MHz . . . . . . . . . . . . . . . . . . . . . . Resource Block in 1869.7 MHz . . . . . . . . . . . . . . . . . Resource Block in 1870.2 MHz . . . . . . . . . . . . . . . . . SINR in 1869.7 MHz . . . . . . . . . . . . . . . . . . . . . . . SINR in 1870.2 MHz . . . . . . . . . . . . . . . . . . . . . . . Received signal strength indication in 1869.7 MHz . . . . . . Received signal strength indication in 1870.2 MHz . . . . . . channel quality indicator in 1869.7 MHz . . . . . . . . . . . . channel quality indicator in 1870.2 MHz . . . . . . . . . . . . Intermodulation distortion on TELIA’s 1800 MHz band starting at the center frequency and move 8 steps with 1 MHz each . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Throughput with different modulation on the intermodulated distorted signal . . . . . . . . . . . . . . . . . . . . . . . . . . RSRQ with different modulation on the intermodulated distorted signal . . . . . . . . . . . . . . . . . . . . . . . . . . Resource Block with different modulation on the intermodulated distorted signal . . . . . . . . . . . . . . . . . . . . . . . SINR with different modulation on the intermodulated distorted signal . . . . . . . . . . . . . . . . . . . . . . . . . .

9.2 9.3 9.4 9.5 9.6 9.7 9.8 9.9 9.10 9.11 9.12 9.13 9.14 9.15 9.16 9.17 9.18

xviii

31 34 35 35 36 36 37 37 37 38 38 39 39 41 42 42 43 43 44 44 44 45 45 45 46 46 47 47 48 48 49

9.19 RSSI with different modulation on the intermodulated distorted signal . . . . . . . . . . . . . . . . . . . . . . . . . . 9.20 channel quality indicator with different modulation on the intermodulated distorted signal . . . . . . . . . . . . . . . . . 9.21 Intermodulation distortion on TELIA’s 1800 MHz band . . . 9.22 Throughput with -40dB attenuation . . . . . . . . . . . . . . 9.23 Throughput with -60dB attenuation . . . . . . . . . . . . . . 9.24 RSRQ with -40dB attenuation . . . . . . . . . . . . . . . . . 9.25 RSRQ with -60dB attenuation . . . . . . . . . . . . . . . . . 9.26 Resource Block with -40dB attenuation . . . . . . . . . . . . 9.27 Resource Block with -60dB attenuation . . . . . . . . . . . . 9.28 SINR with -40dB attenuation . . . . . . . . . . . . . . . . . . 9.29 SINR with -60dB attenuation . . . . . . . . . . . . . . . . . . 9.30 RSSI with -40dB attenuation . . . . . . . . . . . . . . . . . . 9.31 RSSI with -60dB attenuation . . . . . . . . . . . . . . . . . . 9.32 channel quality indicator with -40dB attenuation . . . . . . . 9.33 channel quality indicator with -60dB attenuation . . . . . . .

49 49 50 51 51 51 52 52 52 53 53 53 54 54 54

A.1 Simplified MIMO. Transmitter (Tx0,Tx1), Receiver (Rx0,Rx1) 64 A.2 Multipath with signal diversity. Transmitter (Tx0,Tx1), Receiver (Rx0,Rx1) and obstacles (A,B,C,D) . . . . . . . . . 65 C.1 Channel access using FDMA, TDMA, and CDMA in relation to each other. (Flintoff et al. 2000) . . . . . . . . . . . . . . . C.2 Multiple modulated OFDM subcarriers with constant amplitude. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D.1 1-bit signal . . . . . . . . . . . . . . . . . . . . . . . . . . . . D.2 QPSK modulated signal, which comprises of 2-bit symbols . . D.3 The two QAM schemes. 1) shows 16 QAM while 2) shows 64 QAM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E.1 Measurement test setup for determining the impact of different attenuation while downloading, in the 1800 MHz band. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E.2 Location of UE and eNB. The measurement was performed approximately at position A), while the eNB is located approximately at position B). . . . . . . . . . . . . . . . . . . E.3 Throughput is measured in kbit/s. . . . . . . . . . . . . . . . E.4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E.5 Throughput is measured in kbit/s. . . . . . . . . . . . . . . . E.6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F.1 Test setup for measurement the impact of different attenuator size while downloading in the 1800 MHz band . . . . . . . . . xix

68 69 70 71 71

73 74 75 76 78 79 82

F.2 Location of UE and eNB. The measurement was performed approximately at position A), while the eNB is located approximately at position B). . . . . . . . . . . . . . . . . . . F.3 TP in kbit/s, RSRQ and RB with different level of attenuation F.4 SINR in dB, MCS index, RSSI in dB and CQI with different level of attenuation . . . . . . . . . . . . . . . . . . . . . . . . F.5 TP in kbit/s, RSRQ and RB with different level of attenuation F.6 SINR in dB, MCS index, RSSI in dB and CQI with different level of attenuation . . . . . . . . . . . . . . . . . . . . . . . . F.7 Every 1 minute the frequency was increased by 1 MHz. The extra time was to insure that the interference signal was out of range. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F.8 TP in kbit/s, RSRQ and RB with different frequencies . . . . F.9 SINR in dB, MCS index, RSSI in dB and CQI with different frequencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . F.10 Every 30 sec. the frequency was increased by 1 MHz. The bandwidth was 10 MHz and the extra time was to insure that the interference signal was out of range. . . . . . . . . . . . . F.11 TP in kbit/s, RSRQ and RB . . . . . . . . . . . . . . . . . . F.12 SINR in dB, MCS index, RSSI in dB and CQI . . . . . . . . F.13 TP in kbit/s, RSRQ and RB . . . . . . . . . . . . . . . . . . F.14 SINR in dB, MCS index, RSSI in dB and CQI . . . . . . . .

xx

83 85 86 87 88 89 90 91 92 93 94 95 96

List of Tables 5.1 5.2

Transmission bandwidth configuration. (3GPP 2011a) . . . . Example of Forward Error Correction. Adding redundancy by receiving triplets of the symbol, reduces the possibility of errors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-bit CQI table showing the corresponding modulation type. (3GPP 2011c, Table 7.2.3-1) . . . . . . . . . . . . . . .

14

6.1 6.2

Intermodulation distortion in the 1800 MHz band . . . . . . . Intermodulation distortion in Telia’s 1800 MHz band . . . . .

24 26

8.1

List of equipment for initial measurements. . . . . . . . . . .

33

E.1 List of equipment required to perform the measurement. . . . E.2 List of downloads used for the measurement. . . . . . . . . . E.3 Measurement procedure for attenuation measurement. . . . .

72 73 74

F.1 List of equipment required to perform the measurement. . . . F.2 List of downloads used for the measurement. . . . . . . . . . F.3 Measurement procedure for SMIQ measurement. . . . . . . .

81 82 84

5.3

xxi

15 17

xxii

Chapter 2

Introduction Over the last few decade, wireless telecommunication have increased dramatically, especially after the introduction of the cellular phone, which now can not only transmit voice, but as well receive e-mail, browse the World Wide Web, and much more. In wireless telecommunication different standards are used in order to provide connectivity for the user in the rapid grow in the usage of the frequency spectrum. With the fusion of usage in the wireless telecommunication that include the same task as before only was possible in the normal wired communication(modem, ADSL, broadband and a lot more), the demand for speed and availability from the daily user have become increasingly real. This is seen especially in studies which have shown that up to 88 % of danish families have at least one household computer, and with 86 % of these families having internet access (og Telestyrelsen 2011). Thus the introduction of LTE. With this new technology, a wide range of improvements are brought forward, such as improved connectivity and availibility, as well as higher speeds. In some countries, LTE is still under deployment, however in most Western countries it is currently available by Telecom Service Providers. In Denmark, LTE is provided on two frequency bands. One of these are the 1800 MHz bandwidth, which as of May 1st, 2011, was released for commercial use with other telecommunication technologies, other than Global System for Mobile Communications (GSM). This enables LTE to use this frequency band. The 1800 MHz frequency is divided into two sub frequency bands. 1710-1785 MHz is dedicated to uplink, and 1805-1880 MHz is dedicated to downlink (og Telestyrelsen 2009c). This project deals with the use of Telia Nättjänster Norden AB’s 1761.3-1773.1 MHz uplink and 1856.3-1868.1 MHz downlink. Since Telia bought Orange A/S in 2004, Telia was able to increase their 1800 MHz band to 1761.3-1784.9 MHz uplink and 1856.3-1879.9 MHz downlink (og Telestyrelsen 2009b). The focus of this project is primarily to examine how poor signal strength

1

and intermodulation in Real life measurements affects throughput. At the same time the modulation of the network will be analysed.

2

Chapter 3

Project Goal The purpose of this project is to examine how LTE’s physical layer throughput is affected by disturbances, such as noise and interference, from other communication technologies in the same frequency band. These disturbances are seen from a user applications point of view, by introducing noise in various ways into the channel. This may be done either directly into the channel or by introducing distortion from third order intermodulation products, over the channel bandwidth, afterwhich the effects on the system are analysed. During the project, this report will include; • A pre-analysis of the physical layer in the LTE standard, from preliminary experiments – with focus on LTE’s adaptive modulation mechanism • A theoretical inference of how physical throughput is affected by different levels of Signal-to-Noise Ratio (SNR). • Experiments on deployed LTE networks with the introduction in different levels of SNR. • A comparison of theoretical and experimental results. This project will mainly focus on the 1800 MHz frequency area.

3

Part I

Analysis

4

Chapter 4

Long Term Evolution System Overview The LTE system architecture is designed with the goal of supporting a packet-switched networking paradigm. This allows a highly simplified architecture in which there only exists two types of nodes. The base station, also known as evolved-Node B (eNB) and the Mobility Management Entity (MME). This chapter serves as a brief discussion on the LTE system architecture, inorder to give a quick overview, before going into depth with the physical layer. First is a description of the network architecture, and afterwards the protocol architecture is briefly described.

4.1

Network architecture

System Architecture Evolved (SAE) is the core network architecture of 3rd Generation Partnership Project (3GPP)’s LTE wireless communication standard. It allows for a more simplified architecture, with support for higher throughput/low latency to non-3GPP networks and for a better mobility between 3GPP legacy systems. All network interfaces are based on IP protocols, where the eNBs are interconnected by means of an X21 interface, and the to MME through an S12 interface. Figure 4.1 on the following page shows how the eNBs and MME are interconnected. 1

The X2 interface is a communication protocol, by which the eNBs are interconencted with. 2 The S1 interface, connects the eNB to the Evolved Packet Core (EPC).

5

Figure 4.1: The network architecture in LTE, showing how the system is split and interconnected. The functional split between the eNB and the MME results in two logical gatway entities being defined. The Serving Gateway (S-GW) acts as a local mobility anchor for the user plane, during handovers and anchoring LTE and other 3GPP technologies, while at the same time forwarding and receiving, user data packets. The Packet Data Network Gateway (P-GW) allows interfacing with other external Packet Data Networks (PDNs) such as the Internet, along with other IP functions. Furthermore, the P-GW acts as an anchor between 3GPP and non-3GPP technologies, like WiMAX. The eNB functions mainly by performing header compression, ciphering and providing a reliable delivery of packets.

4.2

Protocol architecture

Besides SAE are the NAS protocols. These form the highest stratum of the control plane between the User Equipment (UE) and MME. The NAS performs functions such as Evolved Packet System (EPS) bearer management, authentication and security control. 6

In figure 4.2, is the user plane protocol stack.

Figure 4.2: The protocol stack in the user plane. It consists of 3 layers. Layer 1 is the Physical layer, Layer 2 consisting of 3 sublayers; MAC, RLC, PDCP, and Layer 3 as the RRC layer. The protocol stack is divided into three layers, where Layer 2 is subdivided into three sublayers, namely the MAC, RLC and PDCP. Figure 4.3 on the following page shows the protocol stack for the control plane. MAC, RLC and PDCP behave exactly as they do in the user plane, however they function mainly to carry control messaging from the RRC.

7

Figure 4.3: The protocol stack for the control plane. It functions exactly as in the user plane, however serves mainly as a carrier for control messaging from the RRC which may contain NAS messaging, rather than user data. As mentioned before, Layer 2 consists of the three sublayers. The main functions of the MAC layer is to perform multiplexing of data from logical channels, which are then to be delivered to the physical layer via the transports channels. Moreover, the MAC performs error correction from HARQ, and diciding which UEs will be allowed to send or receive data on the shared physical resource (Rumney 2009). The RLC acts as an interface between the higher layers of the stack and the MAC layer. Basically it acts more as a router, since its main purpose is to interface and buffer because the MAC has no buffer capabilities (Rumney 2009). Next is the PDCP layer. This layer performs functions such as header compression, and decompression, ciphering and passing Serving Data Units (SDUs) and Packet Data Units (PDUs) (Rumney 2009). Finally, is the Physical layer (PHY). This is the lowest layer in the LTE protocol and covers the downlink transmission from the eNB to the UE, and the uplink transmission from the UE to eNB. The physical layer is of particular interest, and it is in this layer that will be primarily focused on. In figure 4.4 on the next page is a simple overview of the system.

8

Figure 4.4: Overview of the ideal system. The main focus of the measurements will be on the physical layer downlink side. It is the physical layer KPIs that are of main interest. The PHY layer will be discussed in further detail in chapter 5.

9

Chapter 5

LTE Physical Layer In order to better understand the principles in LTE, it is neccessary to understand how the lowest layer of LTE, the Physical layer, works. During this chapter an introduction to the Physical layer will be given, as well as an overview of how the Physical layer is constructed. This chapter is primarily based on Rumney (2009), except where stated otherwise.

5.1

Introduction To The Physical Layer

The Physical layer of LTE covers the downlink and uplink tranmission between the UE and the eNB base transceiver station. The Physical layer supports two multiple access schemes. These multiple access schemes are OFDMA and SC-FDMA, which will be discussed in detail later in this chapter. Addtionally to OFDMA and SC-FDMA, both paired and unpaired spectra are supported by using Frequency-division Duplexing (FDD) and Time-division Duplexing (TDD), respectively.

5.2

Architectural Overview

There are defined two types of Physical layer channels. These two types are; the physical channels, which carry information from the higher layers, as well as data, and the physical signals, which are generated in the physical layer for cell identification, radio channel estimation, and system synchronization. Two types of frames are also defined in the Physical layer; type 1 for FDD and type 2 for TDD.

5.2.1

Frame And Slot Structure

The frame structure defines frame, subframe, slot and symbol in the time domain. Each time length is defined in units of TS = 1/(15000 · 2048) = 32.55 ns. 10

Figure 5.1: Frame structure for type 1 for FDD mode. (Rumney 2009) The frame structure seen in figure 5.1 is frame type 1 defined for FDD mode. Each frame consistes of 10 subframes, which consists for 2 slots. One radio frame is 10 ms long. In FDD mode, both the uplink and downlink scheme use the same frame structure however, they uses different spectra. Frame structure type 2 is defined for TDD mode, and is seen in figure 5.2.

Figure 5.2: Frame structure for type 2 for TDD mode. (Rumney 2009) Frame structure type 2 is also defined for 7 different configurations, where each radio frame is 10 ms long and consists of two half frames. Futhermore, each half frame consists of 5 subframes, which are 1 ms long. The 7 configurations of frame structure type 2 can be seen in table 4.2-2 in 3GPP (2011b). Inter-Symbol Interference (ISI) and cyclic prefixing In OFDM systems, as well as SC-FDMA in this context, one of the key advantages is the introduction of a Guard Period (GP) between each symbol. This GP gives the ability to protect against multipath delay spread, and thus 11

eliminates ISI. If the GP is longer than the delay spread in the channel, and each OFDM symbol is cyclically extended into the GP, then the ISI can be completely removed. In figure 5.3, an example of an OFDM symbol structure can be seen.

Figure 5.3: OFDM symbol structure for normal cyclic prefix case. (Rumney 2009) By cyclic prefixing, the symbol will be prefixed with a repetition of the symbol sequence itself (Haykin 2000). Thus by introducing cyclic prefixing, OFDM and SC-FDMA systems are able to protect against multipath spreads of up to 10 km. In figure 5.4 the last part of the OFDM signal is added in the beginning if the OFDM signal. The length of the cyclic prefix is chosen to accommodate the wireless channel’s maximum delay spread.

Figure 5.4: OFDM symbol versus cyclic prefix + OFDM symbol It should be noted that delay spreads represent the variation in path delay, and can be interpreted as the difference in time of arrival between extreme multipath components, i.e. earliest and latest component.

12

Resource elements and blocks Within the Physical layer, a resource element is the smallest unit and extends over one symbol (OFDM or SC-FDMA) in the time domain, and one subcarrier in the frequency domain.

(a)

(b)

Figure 5.5: Ressource grid for 1 uplink slot (a) and 1 downlink slot (b) The Ressource Block (RB) is the smallest unit, that can be scheduled. It physically occupies 180 kHz in frequency, and 0.5 ms in time. Thus for a channel bandwidth of 10 MHz (including guardspaces, etc.), a maximum of 50 RBs can be alotted. For the full channel bandwidth of 20 MHz, there are 100 RBs available. In most systems the transmission bandwidth is fixed, however OFDM systems enables the possibilty for flexible bandwidths. Subcarrier spacing is determined by an inverse of the FFT intergration time, thus giving LTE the flexibility of having six different transmission bandwidth configurations to choose from. The different transmission bandwidth configurations can be seen in table 5.1 on the following page

13

Channel bandwidth [MHz]

1.4

3

5

10

15

20

Configuration in MHz Configuration in RB

1.08 6

2.7 15

4.5 25

9 50

13.5 75

18 100

Table 5.1: Transmission bandwidth configuration. (3GPP 2011a) The channel bandwidth which is defined in MHz, represents the nominal occupied channel. Basically, this is the bandwidth which the operator, such as Telia, provides. In Denmark, Telia provides LTE with 10 MHz bandwidth, whereas Telenor has 20 MHz. The transmission bandwidth which is defined in units of RB and represent the maximum of RB that can be transmitted, for any given channel bandwidth. The architectural overview, however, does not only cover the the physical allocation of resource blocks and the frame structure of the LTE frame. Next comes the physical layer signalling, which is a key part of the Physical layer, since it contains different error correction methods, among other aspects. Physical layer signalling Besides the physical carriers of data, which are RBs there are two key measures of performance in communications systems; throughput and latency. Throughput is the actual amount of data being tranmitted, and is usually measured in bit per second. In comparison, shipping a box of DVDs overnight would result in a superb throughput. Therefore, high throughput is desired if a user wishes to download large files. However, in the shipping example, the delay would not be acceptable, since the user would have to wait some time before recieving his shipment. In this case, it would result in high latency, because a low latency is desired to inorder to guarantee quick responses to a users requests, in applications such as VoIP, internet gaming, etc. To counteract low throughput and high latency, LTE employs a number of mechanisms in the physical layer. Two of these mechanisms are Hybrid Automatic Repeat Request (HARQ) and AMC. AMC will be discussed in section 5.3 on page 16 along with the modulation schemes. In order to ensure that data is sent reliably from one node to another, Automatic Repeat Request (ARQ) is used. This is an error detection mechanism, which requests a retransmission from the receiver, incase of a timeout. HARQ is a combination of ARQ, and Forward Error Correction (FEC) which is error correction technique by adding redundancy into the transmitted signal (Haykin 2000). A simple example of FEC can be seen in table 5.2 on the facing page.

14

Triplet received

Interpreted as

000 001 010 100 111 110 101 011

0 (error free) 0 0 0 1 (error free) 1 1 1

Table 5.2: Example of Forward Error Correction. Adding redundancy by receiving triplets of the symbol, reduces the possibility of errors. Given the instance that the binary sequence 1012 is to be transmitted, then by using FEC instead of transmitting a single bit at a time, at triplet of the bit is transmitted. Thus the sequence 1012 would result in 1110001112 . The added redundancy allows an error in any of the three samples to be corrected. There exists two types of HARQ. Type I HARQ which is the simplest form of HARQ, and Type II HARQ, whereas it is the Type II which is used in LTE. On the first transmission of the packets life, a subset of the coded bits are transmitted with enough information for the receiver to decode the original information of the packet and the CRC, with only a small amount of redundancy, thus resulting in high efficiency under good channel conditions. However, if the packet is not decoded correctly, a retransmission is triggered. Where the benefits of HARQ comes into light is rather resending the same data, the HARQ chooses another set of encoded bits, still representing the original information bits and the destination node adds this new information to what was received earlier. This HARQ process is a stop-and-wait protocol, meaning that once the HARQ process has sent its packet, it stops and waits for an ACK/NACK from the destination, before sending the next packet. The different transmitted packet versions from the HARQ process contains different mixes of redundancy and systematic bits. These versions are called Redundancy Versions (RVs). These RVs er sequenced through in LTE by the HARQ process, until the packet has either been received correctly, or the maximum of retransmissions have been reached, in which case HARQ declares a failure and hands it over to the ARQ running in the RLC layer. At this point AMC takes over.

15

5.3

Modulation Scheme and Coding

LTE introduces the use of different modulation schemes, depending on the uplink and downlink. In appendix C on page 68, OFDM and SC-FDMA will be discussed more thoroughly. In this section the two modulations schemes and AMC are described.

5.3.1

AMC

When HARQ has declared a failure in retransmissions, it hands the packet over to ARQ and AMC takes over. It attempts to match the transmissions from the HARQ process to the channel conditions in order to choose the appropriate coding. During good channel conditions AMC would employ a higher modulation, such as 64-Quadrature Amplitude Modulation (QAM) which uses less redundancy in the transmission. This would results in a larger transport block to be carried in the allocated channel. However, if the channel suffers from poor conditions, AMC would choose a lower order of modulation. Such a modulation would be Quadrature Phase-Skift Keying (QPSK). With QPSK more redundancy bits would be sent to in order to improve the probability of reception, but then employing a smaller transport block. If the packet error rate is very low, it would imply the the modulation depth is to high or to much redundancy is used. This results in a smaller transport block size, and thus ultimately reduces throughput. Moreover, if the packet are large, then the packet error rate would be high, and again result in reduced throughput. In order for AMC to work it is required that the eNB is informed about the channel quality, seen by the UE. This is done through Channel Quality Indicator (CQI) information, reported by the UE in the uplink. The CQI index and its corresponding modulation scheme can be seen in table 5.3 on the facing page.

16

CQI Index

Modulation

0 1 2 3 4 5 6

Out of range QPSK QPSK QPSK QPSK QPSK QPSK

7 8 9

16-QAM 16-QAM 16-QAM

10 11 12 13 14 15

64-QAM 64-QAM 64-QAM 64-QAM 64-QAM 64-QAM

Table 5.3: 4-bit CQI table showing the corresponding modulation type. (3GPP 2011c, Table 7.2.3-1)

5.3.2

Downlink: Orthogonal Frequency Division Multiple Access (OFDMA)

Since the introduction of small but powerful DSP(Digital signal processor) the OFDM techniques can be utilized in modern telecommunication. In the downlink, LTE utilizes OFDMA, which is a variant of OFDM C.2 on page 69 allowing multiple users to access the network. In OFDMA the channel is divided into many narrow subchannels and transmitted parallel. This increase the symbol duration and reducing the intersymbolinterference(ISI). The main advantage in OFDMA is its ability to allocate subcarriers dynamically, allowing users to access the network. This principle can be seen in figure 5.6 on the following page.

17

Figure 5.6: Subcarrier allocation in OFDM and OFDMA. By assigning different OFDM sub-channels, FDMA is achieved.

5.3.3

Uplink: Single-Carrier Frequency Division Multiple Access (SC-FDMA)

Two of the main concerns to the LTE uplink, however, was the power consumption in the UE terminals, as well as high Peak-to-Average Power Ratio (PAPR) which is a comparison of the peak power detected over a period of samples at the time period. SC-FMDA can be seen as a DFT-spread OFDMA by using the time domain data signals and transform it to frequency domain by a DFT before parsing through OFDMA modulation. This techniques reduce the instantaneous transmit power implying increase power-amplifier efficiency, low-complexity and flexible bandwidth assignment. Using SC-FDMA allows the usage of a single carrier transmission system such as GSM and Code-Division Multiple Access (CDMA). These types of systems have a low PAPR. SC-FDMA utilizes a single-carrier transmitting signal in contrast to OFDMA that use a multicarrier transmission scheme. In 5.7 on the next page a graphical comparison of OFDMA and SC-FDMA are shown.

18

Figure 5.7: Transmission of a series of QPSK symbols in both OFDMA and SC-FDMA In SC-FDMA, signals are built up in units of 12 subcarriers. However in figure 5.7 there are only four subcarriers used over two symbol periods represented by QPSK modulation. The obvious difference between OFDMA and SC-FDMA is that OFDMA transmit the four QPSK data symbol in parallel, while SC-FDMA transmit the four QPSK data symbols in series. An overall model of how the data bits get through SC-FDMA and OFDMA is to find in figure 5.8 on the next page. Both the SC-FDMA and OFDMA techniques is represented. First the data gets and transforms to a time domain waveform. By using a DFT the signal gets to the frequency domain and map one more time. This time the same same is existing in the OFDMA that the symbols gets map to subcarriers. Then the IFFT gets the subcarriers back in a time domain to be unconverted for the transmission. When received the inverse process take place.

19

Figure 5.8: Simplified signal generation of SC-FDMA and OFDMA

5.4

Summary

In this chapter details of the Physical layer has been described. At first an architectural overview of the physical layer in LTE has been given. The Physical Layer consists of several error correction mechanisms, which enables the opportunity for a low probability of error. The use of both OFDMA for downlink and SC-FDMA for the uplink is only possible since the introduction of powerful and small DSP(Digital signal processor). This give a range of improvement by access mode seen in OFDMA and low power consumption as in SC-FDMA.

20

Chapter 6

Concepts Of Interference In this chapter, some concepts of interference will be discussed. Interference can be found everywhere in a communication system, and can cause errors stemming from transmitted symbols interfering with eachother or from a noisy channel resulting in a receiver having trouble with distinguishing between wanted signal and background noise. In the following sections, noise and interference will be discussed, and afterwards some aspects on intermodulation, and what influences if makes on a communication system.

6.1

Noise

One way that interference can be regarded as, is noise. Noise comes in different forms and can be defined as an unwanted and random signal or disturbance. It can originate from different places, and given a communication system, can be introduced either before or after the decoder/encoder. In this section, two types of noise will be considered. These are Johnson-Nyquist noise and Gaussian noise.

6.1.1

Johnson-Nyquist noise

Thermal noise, which it is also known as, arises from the random motion of electrons in a conductor. Thermal noise is expressed as PdB = 10 · log10 (kB · T · ∆f · 1000)

21

(6.1)

where, P is the noise power, kB is Boltzmann’s constant, T is the absolute temperature ∆f is the frequency bandwidth

6.1.2

[dBm]   J K [K] [Hz]

Gaussian noise

Another type of noise is Gaussian noise in which its probability density function (pdf) is equal to that of the normal distribution. The normal distribution has a “bell-shaped” pdf, and is regarded as one of the most prominent probability distributions in statistics due to its applicability as a simple model on complex systems. A special form of this type of noise is white Gaussian noise, in which all values in any pairs are uncorrelated. However, Gaussian noise is most commonly used in applications as additeve white noise, in order to yield additive white Gaussian noise. Additive white Gaussian noise The Additive white Gaussian noise (AWGN) channel model which is widely used in communications (Land and Fleury 2007). Having a transmitted signal X(t), then X(t) will be superimposed by a stocastic noise signal W (t), in such a way that the transmitted signal will be received as, Y (t) = X(t) + W (t)

(6.2)

This means that some noise signal will be directly added to the transmitted signal.

6.1.3

Signal-to-Noise ratio

A way of measuring noise can be donw through SNR. SNR is a measure of comparing the level of desired signal to the level of background noise, and is defined as the power ratio between signal and noise. SNR =

Psignal Pnoise

(6.3)

where P is the average power. SNR is also most often expressed in decibels. Thus in dB, SNR is defined as,   Psignal SNRdB = 10 · log10 (6.4) Pnoise

22

6.2

Interference

When dealing with communication systems interference will always occur, in some way or another. It can be anything in which alters, modifies or disrupts the signal as it travels along a channel. In this section, mainly Co-Channel Interference (CCI) and Inter-Symbol Interference (ISI) will be regarded.

6.2.1

Co-Channel Interference

CCI, also known as crosstalk stems from when two different tranmitters attempst to transmit using the same frequency. Since the frequency spectrum over the last decade has become more and more crowded due to other technologies, it is becoming increasingly more difficult to divide the different frenquency bands required for these technologies. In mobile communications, the frequency spectrum is divided into non-overlapping cells. However, due to the crowded spectrum, it is neccessary to reuse frequencies. It is here where CCI arises. Even though two cells using the same frequency and situated far away from eachother, a signal from the undesired transmitters may still arive. This will lead to the signal from far away will be received and interfere with closer and “correct” signals.

6.2.2

Intersymbol interference

In contrast to CCI, ISI is a form of interference where one transmitted signal is blended with subsequent symbols. This is an unwanted effect, since it can be catagorised as noise, and making communication unreliable. And since, ISI is usually caused by a multipath propagation in a environment prone to reflections, it is especially evident in Multiple Input-Multiple Output (MIMO) systems. A method of counteracting ISI is by separating symbols with guard periods, as mentioned in section 5.2 on page 10

6.3

Intermodulation

In non-linear systems, all signals will produce second and third order products around their centerfrequencies. Given two frequencies f1 and f2 , these will produce second order products at 2f1 , f1 + f2 , 2f2 and the inverse. However, since the second order product are situated far away from their main frequencies, they will no immediate significance, and can thus be filtered away. On the other hand, with second order products removed, third order products are still in range of the signal of interest. Figure 6.1 on the next page shows how the problem of third order products still are in effect.

23

Figure 6.1: Intermodulation distortion from third order product Having third order products close to the main frequencies is not desired, since they will influence whatever signal may recide on the main frequency, and thus results in Intermodulation Distortion. In table 6.1 the intermodulation products which are produced in the 1800 MHz area, are listed. RX2 1805 1815 1825 1835 1845 1855 1865 1880

TX1

1710

1720

1730

1740

1750

1760

1770

1785

1900 1920 1940 1960 1980 2000 2020 2050

1890 1910 1930 1950 1970 1990 2010 2040

1880 1900 1920 1940 1960 1980 2000 2030

1870 1890 1910 1930 1950 1970 1990 2020

1860 1880 1900 1920 1940 1960 1980 2010

1850 1870 1890 1910 1930 1950 1970 2000

1840 1860 1880 1900 1920 1940 1960 1990

1825 1845 1865 1885 1905 1925 1945 1975

Table 6.1: Intermodulation distortion in the 1800 MHz band All frequencies that are in red, are those of interest. From the table it can be seen that in the case of having a TX1 in 1730-1785 MHz, and at the same time having an RX2 at 1805-1825 MHz, will directly results in third order intermodulation distortion on the RX1. In figure 6.2 on the facing page the two times downlink from a second provider RX2 minus one uplink from first provider TX1, will create intermodulated interference on the downlink of first provider. This is in contrast to figure 6.1 where the entire 1800 MHz downlink becomes intermodulated with the uplink area in the 1700 MHz frequency area. 24

Figure 6.2: intermodulation interference In the figure 6.2 the case of intermodulation distortion is shown. For this to happen it is a requirement that, 2 × RX2low band-edge − 1 × T X1high band-edge

(6.5)

is met. This means that intermodulation distortion will directly occur on the high band-edge of RX’s downlink area, given the case that some signals are generated on the high band-edge of the frequency area dedicated to uplink, at the same time with a signal on the low band-edge on the frequency area dedicated to downlink.

6.4

In Denmark

With Hi3G and TDC together, offering LTE downlink on 1805.1-1836.9 MHz (og Telestyrelsen 2009a, 2010), and Telia having the downlink on 1856.3-1879.9 MHz (og Telestyrelsen 2009c) and uplink on 1761.3-1784.9 MHz, this phenomenon is very much evident.

25

RX2 Hi3G 1805 Hi3G 1815 TDC 1825 TDC 1835

Telia TX

1750

1760

1770

1785

1860 1880 1900 1920

1850 1870 1890 1910

1840 1860 1880 1900

1825 1845 1865 1885

Table 6.2: Intermodulation distortion in Telia’s 1800 MHz band In table 6.2 the direct impact of the two’s providers downlink frequencies is highlighted in read. In the case where a number of UEs are in the same cell, the Telia users downlink throughput is significantly impaired, if there are a number of TDC and/or 3 users uploading large amounts of data. The expectation is to reproduce this phenomenon with different signals level and modulation to see the impact of intermodulation distortion in a real life measurement. A range of different frequencies will be used to observe the effect of intermodulation distortion.

6.5

Summary

In this chapter, different aspects of noise and interference, has been discussed. It can be seen that inorder to be able to perform more indicative measurements it is neccessary to include the these different aspects of noise and interference. In the following chapters, two types of measurements will be performed. The basis of these measurements are introducing noise to the channel. However, the type of noise that will be used is based on inserting some attenuation. A more optimum way of doing this would be to introduce a better defined noise or interference channel, such as superimposing a AWGN channel model, or emulate CCI or ISI. Moreover, during a previous internship at Agilent Technologies, it was noticed that a phenomenon indicative of Intermodulation Distortion may be present. Through the analysis seen in section 6.3 on page 23 showed that theorethically this could occur. Thus creating a basis for an Intermodulation Distortion measurement.

26

Part II

Measurements

27

Chapter 7

Introduction Two main types of measurements were performed. The initial tests were made in order to establish what channel condition which are being dealt with, as well as to confirm that the network responds as expected. The second type of measurement was made to determine what would happen in case some unwanted signal would move in on the center frequency and cross it.

7.1

Throughput measurement

The purpose of this measurement is to give a preliminary indication as to which KPIs may be of special interest, as well as examine how LTE reacts to a simple attenuation of the downlink signal, seen from the UE. Additionally, this measurement will be able to give an indication as to what follow up measurements could be perforemed. However, the main focus of the measurement will be to give a basic idea as to what happens in the LTE system, when a channel becomes more an more impaired.

7.2

Intermodulation Distortion measurement

A blocker is a connection in which an unwanted signal is superimposed on the desired signal, by using a directional coupler.

7.2.1

Directional coupler

The directional coupler work by having two transmission lines close to each other see figure 7.1 on the next page. These closely align transmission lines passing energy through the one that is not block in the end.

28

Figure 7.1: The basic construction of a -20dB directional coupler

7.3

Case: Downlink blocks Uplink

In figure 7.2 is graphical representation of the User Case can be seen.

Figure 7.2: Block diagram of how the user case is assumed. It shows the specific phenomenon which the final measurement is based 29

on. By first determining an initial measurement inorder to gain a standpoint on what to do next, the attenuation measurement was performed. This measurement gave indications that in accordance to the background theory from chapter 5 on page 10 some changes in throughput should be seen, when some distortion is added to the channel. Afterwards, the Intermodulation Distortion measurement was set up. Based on the analysis in section 6.3 on page 23 the following case study is asserted. Is it possible to measure the effects of intermodulation distortion in a real life environment, based on the assumptions given in section 6.3 on page 23. Since the measurement is performed in a reallife environment the neccesity of testing the LTE system is an important step.

7.4

Key Performance Indicator (KPI)

RB RBs are the physical amount of bandwidth which can be scheduled on the eNB and are allocated to the UE. These were discussed in further detail in chapter 5 on page 10.

Physical Throughput Physical throughput can be defined as the actual throughput of data being transmitted in the physical layer. It is measured in kbit/s.

Reference Signal Receive Power (RSRP) RSRP is the most basic of the Physical layer measurements. It is an expression of the linear average of the downlink Reference Signals, in watts, across the channel bandwidth. Providing the UE with knowledge of absolute RSRP, is essential, since it provides information about the strength of cells from which path loss can be calculated, and afterwards used in optimization algorithms. However, the measure of RSRP give no indication of the signal quality.

Received Signal Strength Indicator (RSSI) RSSI represents the entire recieved power, which is radiated onto the UE, including wanted power from the serving cell, as well as all other co-channel power and noise.

30

Reference Signal Receive Quality (RSRQ) Given RSRP and RSSI, the RSRQ is an important measure, since it is defined as a ratio between RSRP and RSSI. A mathematical expression of RSRQ can be seen in equation 7.1 RSRP RSSI = 10 · log10 (50) + (RSRPdB − RSSIdB )

RSRQ = #RBdB +

(7.1) (7.2)

Signal-to-Interference plus Noise Ratio (SINR) SINR is a measure which calculates the ratio between the wanted signal and levels of interference and noise. It can be expressed mathematically as, SINR =

P I+N

(7.3)

where, P is the signal power, I is interference power N is the noise power

CQI The CQI report, uses measurements performed on the downlink conditions, inorder to report to the scheduler on which combination of modulation and coding would have resulted in a 10 % Block Error Ratio (BLER), if this combination had been used. In figure 7.3 the method how the UE reports CQI to the eNB is shown.

Figure 7.3: UE reporting CQI to the eNB, which afterwards sends a request of which modulation and coding the UE should use next.

31

Modulation and Coding Scheme (MCS) After the CQI has been reported, the eNB responds with an MCS index. MCS is an index from 0 to 31 which indicates to the UE, what the modulation and coding it should transmit on next.In figure 7.3 on the previous page the UE receives the MCS index and on the basis of this information, the data can be transmitted back with the chosen modulation and coding.

32

Chapter 8

Initial test Before being able to determine how LTE will react when the connection is introduced to a blocker, it is neccessary to determine how it reacts when LTE begins to suffer from a bad channel conditions. Therefore it is neccessary to find out, a method of introducing noise in the channel. One way to do this, is by attenuating the signal from the base station to the UE. Doing this will hopefully result in LTE attempting to perform rate adaption, by either changing the modulation scheme, requesting retransmission or in some other manner, inorder to sustain a reliable and stable conenction. The goal with this initial measurement is to gain some insight into the effects on throughput, SNR, and RSSI among others. Especially identify what happens, when some attenuation is introduced into the communications path. Inorder to realise this measurement, some different equipments and measurement tools are needed. In table 8.1, an overview of the equipment which has been utilised is listed. Device

AAU-nr.

Attenuators – 3 dB attenuator – 6 dB attenuator – 10 dB attenuator – 20 dB attenuator 1800 MHz IFA antenna

Note

Optimised only for downlink

Table 8.1: List of equipment for initial measurements.

8.1

Setup and test procedure

Figure 8.1 on the next page shows the test setup. To perform this measurement, the attenuation in the communication path was increased 33

with 10 dB intervals, except at the maximum attenuation. At this point the total attenuation was 39 dB.

Figure 8.1: Test setup for measurement the impact of different attenuation while downloading in the 1800 MHz band. With the attenuators in place, what this means is that generally less power which is radiated on the antenna will be transferred to the dongle. To perform these measurements it is neccessary to create some traffic, because it is imperative to make sure that as many ressource blocks are allocated, since the resource blocks aloocates the amount of bandwidth available. Basically, it is neccessary, to attempt to force the eNB to schedule as many resources as possible. When a connection has been established, the measurements are performed. The first measurement is made with no attenuation. This is to have control measurment as a comparison to the ones made with attenuation. Two sets of measurements with the attenuators were performed over the course of two days. They were performed in 2011 on December 8th and December 11th , and will be referred as Meas-1 and Meas-2, accordingly. Common for all plots in the following chapter, is that the time in minutes are plotted on the X-axis. In Meas-1, a simple explanation as to why each measurement run stands out, regarding to time duration is that during the measurements it was deemed that a 4 minute measurement was more then enough. So inorder to have time for other measurement they were shortened. Since these were static test, it seemed not to make any difference. Of course, this goes against 34

common measurement practice, and in hindsight probably should have done otherwise, these measurements were preliminary tests, meant to be used as a basis for planning future measurements. Meas-2, which was performed a few days later, are based on experiences from Meas-1. In Meas-2, a second control measurement was performed.

8.2

Observations

After performing the measurement 1 it could be seen from figure 8.2, that the attenuators clearly had some effect.

Figure 8.2: Throughput from Meas-1. Throughput is measured in kbit/s. An interesting part of these results are that between the measurement runs with 0 dB and 10 dB are some very distinct similarities. Throughput for 0 dB and 10 dB both stabilise at around 25Mbit/s , whereas SINR and CQI, which can be seen in figures 8.3 and 8.4 on the following page. However, it is also noted that in the SINR and CQI plots, also remains high for the 20 dB measurement.

Figure 8.3: Measured SINR from Meas-1.

35

Figure 8.4: Reported CQI index from Meas-1. In the meantime, it is also evident that as more attenuation is inserted, the CQI index falls. This is expected, since UE reports a CQI index which the UE believes is neccessary to withhold less then 10 % BLER. It can be seen in figure 8.4, that the UE requests a lower modulation, since the CQI index falls between every measurement run, which can be seen by comparing the results to table 5.3 on page 17. When no attenuation is inserted, the UE requests 64-QAM, while the UE requests a QPSK when the maximum attenuatation is inserted. At the same time the RSSI, differs with approximately 10 dB between every run, as seen in figure 8.5, which clearly indicates that the attenuators are lowering radiated power.

Figure 8.5: Measured RSSI in Meas-1. However, looking at the other KPIs, they show that perhaps only 10 dB attenuation does not force LTE as far down the BLER curve, in order to force LTE to attempt to uphold the connection, by performing rate adaption. The same can seen as well in the results from Meas-2. In figure 8.6 on the facing page can the throughput from Meas-2 be seen.

36

Figure 8.6: Throughput from Meas-2. It is measured in kbit/s. By looking at plots of SINR in figure 8.7 and CQI in figure 8.8 from Meas-2, most of the same observations can be made, as those seen in Meas1.

Figure 8.7: SINR from Meas-2.

Figure 8.8: Reported CQI index, from Meas-2. Just as in Meas-1, it can be seen that SINR is approximately the same level for 0 dB and 10 dB, with the same going for CQI. A note on Meas-2 is as mentioned earlier, that a second 0 dB measurement was performed. This measurement is a control measurement, whose main purpose is make sure that channel condition before and after the Meas-2 are the same. Of course, this is no guarantee, since conditions could have changed between the two 0 dB measurements, and change back. Inorder to be sure that tis 37

would not occur, the control measurement should have been done another way. This will be discussed in the final conclusion in chapter 10 on page 57. However, when looking at figure 8.8 on the preceding page, one may notice that suddenly around 3 minutes into the measurement the CQI index begins to rise. Recalling figure 8.6 on the previous page, throughput begins to fall, around the same time. This may seem unexpected, but when looking at the RBs in figure 8.9, drops significantly to slightly above an average of 40 RBs.

Figure 8.9: Resource block allocation in Meas-2 While the same goes for the RB allocation in Meas-1 which is seen figure 8.10.

Figure 8.10: Resource block allocation in Meas-1 This indicates that less bandwidth is scheduled to the user. Precisely what causes RBs to drop, is unsure. Perhaps some network issue occurs, or another UE with better condition is accessing the basestation. Now drawing the attention over to RSRQ in figure 8.11 on the facing page one may notice that the receive quality for is infact better.

38

Figure 8.11: Measured RSRQ from Meas-2. According to RSRQ the receive quality is much better then the others, so even though less bandwidth is available, the UE increases the modulation as seen by the CQI index in figure 8.8 on page 37 perhaps due to the better receive quality. RSRQ from Meas-1 can be seen in figure 8.12.

Figure 8.12: Measured RSRQ from Meas-1. In this figure it can be seen that RSRQ almost does not differ between each measurement run, in contrast to figure 8.11 from Meas-2.

8.3

Summary

In summary of the observations in the attenuations measurements, be concluded that a clear affect on Physical layer throughput can be measured, when introducing attenuation into the channel. By adding attenuation of 0dB,10dB and 20dB the performance is close to each other. 30dB and 39dB attenuation decreases the throughput drastically and some of the correlations between the RSSI, RSRP, CQI and SINR is striking.

39

Chapter 9

Intermodulation Distortion test 9.1

Introduction

In section of intermodulation distortion 6.3 on page 23 it has been shown how certain frequencies affect the upper 1800 MHz band. In this section the purpose is to investigate the effect of intermodulation distortion by adding a intermodulated distorted signal with different level of attenuation and modulation to the traffic. The SMIQ 06b is an signal generator able to produce a 5MHz signal with a range of different modulation applied to the signal. In these test the focus was to see; first the effect of the intermodulation distortion and subsequently to see the impact of the level of attenuation of these intermodulated distortion signals in LTE.

9.2

Setup and test procedure

In figure 9.1 on the next page the test setup. The setup was only performed in a SISO configuration due to lack of coupler. This is an very important aspect in this measurement case since this half of the throughput. In good signal condition we might theoretical be able to see throughput of 20M bit/s.

40

Figure 9.1: Test setup for measurement the intermodulation on 1800 MHz band

9.3

Observation

In this section three measurement will be evaluated and compared with the theory of LTE Physical layer from chapter 5 on page 10. All the information that is being used in the following subsection can be found in the measurement journal at the end of Appendix chapter.

9.3.1

13-01-2012

By selecting two frequencies 1869.7 MHz and 1870.2 MHz the impact of intermodulation distortion was measured. The only variable in these measurement was the attenuation on the signal (that used 1869.7 MHz and 1870.2 MHz) generated from the SMIQ. The intermodulated signal from the SMIQ was set to a level of attenuation of -80dB from the start and every minute decreased by -20dB ending with -40dB after 5 minutes.

41

Figure 9.2: Throughput in 1869.7 MHz

Figure 9.3: Throughput in 1870.2 MHz The throughput in figure 9.2 is first effected by the intermodulated distortion signal two minute in the measurement. In figure 9.3 the effect is seen slowly after the start of the measurement. At two minute in figure 9.2 the level of attenuation of the signal was -60dB. Further decrease of attenuation on the intermodulated signal from the SMIQ decrease the overall performance of the throughput significantly in both measurement. At the start throughput of 18–20M bit/s was reached but in the end of the measurement the throughput is between 5–8M bit/s.

42

Figure 9.4: RSRQ in 1869.7 MHz

Figure 9.5: RSRQ in 1870.2 MHz The same observation that was made during throughput can be seen in RSRQ in figure 9.4. First after -60dB the RSRQ value decreased to -13. But in figure 9.5 the RSRQ decrease later in the measurement and drops down to -25dB. Since RSRQ is a relation between RSRP and the RSSI the conclusion is that the reference signal power dos not vary in contrast to RSSI. In RSSI on figure 9.11 on page 45 the increase of received signal strength is more than 10dB. This decrease the quality of the channel by almost the same amount.

43

Figure 9.6: Resource Block in 1869.7 MHz

Figure 9.7: Resource Block in 1870.2 MHz Resource block from figure 9.6 and figure 9.7 is stable on the 49 resource block’s. This indicates that the above fading was not because of some other UE was using the same eNB. From table 5.1 on page 14 it is seen that the maximum of resource block is 50.

Figure 9.8: SINR in 1869.7 MHz

44

Figure 9.9: SINR in 1870.2 MHz In the beginning of the SINR measurement see figure 9.8 on the facing page and figure 9.8 on the preceding page the signal to interference and noise ratio is positive at a level of 15. After two minute at -60dB the SINR decrease to almost -15. This indicates that the interference and noise ratio has increased in both the measurement.

Figure 9.10: Received signal strength indication in 1869.7 MHz

Figure 9.11: Received signal strength indication in 1870.2 MHz In figure 9.10 and figure 9.10 RSSI start’s at -58dB to -56dB and by increasing the intermodulated distortion signals power, the total received signal strength increase to.

45

Figure 9.12: channel quality indicator in 1869.7 MHz

Figure 9.13: channel quality indicator in 1870.2 MHz Channel quality indicator in figure 9.12 and figure 9.12 reports the strength of the signal that the UE has to the eNB. In these two measurement we see how the channel quality decrease when increasing the intermodulation distortion signal. In figure 9.12 the decease is much greater ranging from 64-QAM and to QPSK, according to table 5.3 on page 17. However, in figure 9.12 the increase at the end comes from the RSRQ index that rise in level at the end.

9.3.2

14-01-2012

In this measurement setup the goal was to see if the modulation of the intermodulation distortion signal from the SMIQ 06B signal generator, had any impact on the performance on the network. The center frequency in Telia’s 1800 MHz ban is1868.2 and it was used to start the measurement by moving the intermodulated distorted signal from the center of the frequency and out of the range of Telia’s bandwidth. Every minute the intermodulated distorted signals was moved 1 MHz up in the frequnecy band with 8 steps. All the steps was performed with a -40dB on the intermodulated signal. Figure 9.14 on the next page illustrate the concept in this measurement. After 8 step with 1 MHz each the intermodulated distorted signal is out of the bandwidth of Telia’s downlink frequency.

46

Figure 9.14: Intermodulation distortion on TELIA’s 1800 MHz band starting at the center frequency and move 8 steps with 1 MHz each

Figure 9.15: Throughput with different modulation on the intermodulated distorted signal In the first measurement figure 9.15 the throughput decrease by all types of modulation. The only difference is that the higher modulation order 16QAM and 64QAM has a slighter greater throughput but is affected by the intermodulated distorted signal seeing an increase of the throughput at the end of the measurement. In section Intermodulation 6.3 on page 23 the direct effect of the distortion on the Telia downlink frequency is seen.

47

Figure 9.16: RSRQ with different modulation on the intermodulated distorted signal RSRP level in figure 9.16 show in this measurement that both the 16QAM and 64QAM modulation has a higher level.

Figure 9.17: Resource Block with different modulation on the intermodulated distorted signal The Resource Block number stays relative stable on the maximum number that this 10 MHz bandwidth can provide. The different modulations form did not cause any unexpected drops in the Resource Block count since the only disturbance is on the UE.

48

Figure 9.18: SINR with different modulation on the intermodulated distorted signal In the SINR on figure 9.18 the first real indication on that the type of modulation on the intermodulated distorted signal has no effect. The SINR is only effected by the moving of the intermodulated distorted signal throug the half bandwidth. At the end of grafe the SINR value increase above zero and increase almost to 10. This shows that the signal strength is better than the interference and noise that is present.

Figure 9.19: RSSI with different modulation on the intermodulated distorted signal By moving the intermodulated distorted signal through the upper Telia 1800 MHz band the received signal strength increase to see figure 9.19. This is the result of intermodulation distortion. Both the -60dB and the intermodulated distorted signal strength in merge together.

Figure 9.20: channel quality indicator with different modulation on the intermodulated distorted signal 49

In figure 9.20 on the previous page both BPSK and QPSK stays together and decrease from 8 and down to 3 and up again. 16QAM and 64QAM more or less independent on each other. The range from 16QAM starts at 5 CQI and decrease down to 3 and moves up to 8. 64QAM stays in the range of 8 to 5.

9.3.3

15-01-2012

This measurement is an semi reproduction of the measurement above (1401-2012). In figure 9.21 the basic of the measurement setup is displayed. In this measurement the whole bandwidth of 10 MHz is affected by the intermodulated distorted signal. This gives a range from 1863.2–1873.2MHz that the intermodulated signal is moved through. The x-axis increase in 1 MHz every 30sec. This was don because of the length of the measured range of 16 MHz and to ensure that the 5 MHz modulated distorted signal of 5 MHz was out of the 10 MHz bandwidth from Telia’s downlink range. The measurement was executed in to level of attenuation -40dB and -60dB.

Figure 9.21: Intermodulation distortion on TELIA’s 1800 MHz band

50

Figure 9.22: Throughput with -40dB attenuation

Figure 9.23: Throughput with -60dB attenuation The throughput from both figure 9.22 and figure 9.23 decrease when the intermodulated distorted signal is crossing the Telia’s downlink 10 MHz range. At the end of both the measurement the throughput increase to 13–14M bit/s

Figure 9.24: RSRQ with -40dB attenuation

51

Figure 9.25: RSRQ with -60dB attenuation RSRQ in figure 9.24 on the previous page and in figure 9.24 on the preceding page dos not give any indication on how the modulation and attenuation level effect the measurement.

Figure 9.26: Resource Block with -40dB attenuation

Figure 9.27: Resource Block with -60dB attenuation The only interrested thing in figure 9.26 and figure 9.27 is that in figure 9.27 the BPSK measurement (red) dos not stay at the 49 Resource block’s. The eNB can have some other UE using the channel and therefore 52

only in this time period down crease the Resource block’s.

Figure 9.28: SINR with -40dB attenuation

Figure 9.29: SINR with -60dB attenuation The SINR value in figure 9.28 and figure 9.29 differ from each other. In -40db the interference and noise level is hight and therefore the ratio between the signal strength and the disturbances from interference and noise is influencing the measurement. At the end of figure 9.28 with -40dB the SINR value increse to 15. The -60dB in figure 9.29

Figure 9.30: RSSI with -40dB attenuation

53

Figure 9.31: RSSI with -60dB attenuation The RSSI in figure 9.30 on the preceding page indicate the strength of the added intermodulated distorted signal. In contrast to this is in figure 9.31 the signal strength from the intermodulated distorted signal not a contributor to the increase of all the received signal strength.

Figure 9.32: channel quality indicator with -40dB attenuation

Figure 9.33: channel quality indicator with -60dB attenuation Both in figure 9.32 and figure 9.33 there is no direct effect on which modulation type is used and the level of noise and interference that the receiver is experienced. Both is decreasing and increase after leaving the Telia bandwidth.

9.4

Summary

In the results from the 13-01-2012 the attenuation level of the intermodulated distorted signal influence the measurement significantly. Both in 54

1869.7 MHz and in 1870.2 MHz the throughput decrease while increasing the signal power from the SMIQ 06B to the LTE dongle. RSRQ decrease (in 1870.2 dramatically) while RB stays untouched at 49. SINR moves from approximately +15dB to -15dB. MCS index increase in both measurement op to 20 and stabilising. From the results 14-01-2012 the no major impact on the difference modulation types that was use on the intermodulated distorted signal affected the LTE signal. Only the level of the intermodulated distorted signal affected the performance on the network. The power level of the intermodulated signal was visible in the RSSI measurement. In the last measurement form the 15-01-2012 the whole Telia 10 MHz bandwidth was used to cross with the intermodulated distortion signal from the SMIQ. Two levels of attenuation on the intermodulated distortion signal was used and the impact of the signal was visible both in throughput and in SINR

55

Part III

Assesment

56

Chapter 10

Final conclusion The goal of this project was to examine how LTE’s physical layer throughput was affected by outside disturbance on the channel. These disturbances could stem from different sources, and was in this project regarded as disturbances from noise and interference from other communication technologies, present in the same frequency band. During this project two main sets of measurements were orchestrated in order to examine the effects on LTE physical layer throughput, as could be seen in chapters 8 on page 33 and 9 on page 40. From the initial test in chapter 8 on page 33 it could overall be concluded that adding attenuation directly in to the channel had a real affect on how LTE adapts to poor channel quality. In figures 8.2 on page 35 and 8.6 on page 37, it could clearly be seen that when adding more attenuation into the channel, physical throughput fell, between every measurement run. Experiences from the initial measurement, were the basis on the following measurement, with the blocker circuit. During initial studies it was noticed that in the case stated in section 7.3 on page 29, is indeed real, and can be approximated in real-life on a deployed LTE network. This can directly be seen in figure 9.12 on page 46. As the attenuation of the blocking signal decreases, it can clearly be seen that the CQI index falls from what appears to be a 16-QAM modulation, and as the blocking signal becomes more distinct, CQI index drops to a level indicating QPSK or the like. E.g. a lesser efficient modulation. A better understanding on how to perform real-life measurements on deployed communication networks, mostly due to “learning-by-doing”-aspect of performing the measurements. However, it should be noted that to be able to actually find the phenomenon it was based on knowledge gained from

57

a prior internship at Agilent Technologies.

58

Chapter 11

Perspective The project started by focusing on other aspect than measuring intermodulation. This progress in learning by doing has led to some of the problems seeing in this project by different kind of measurement. There has been many initialize measurement leading to this project goal. Usually when making measurements a plan for how it should be done is produced first and then the measurement could be produced. In our case this was most of all not possible since the shorten of equipment. This lead to measurement performed with no goal. One of the main improvement to perform these kind of measurement with both more precision and reliable data is by selecting the level of attenuation on a smaller scale than the -10dB intervals. The measurement of the impact with modulation was only produced since the post processing was after that the equipment had to be delivered back three day after borrowing it. Since it was over a weekend some of the questions had to be resolved after the measurement and then it was to late. One of the interesting aspect in this case studies is that the intermodulated distorted signal did not react on the modulation but only in the level of attenuation. In further studies the smaller interval on attenuation in a intermodulated distorted signal interference could show some more accurate data collection. But at the same time the real life measurement is a dangerous area to perform measurement on since a wide range of variable has to be under control or needed to be suppress. All the measurement was performed sequential while the network performance varies in time and in the end all the measurement was placed above each other (look like parallel). This is one of the main risk which one must consider before stating the measurement. These uncertainties makes the measurement hard to reproduce and can only been as a case study. Laboratories measurement would eliminate these uncertainties and give a wider control to all of the variables that is not under control in a real life measurement. This project only focus on real life measurement since the experiences from Agilent Technologies internship. In the future the use of MIMO will increase since the performance of the

59

network only will be better. LTE supports an 4x4 MIMO connection and this will increase the throughput dramatically in urban environments since the best condition of multipath. The use of IFA antenna in the measurement is maybe a problem to represent a standard UE since the design of the IFA direct emphasize download frequencies in the 1800 MHz band. Normally this would not be the case and other antenna designs would be used to accommodate tx and rx antenna in the big range of 1700–1800 MHz band on one antenna. By using the IFA antenna only the downlink communication is preferred and dos not reflect an normally UE device. One of the interesting aspect in the intermodulation distortion in LTE on the upcoming transmission to the released 800 MHz band. In this band the television is sending its long range and with another power level than the LTE network dos. Perhaps some intermodulation distortion will occur if not certain requirement is met.

60

Bibliography 3GPP. E-UTRA UE radio transmission and reception. Technical specification Release 10.4.0, 3GPP, October 2011a. URL http://www. 3gpp.org/ftp/Specs/archive/36_series/36.101/36101-a40.zip. 3GPP. Physical Channels and Modulation. Technical specification Release 10.3.0, 3GPP, September 2011b. URL http://www.3gpp.org/ftp/ Specs/archive/36_series/36.211/36211-a30.zip. 3GPP. E-UTRA Physical layer procedure. Technical specification Release 10.3.0, 3GPP, September 2011c. URL http://www.3gpp.org/ftp/ Specs/archive/36_series/36.213/36213-a30.zip. Ian D. Flintoff, Martin P. Robinson, Stuart J. Porter, and Andrew C. Marvin. Addressing the Risk of EMC Problems with Mobile Radio Transmitters. Compliance Engineering, 2000. URL http://ce-mag.com/ archive/2000/sepoct/flintoft.html. Simon Haykin. Communication Systems. Wiley, 4th edition, May 2000. ISBN 0-471-17869-1. Ingmar Land and Bernard H. Fleury. Digital modulation 1, February 2007. URL http://kom.aau.dk/project/navcom/CourseWebSites/ DigitalModulation1/notes.pdf. Lecture note. IT og Telestyrelsen. Tilladelse til brug af frekvenser til oprettelse og drift af radioanlæg i DCS1800, December 2009a. URL http: //www.itst.dk/frekvenser-og-udstyr/frekvenstilladelser-mv/ tilladelser-udstedt-efter-auktion-eller-udbud/ 1800-mhz-frekvensbandet/filarkiv-tk7/TK7%201800-tilladelse% 20til%20Hi3G%20Denmark%20ApS.pdf. Article. IT og Telestyrelsen. Tilladelse til brug af frekvenser til oprettelse og drift af radioanlæg i DCS1800, December 2009b. URL http: //www.itst.dk/frekvenser-og-udstyr/frekvenstilladelser-mv/ tilladelser-udstedt-efter-auktion-eller-udbud/ 1800-mhz-frekvensbandet/filarkiv-tk4/Tilleg%20nr.%205%20til% 20TK4%201800%20frekvenstilladelse.pdf. Article. 61

IT og Telestyrelsen. Tilladelse til brug af frekvenser til oprettelse og drift af radioanlæg i DCS1800, December 2009c. URL http: //www.itst.dk/frekvenser-og-udstyr/frekvenstilladelser-mv/ tilladelser-udstedt-efter-auktion-eller-udbud/ 1800-mhz-frekvensbandet/filarkiv-tk3/Tilleg%20nr.%204%20til% 20TK3%201800%20frekvenstilladelse.pdf. Article. IT og Telestyrelsen. Tilladelse til brug af frekvenser til oprettelse og drift af radioanlæg i DCS1800, October 2010. URL http: //www.itst.dk/frekvenser-og-udstyr/frekvenstilladelser-mv/ tilladelser-udstedt-efter-auktion-eller-udbud/ 1800-mhz-frekvensbandet/filarkiv-tk1/Tilleg%20nr.%203%20til% 20TK1%201800%20frekvenstilladelse.pdf. Article. IT og Telestyrelsen. Det digitale samfund 2010, September 2011. URL http://www.itst.dk/statistik/publikationer/ det-digitale-samfund/2010/det-digitale-samfund-2010. Article. Moray Rumney, editor. LTE and the evolution to 4G: Design and test. Agilent Technologies, 2nd edition, May 2009. ISBN 978-988-17935-1-5. Jochen Schiller. Mobile Communications. Addison-Wesley Professional, 1st edition, January 2000. ISBN 0-201-39836-2. Alan Way. What is MIMO?, December 2009. URL http://www.youtube. com/watch?v=VLAgYUQCgD8. Sprirent Communications. Jim Zyren. Overview of the 3GPP Long Term Evolution Physical Layer, July 2007. URL http://www.freescale.com/files/wireless_comm/ doc/white_paper/3GPPEVOLUTIONWP.pdf. Freescale Semiconductor.

62

Part IV

Appendices

63

Appendix A

MIMO MIMO is one of the main advantages that LTE is using to increase/improve the data throughput and link range without using more bandwidth or by increasing the transmitted power. In figure A.1 the simplified MIMO configuration send’s the data in half of the time (Grey sends DA and Blue sends TA) by sending it parallel and making use of the both the antenna. In this case the throughput is being increased by a factor of two. To illustrate the increase of link range A.1 Tx0 and Tx1 both sends first DA and then both sends TA. This increase the range but decrease the throughput to the same speed as a singe antenna configuration.

Figure A.1: Simplified MIMO. Transmitter (Tx0,Tx1), Receiver (Rx0,Rx1) But MIMO is more complex than that. One of the main principle in MIMO is multipath see figure A.2 on the next page. Multipath change the property of the wave and is very location specific. Rx0 and Rx1 has different multipath property. At the transmitter side (Tx0 and Tx1) each of the antenna is operating at the same frequency but transmitting different streams of data. At the receiver side we.. are able of differentiate between the two streams and then recombine the data into a single serial stream. This is don by using Digital signal processors (DSP). But it is only possible when 64

the signal is arriving in different interval of time called “Signal diversity”. Signal diversity is caused by object like buildings, cars, train, people, natural obstacles and so on, and that the signal is not only taking a direct path between the transmitter and receiver. In this way of diversity of signal arrival makes it possible for the DSP to unscramble the signal (Way 2009)

Figure A.2: Multipath with signal diversity. Receiver (Rx0,Rx1) and obstacles (A,B,C,D)

65

Transmitter (Tx0,Tx1),

Appendix B

Duplexing and Multiplxing In short, duplexing is the method of communication between 2 devices, while multiplexing is the method of communication among multiple devices. In communication, duplexing is performed between two devices which are connected to eachother and communicates in both directions. Such a duplex system can be made in two ways; half-duplex or full-duplex. Half-duplexing is where each terminal alternates its turn to transmit, wheres a full-duplex system can both transmit and receive at the same time. Multiplexing can be performed in four different dimensions; space, time, frequency and code (Schiller 2000). When a number of independant signals are multiplexed, they are combined into one composite signal, which then is suitable for transmisison over a channel (Haykin 2000). However, in order to transmit these signals over the same channel, they must be kept apart, in such a way that they can be separated on the receiving end. This seperation can be done either in space, frequency, time and code as mentioned before. Space-Division Multiplexing (SDM) is the first of the four schemes of multiplexing. In this method, each channel medium is separated in space.FrequencyDivision Multiplexing (FDM) describe the scheme in which signals are subdivided into several frequency bands. Each channel is then allotted into its own frequency band, where senders can continuously use a certain frequency band (Schiller 2000). As opposed to allocating a portion of the whole bandwidth to each signal, Time-Division Multiplexing (TDM) is a more flexible multiplxeing scheme (Schiller 2000). In this scheme, the entire bandwidth is allocated for a certain amount of time. All senders use the same frequency however, they only use it at different points in time. Giving the entire bandwidth, enables the sender to transmit more data at a time, rather than being supplied with its own frequency. While TDM has its advantages, it also has its disadvantages. TDM requires a precise synchronization, in order to prevent interference among the senders, since two or more transmissions wound overlap (Schiller 2000).

66

Code-Division Multiplexing (CDM) is a relatively new scheme in communication, which inherents some security features. In CDM, tranmisison channels use the same frequency, coherent in time. The seperation of channels is done by assigning each channel its own “code” (Schiller 2000). This is an advantage, due to the fact that the code space is huge, in comparison to the frequency space (Schiller 2000). Different codes have to be assigned to each receiver, and therefore provides a form of security. On the other hand, CDM has the disadvantage of requiring a more complex receiver. It must know the code and be able to separate the channel with user data from background noise combined with other signals and environmental noise. In adiition to this, the receiver must be precisely synchronized with the transmitter, in order to perform the decoding correctly.

67

Appendix C

Channel Access Methods C.1

Basic Channel Access Methods

A channel access method is a method of accessing several terminals which are connected on the same physical ressource used for communication. There exists four fundamental methods to separate the different channels. They can be divided into either frequency, time, coding, and physical space, and are based on their corresponding multiplexing schemes. The first three methods are illustrated in figure C.1.

Figure C.1: Channel access using FDMA, TDMA, and CDMA in relation to each other. (Flintoff et al. 2000) These three access methods are based on their matching multiplexing schemes, FDM, TDM and CDM. FDMA is a channel access method which is based on FDM and provides different frequency bands to each data stream of the communication protocol. Time-Division Multiple Access (TDMA) is based on TDM and thus provides different timeslots to different data streams, in a repetative framestructure. For example, User A is allotted timeslot 1 and User B is given timeslot 2, etc. untill the final user has been reached. However, for TDMA to work efficiently it requires a precise synchronization. Otherwise transmissions may collide and result in corrupt data being transferred. 68

The last methods where channel access is separated in space, can be descibed as when each set of terminals are in close proximity, but at a distance from all other terminals.

C.2

OFDM

The scheme is an digital multi-carrier that use many closely-spaced subcarriers to transmit data. The main principle is to use numerous closely orthogonal sub-carrier signal to modulate the data. The modulation on each subcarrier is the conventional format like QAM, Binary Phase-Shift Keying (BPSK) or QPSK The use of many subcarrier performs similar to conventional single-carrier modulation. The OFDM is used from television to audio broadcasting and network systems. The way OFDM works is to use a square wave phase modulation represented by a sinc or sin(x) that x is convolved in the range of the subcarrier frequency. In figure C.2 the spectrum of multiple truncated modulated OFDM subcarriers with constant amplitude. The constant amplitude modulation is used in BPSK and QPSK.

Figure C.2: Multiple modulated OFDM subcarriers with constant amplitude. In contrast to the constant amplitude modulation of the multiple subcarriers LTE also supports 16-QAM and 64-QAM that enables the use of varying in amplitude. The standard for each modulated symbol in the LTE last 66.7 µs and by setting up the subcarrier spacing to 15 kHz. This enables that the peaks and null are of the subcarriers are orthogonal and therefore no interference between them (Rumney 2009). One of the main advantage in OFDM is the high data rates without increasing the symbol rates like in single carrier system. This simplifies the managing of the ISI. Eliminating the ISI is done by introducing the Cycle Prefix (CP). CP ensures that the symbol will be undistorted in its normal symbol length, when multipath. Unlike single carrier system the symbol length is set to the reciprocal of the channel bandwidth, OFDM’s symbols length is determined by the subcarrier spacing. One other aspect of dealing with multipath is to tightly space the subcarriers to effectively use of the bandwidth (Zyren 2007).

69

Appendix D

Modulation schemes This chapter is based on Schiller (2000), except where noted otherwise.

D.1

BPSK

BPSK is the simples form of Phase Shift Keying (PSK). The use of two separated (180◦ ). The BPSK modulation is the most robust form because it takes the highest level of noise. A drawback is of course that it is only able to modulate a 1 bit/symbol. In the figure D.1 the two phases of the modulation is representated.

Figure D.1: 1-bit signal

D.2

QPSK

QPSK encode two bit per symbols. The main advantage of using QPSK is that it double the data rate with the same bandwidth compared with BPSK. In another and very important way of using QPSK is to use the same data rate as BPSK, however only use the half the bandwidth while the Bit Error Ratio (BER) is exactly the same in both case. This make the QPSK an interesting modulation to wireless communication as BER is an very important factor. D.2 on the next page

70

Figure D.2: QPSK modulated signal, which comprises of 2-bit symbols

D.3

QAM

QAM is both an analog and a digital modulation scheme. In this section, only the digital modulation scheme is considered. In QAM signal are two carriers with same frequency but differs in phase with 90◦ . The first signal is called the I signal and the other is the Q signal. Mathematical representation is done by a sine wave and a cosine wave. In the end the two modulated carrier signal is composed at the source for transmitting. At the receiver side the carriers are separated and the data is then extracted from within, by combining the data into the original modulated information. There are several forms of QAM, however only two of them is used in the LTE. Figure D.3 shows both 16-QAM and 64-QAM. 16-QAM can transport four bits per symbol while 64-QAM is able to transport 6 bits of information per symbol. The drawback, of increasing the bit information, however, is that the modulations schemes becomes increasingly sensitive to noise (Rumney 2009).

Figure D.3: The two QAM schemes. 1) shows 16 QAM while 2) shows 64 QAM.

71

Appendix E

Attenuation measurement report E.1

Purpose

The purpose of this measurement is to examine effects on LTE Physical throughput, by introducing attenuation between antenna and USB LTE dongle. In some test the complete attenuation was 39 dB.

E.2

Requirements and equipement

Device

AAU-no.

Note.

Laptop 4G LTE USB dongle Attenuators – 3 dB – 6 dB – 10 dB – 20 dB 1800 MHz IFA antenna Downloads

N/A N/A

Externally borrowed. Externally borrowed.

01299-03 01048-00 00328-10 01340-00 N/A N/A

& & & &

Unknown 01339-18 01328-06 01340-01

Externally borrowed. See table E.2 on the next page

Table E.1: List of equipment required to perform the measurement. The downloads which are used during the test are listed below in table E.2 on the facing page. These downloads are chosen because it is track the packets way through the network and track ressources allocated from the network beyond the eNB.

72

Name

Size

Debian DVD Iso TheCamp.dk video fil

4.4 GB 1.9 GB

Filename

Server

URL

debian-6.0.3-i386-DVD-2.iso 00001.MTS

dotsrc.org dotsrc.org

Link Link

Table E.2: List of downloads used for the measurement.

E.3 E.3.1

Test setup Connectivity setup

Figure E.1: Measurement test setup for determining the impact of different attenuation while downloading, in the 1800 MHz band.

E.3.2

Location

The measurements were performed at,

E.4

Performing the measurement

In table E.3 on the next page, a summary of the test procedure is shown. Two control measurements were performed before and after, in order to determine the initial connection quality. These control measurements are used for a comparison of the affects the attenuators. The goal is to examine the effects in LTE, when channel quality is impaired. To reduce the channel quality, the connection will be attenuated inorder to attempt to mimic real-life degradation of channel quality. After the initial control measurement a 10 dB attenuator1 was connected 1

Previous attenuation measurements showed that there was no significant difference with only 3 or 6 dB attenuation.

73

Room 4-309 Niels Jernes Vej 12-14 9220 Aalborg Øst

Figure E.2: Location of UE and eNB. The measurement was performed approximately at position A), while the eNB is located approximately at position B). between the LTE dongle and the antenna. The attenuators were inserted between the dongle and the antenna coaxial cable. Each measurement was performed with increasing attenuation, starting from 10 dB to 39 dB. After the 39 dB attenuation the second control measurement was performed. Step

Action

1 2 3

Connect laptop and LTE device, and external antennas. Establish connection to the internet. Verify LTE 4G connection and speedtests are at their expected levels. Begin download. Perform control measurement. Connect attenuator. Restart download. Perform measurement. Go to step 6, when 39 dB is reached proceed to step 10. Perform second control measurement.

4 5 6 7 8 9 10

Table E.3: Measurement procedure for attenuation measurement.

E.5

Results 08-12-2011

In the following figures, the results from the first measurement run, are seen. All measurements are plotted with a smooth-factor2 =1001. Figure E.3 on the facing page contains Throughput, measured MCS index, instantaneous 2

Moving avarage function in MATLAB

74

RSRQ and assigned resource blocks. Figure E.4 on the next page contains SINR, instantaneous RSSI, and reported CQI.

Figure E.3: Throughput is measured in kbit/s.

75

Figure E.4

E.5.1

Summary

This first measurement run shows clearly how the physical throughput decreases between each measurement, meanwhile upholding a high amount of resource blocks. From the plots, it is evident that introducing attenuation into the system has the expected result of decreasing channel quality. With the rising attenuation the measured SINR falls as well, indicating that 76

the signal becomes more noisy. It is also seen that with the increasing attenuation CQI reports a lesser efficient modulation.

E.6

Results 11-12-2011

In this section are the results from the second measurement run. All measurements are plotted with a smooth-factor = 1001. Figure E.5 on the following page show Throughput, measured MCS index, instantaneous RSRQ, and assigned resource blocks, while figure E.6 on page 79 shows SINR, instantaneous RSSI, and reported CQI index.

77

Figure E.5: Throughput is measured in kbit/s.

78

Figure E.6

E.6.1

Summary

In contrast to the results from december 8th, is can be seen in figure E.5 on the facing page, that a better throughput is actually achieved by introducing 10 dB. However, as time progresses the throughput stabilises at approximately the same level, as those seen in figure E.3 on page 75. This can probably be explained due to settling time. 79

E.7

Conclusion

All in all, the results were more or less expected. The reason as to why 0 dB, 10 dB and 20 dB measurements are so similar is that the connection was so “high up” on the BLER curve, thus resulting in no significant difference. In order to see some distinction it is neccessary to force the “quality” further down the BLER curve. It can be seen that LTE attempts to uphold a high throughput, by trying to find a proper modulation and coding scheme, as seen in the MCS index. The CQI reports, show that with a low attenuation a “good” modulation is chosen by the eNB inorder to uphold a 10 % BLER. As the attenuation rises, the eNB, lowers the modulation inorder to counteract errors in the datastream.

80

Appendix F

Blocker measurement report F.1

Purpose

The purpose of this measurement journal is to examine effects on LTE Physical throughput, by introducing intermodulation distortion signal between antenna and USB LTE dongle. To combine the interference signal with the antenna and the dongle an coupler of -20dB is used. Due only one coupler could be borrowed all the measurement were performed in SISO mode. This decrease the throughput significantly.

F.2

Requirements and equipement

Device

AAU-no.

Note.

Laptop 4G LTE USB dongle SMIQ 06B Coupler 1800 MHz IFA antenna Downloads

N/A N/A

Externally borrowed. Externally borrowed.

N/A N/A N/A

Externally borrowed. Externally borrowed. See table F.2 on the following page

Table F.1: List of equipment required to perform the measurement. The downloads which are used during the test are listed below in table F.2 on the next page. These downloads are chosen because it is track the packets way through the network and track ressources allocated from the network beyond the eNB.

81

Name

Size

Debian DVD Iso

4.4 GB

Debian DVD Iso

4.4 GB

Debian DVD Iso

4.4 GB

Debian DVD Iso

4.4 GB

Debian DVD Iso

4.4 GB

PES 2012 Demo CABAL Online Episode VII: Saint’s Requiem Client

1.24 GB 1.14 GB

Filename

Server

debian-6.0.3-amd64debian.org DVD-1.iso debian-6.0.3-amd64debian.org DVD-2.iso debian-6.0.3-amd64debian.org DVD-3.iso debian-6.0.3-amd64debian.org DVD-4.iso debian-6.0.3-amd64debian.org DVD-5.iso PES2012DEMO.zip gamershell.com gamershell.com CABAL_Online_ Saint_s_Requiem_Full_ Client.zip

Table F.2: List of downloads used for the measurement.

F.3 F.3.1

Test setup Connectivity setup

Figure F.1: Test setup for measurement the impact of different attenuator size while downloading in the 1800 MHz band

F.3.2

Location

The measurements were performed at, 82

URL Link Link Link Link Link Link Link

Room 4-309 Niels Jernes Vej 12-14 9220 Aalborg Øst

Figure F.2: Location of UE and eNB. The measurement was performed approximately at position A), while the eNB is located approximately at position B).

F.4

Performing the measurement

In table F.3 on the following page below, a summary of the test procedure is shown. There were three different measurement performed on the 1301-2012, 14-01-2012 and 15-01-2012. The first day (13-01-2012) a variety of different measurement was performed. First a initiating test on the bandwidth ranging from 1863.2 MHz to 1873.2 MHz to see in with frequency the intermodulated signal had the biggest impact. Two frequencies was selected, 1869.7 MHz and 1870.2 MHz. These frequencies where measured with different attenuation. On the 14-01-2012 the network was affected by an intermodulated signal beginning from the center frequency 1868.2 MHz and moved up to 1875.2 MHz. All this was don with different modulation on the intermodulated signal from the SMIQ 06B (BPSK, QPSK, 16QAM and 64QAM). The level of attenuation from intermodulated signal was 40dB. The last measurement on 15-01-2012 the whole bandwidth was used to perform almost the same measurement as 14-01-2012. The interval of MHz/minute was reduced to the half of time. This enables the analysis of the whole bandwidth. Two set of measurement was performed.

83

Step

Action

1 2 3 4 5 6 7 8 9 10

Connect laptop and LTE device. External antenna is connected with the coupler and the LTE device. SMIQ 06Bsignal generator is connected to the -20dB on the coupler. Establish connection to the internet. Verify LTE 4G connection and speedtests are at their expected levels. Begin download. Perform control measurement. Start SMIQ 06B and select the given modulation and level of attenuation Restart download. Perform measurement. Table F.3: Measurement procedure for SMIQ measurement.

F.5

Results from the 13-01-2012

In the following figures below, results of the different measurements are seen. All measurements are plotted with a smooth-factor1 =1001. Figure F.3 on the next page and F.4 on page 86 displays; Troughput, RSRP, RB, SINR, MCS index,RSSI and CQI in 1869.7 MHz band with a interference signal that had different attenuation ranging from -80dB, -70dB, -60dB, -50dB and -40dB with one minute in between(the couplers -20dB is included). In figure F.5 on page 87 and F.6 on page 88 the same procedure as by the first two. The only difference is that the frequency is 0.5MHz above 1870.2 MHz. 1

Moving avarage function in MATLAB

84

Figure F.3: TP in kbit/s, RSRQ and RB with different level of attenuation

85

Figure F.4: SINR in dB, MCS index, RSSI in dB and CQI with different level of attenuation

86

Figure F.5: TP in kbit/s, RSRQ and RB with different level of attenuation

87

Figure F.6: SINR in dB, MCS index, RSSI in dB and CQI with different level of attenuation

F.5.1

Conclusion

The attenuation level regulated by the SMIQ 06B influence the performance of the throughput dramatically. In almost every KPI the increased of the attenuation from the intermodulated signal makes an impact. However, only

88

RB remains stable at approximately 49.

F.6

Results from the 14-01-2012

In this measurement setup the goal was to see if the modulation of the intermodulated signal had any impact on the performance on the network. All measurements are plotted with a smooth-factor2 =1001. The network frequency had a 10 MHz bandwidth ranging from 1863.7 MHz to 1873.7 MHz and with a center frequency of 1868.2 MHz. In this measurement the start frequency of the measured bandwidth was from the center frequency 1868.2 and with every one minute increased the frequency of the interference signal see F.7. In figure F.8 on the following page the throughput, RSRP and RB is displayed. The next figure F.9 on page 91 displays SINR, MCS, RSSI and CQI.

Figure F.7: Every 1 minute the frequency was increased by 1 MHz. The extra time was to insure that the interference signal was out of range. 2

Moving avarage function in MATLAB

89

Figure F.8: TP in kbit/s, RSRQ and RB with different frequencies

90

Figure F.9: SINR in dB, MCS index, RSSI in dB and CQI with different frequencies

F.6.1

Conclusion

In this measurement we see that the throughput between the QAM(16 and 64) and BPSK and QPSK differs from each other. 16QAM and 64QAM has a higher throughput and RSRP level of -13 to -14. But in the other 91

KPI’s the difference between them is not visible. In the CQI index BPSK and QPSK follows each other all the way and both the QAM(16 and 64) differ over time. From the RSSI measurement the modulation had no impact on the total received power. It’s clear that the start signal power level is around -60dB and by increasing the frequency on the intermodulated signal (and thereby crossing Telia’s 10 MHz bandwidth) the -60dB–40dB gives an increase of the total power by 20db above the -60dB start total power received.

F.7

Results from the 15-01-2012

The measurement that was perform in this section was over the whole 10 MHz band ranging from 1863.2 MHz to 1873.2 MHz. All measurements are plotted with a smooth-factor3 =1001. Start frequency was 1860.7 since the center frequency of the interference signal from the SMIQ 06B was 2.5 MHz. Every 30 sec. the frequency was increased by 1 MHz until the interference signal was out of the 10 MHz bandwidth to 1876.7 MHz. An illustration on how this was don see figure F.10. In figure F.11 on the next page and F.12 on page 94 an -40dB interference signal is used to block the frequency area. On the two last figures F.13 on page 95 and F.14 on page 96 the level of attenuation on the intermodulated signal was increased to -60dB.

Figure F.10: Every 30 sec. the frequency was increased by 1 MHz. The bandwidth was 10 MHz and the extra time was to insure that the interference signal was out of range. 3

Moving avarage function in MATLAB

92

Figure F.11: TP in kbit/s, RSRQ and RB

93

Figure F.12: SINR in dB, MCS index, RSSI in dB and CQI

94

Figure F.13: TP in kbit/s, RSRQ and RB

95

Figure F.14: SINR in dB, MCS index, RSSI in dB and CQI

F.7.1

Conclusion

The difference between the modulations impact on the performance on the LTE network was not as significant than first expected. From figure F.11 on page 93 the throughput range of the different measurement was between 0.5kbit/s. But as the intermodulated signal was moved through the Telia’s 96

10 MHz bandwidth first the throughput decreased and at the end of the 10 MHz the throughput increased again. To verify this the RSSI start at -60dB and moved slowly up to -42dB. This indicated that the total received power has increased by +20dB (-60dB–40dB). On the second recorded measurement with -60dB intermodulated signal the SINR only dips under 0dB in contras from the first measurement were the SINR get close to 20dB. The RSSI moves from -62dB and up to -59dB. This comes that the -64–60 gives a 4dB increase of the total power. Overall the modulation from the intermodulated signal had no impact on the performance on the LTE network. RB from the last measurement had some strange problems to connecting to the eNB. This had only been seen this time and therefore must be something from the server.

97

98