pw benali imcne06

Master Report: WCDMA RF Optimization

Sadok Ben Ali 1

Contents

Introduction…………………………………………………………………………………….……1

Chapter I: RF Optimization Introduction

� �� ............................................................................................ 4 �� .................................................................................................. 4 �� ....................................................................................... 5

�� ................................................................................................................ 5 �� .................................................................................................................... 6 �� !��"��................................................................................................................. 6 � ��......................................................................................................................... 6 �� #��$��...................................................................................................... 6 �� %��#��&��...................................................................................................................... 9 �� #�� &��............................................................................................................... 10 ��' ��#��........................................................................................................11

��'�� #��............................................................................................................11 ��'�� #�(��...........................................................................................................11 ��'�� #��...................................................................................................................11 � ��........................................................................................................................... 12 �� %��............................................................................................................................... 12

�� "��........................................................................................................................... 12 �� #��)��*��........................................................................................................... 13 �� )��............................................................................................................................. 13 �� #� +��#&��........................................................................................ 13�

Chapter II: Network Problem analyzing

�

� �%��#�� !��"��........................................................................................................ 15 �� %��#�� "��........................................................................................................ 15

�� ,��-��%��#�.................................................................................................................. 15 �� .��%��#�........................................................................................................... 16 �� /� ��/��-��-..................................................................................... 16 ��' +��"��................................................................................................................ 17 �� %��#��!��"��................................................................................................. 17

�� -��%��#��!��"��............................................................................................. 17 �� /��-��%��#��!��"��.................................................................................................. 20


Sadok Ben Ali 2

�� %��#�� ........................................................................................................ 21 �� ,��-��%��#��&��)��0�#��#��.................................... 21 �� .��%��#��&��)��+��1��..................................................... 22 �� %��#�� &��)��)��!��...................................... 23 � ��&�� !��"��................................................................................................. 24 �� &��2&�#��................................................................ 24

�� .......................................................................................................................... 24 �� 2&�#��......................................................................................................... 24 �� &��)��&��!��"��............................................................................................... 24

�� &��!��"��................................................................................................................. 24 �� )��&��!��"��............................................................................................................. 25 �� &��&��..................................................................................................... 26

�� !��!�3&��......................................................................................................... 26 �� )�(��!�3&��.................................................................................................... 27 �� /��#� /��4��.................................................................................................. 28 ��' ��!��"��#��&�� ........................................................................ 29

� (��%�� !��"��....................................................................................................... 29 �� +��#� ��.................................................................................................... 30

�� !��"��............................................................................................................... 30 �� %��#� ��&��+��#� ��................................................................................ 34 �� (�� !��"��.......................................................................................... 34

�� (�� .............................................................................. 34 �� $��4��............................................................... 34 Chapter III: Work Process�

�

� )�� .......................................................................................................................... 36 � 5�)�!��"��............................................................................................................................... 37 �� ...................................................................................................................................... 37 �� 0�6+�!��"��7...................................................................................................................... 39 �� /0��*��............................................................................................................ 40 ��' �10 ....................................................................................................................................... 42 ��8 0%��.................................................................................................................................... 45 ��9 �� &��....................................................................................................................... 46 ��: �$�;�<��............................................................................................................ 47 ��= ��&��...................................................................................................................... 48

Conclusion…………………………………………………………………………………….……50


Sadok Ben Ali 3

Introduction Analog cellular systems are commonly referred to as first generation systems. The digital systems currently in use, such as GSM, PDC, cdmaOne (IS-95) and US-TDMA (IS-136), are second generation systems. These systems have enabled voice communications to go wireless in many of the leading markets, and customers are increasingly finding value also in other services, such as text messaging and access to data networks, which are starting to grow rapidly. Third generation systems are designed for multimedia communication: with them personto-person communication can be enhanced with high quality images and video, and access to information and services on public and private networks will be enhanced by the higher data rates and new flexible communication capabilities of third generation systems. This, together with the continuing evolution of the second generation systems, will create new business opportunities not only for manufacturers and operators, but also for the providers of content and applications using these networks. In the standardization forums, WCDMA technology has emerged as the most widely adopted third generation air interface. Its specification has been created in 3GPP (the 3rd Generation Partnership Project), which is the joint standardization project of the standardization bodies from Europe, Japan, Korea, the USA and China. Within 3GPP, WCDMA is called UTRA (Universal Terrestrial Radio Access) FDD (Frequency Division Duplex) and TDD (Time Division Duplex), the name WCDMA being used to cover both FDD and TDD operation. Throughout this book, the chapters related to specifications use the 3GPP terms UTRA FDD and TDD, the others using the term WCDMA. This book focuses on the WCDMA FDD technology. The WCDMA TDD mode and its differences from the WCDMA FDD mode are presented in Chapter 13, which includes a description of TD-SCDMA.

In this report we will try to present a manual for UMTS data collection and for optimization also solutions for analyzing and the way for resolving problems in RF part of the WCDMA Network. Many cases of this manual are token from the UAE Etisalat Network.


Sadok Ben Ali 4

Chapter I: RF Optimization Introduction

1 Basic Processes for RF Optimization

Once all the sites are installed and verification is complete, RF optimization starts. In some situations for a tight schedule, RF optimization might start after the construction of partial sites is complete. RF optimization is usually performed after 80% of total sites in a cluster are constructed. RF optimization stage is one major stage of RNO. It aims at the following aspects:

� Optimizing signal coverage � Control pilot pollution � Control SHO Factor based on DT

RF optimization also involves optimizing list of neighbor cells.

When the indexes like DT and traffic measurement after RF adjustment meets KPI requirements, RF optimization stage ends. Otherwise you must reanalyze data and adjust parameters repeatedly until all KPI requirements are met. After RF optimization, RNO comes to parameter optimization stage.

1.1 Flow Chat of RF Optimization

RF optimization includes the following four parts:

� Test preparations � Data collection � Problem analysis � Parameter adjustment

0 shows the RF optimization flow chat.


Sadok Ben Ali 5

Figure 1.1: RF optimization flow chat

In 01, the data collection, problem analysis, and parameter adjustment might be repeatedly performed according to optimization goal and actual on-site situations until RF indexes meet KPI requirements.

1.2 Detailed Sections of RF Optimization

1.2.1 Test Preparations

During test preparations, proceed as below: � Decide KPI goals for optimization according to the contract � Divide clusters properly and decide test route with the operator � The KPI test acceptance route is especially important. � Prepare tools and materials for RF optimization


Sadok Ben Ali 6

� This ensures smooth RF optimization.

1.2.2 Data Collection

Collect the following data:

� UE and scanner data � DT � Indoor test � Signaling tracing

� Call data tracing at RNC side � Configuration data

The configuration data and the call data tracing help to locate problems. Data collection is a precondition for problem analysis.

1.2.3 Problem Analysis

You can locate problems by analyzing collected data. After analyzing coverage problems, pilot pollution problems, and handover problems, provide corresponding adjustment solutions. After adjustment, test the adjustment result. If the test result cannot meet KPI requirements, reanalyze problems and readjust parameters until all KPI requirements are met.

Due to weak coverage, pilot pollution, and missing neighbor cells, the following problems are related to location:

� Downlink interference � Access problems � Call drop problems

The previous problems occur regularly. You can solve them by repeated optimization.

If the coverage is good, pilot pollution and missing neighbor cells are not present, the access and call drop problems need to be solved during parameter optimization stage. You can refer to corresponding guidebooks. The period for solving uplink interference problems (RTWP is over high but no high traffic matches it) is long, even as long as the RF optimization ends. Output an updated list of engineering parameters and list of cell parameters after RF optimization. The list of engineering parameters reflects adjustment of engineering parameters (such as down tilt and azimuth) during RF optimization stage. The list of cell parameters reflects the adjustment of cell parameters (such as neighbor cell configuration) during RF optimization stage.

2 Test Preparations

Test preparations include the following four aspects: � Deciding optimization goal � Dividing clusters � Deciding DT route � Preparing tools and data

2.1 Deciding Optimization Goal

The key of RF optimization is to solve problems as below: � Weak coverage � Pilot pollution � High SHO Factor based on DT


Sadok Ben Ali 7

Actually, different operators might have different standards on KPI requirements, index definition, and attention. Therefore the RF optimization goal is to meet the coverage and handover KPI requirements in the contract (commercial deployment offices) or planning report (trial offices). Define the indexes as required by contract as below:

The index definition is the percentage ratio of the sampling points with the index (such as CPICH Ec/Io) greater than the reference value in all sampling points.

Usually after RF optimization, the network must meets the index requirements listed in 0. 0 lists the RF optimization objectives according to analysis of and suggestion to coverage by existing network.

Index Reference Remarks

� 97% in urban area

CPICH Ec/Io � –9dB � 97% in suburban area

According to test result from the scanner, in unloaded and outdoor conditions, in planning coverage areas, test in a grid-like route to cover all cells.

� 98% in urban area

CPICH RSCP � –95dBm � 95% in suburban area

According to test result from the scanner, in unloaded and outdoor conditions, in planning coverage areas, test in a grid-like route to cover all cells. The coverage level request is basic. If operators have penetration loss request, add the penetration loss to the coverage level.

SHO Factor based on DT 30%–40%

The SHO Factor based on DT should be 5% to 10% lower than the goal, because the following optimizations cause the soft handover factor to increase

Pilot pollution ratio � 5% –

Table 1-1: List of RF optimization objectives in R99 networks

The RF optimization of HSDPA services aims to improve the distribution of UE CQI.

According to theoretical analysis, the CQI reported by UE and PCPICH Ec/Nt have relationship as below:

CQIUE = Ec/NtPCPICH + MPO + 10log16 + 4.5dB Wherein,

� Nt = (1- a) * Ior + Ioc + No � a is the orthogonal factor � lor is the signals of serving cell � loc is the interference signals from neighbor cells � No is the thermal noise � Io = Ior + Ioc + No

Therefore, PCPICH Ec/Nt is approximately equal to PCPICH Ec/Io.

MPO = Min (13,CellMaxPower –PcpichPower – MPOConstant)


Sadok Ben Ali 8

The maximum transmit power of a cell is usually 43 dBm, and the pilot channel power is 33 dBm. When MPOConstant is 2.5 dB, the default configuration by RNC, the MPO is 7.5 dB.

The 4.5 dB is obtained according to the linear relationship between the SNR of all the subscriber's HS-PDSCHs and the corresponding CQIs. Namely, SNR = –4.5dB + CQIUE, and SNR = Ec/NtHS-PDSCH + 10log16. When calculating CQIUE at UE side, the UE assumes that the total transmit power of HS-PDSCH is PHS-DSCH = PPCPICH + MPO. Wherein, PPCPICH is the transmit power of PCPICH. Therefore, Ec/NtHS-PDSCH = Ec/NtPCPICH + MPO. As a result, the CQI reported by UE is as below: CQIUE = Ec/NtPCPICH + MPO + 10log16 + 4.5dB According to previous analysis, the offset between CQIUE and PCPICH Ec/Io is 24 dB. Therefore, in terms of actual optimization, to optimize CQI is to optimize Ec/Io. Assume that the cell power is dynamically distributed between R99 and HSDPA networks. After receiving CQIUE from UE, the NodeB adjust the CQI as below: The CQI adjusted by NodeB, CQINodeB = ( Pcell - Pcommon – PR99 – PHS-SCCH – (PPCPICH + MPO ) + CQIUE.

Wherein,

� Pcell is the maximum transmit power of cell � Pcommon is the CCH power of cell � PR99 is the power of downlink associated DPCH for R99 or HSDPA subscribers. � PHS-SCCH is the HS-SCCH power. Assume:

� Pcell = 43 dBm � Pcommon is 20% of total power of cell � No R99 subscribers are in the cell � PR99 is too low to neglect � PHS-SCCH is 5% of total power of cell

Therefore,

CQINodeB = 1 + CQIUE

According to experience in actual test, based on the difference between the Ec/Io from scanner and the Ec/Io from UE, reserve a margin of 1 dB. At the edge of cell, an HSDPA subscriber may occupy total power of cell, so the throughput rate at cell edge is equivalent to the throughput rate at cell edge for the single subscriber. Errore. L'origine riferimento non è stata trovata. lists the relationship among the CQI reported by UE, pilot Ec/Io, and throughput rate at MAC-HS layer (MPO = 7.5 dB).

9 > CQI 15 > CQI � 9 CQI � 15

Subscribers' feeling Poor Fair Good

throughput rate at MAC-HS layer for single subscriber

0–320 kpbs 320 kbps to1.39 Mbps > 1.39 Mbps

Ec/Io > –15dB –15dB to –9dB � – 9dB

Table 1-2: Relationship among the CQI reported by UE, pilot Ec/Io, and throughput rate at MAC-HS layer

The throughput rate provided in Errore. L'origine riferimento non è stata trovata. is based on the test in the following conditions:


Sadok Ben Ali 9

� The codes and lub are not restricted. � The category 12 UE has a subscribed rate of 2 Mpbs. � The subscribed type is background or interactive service � Power is dynamically distributed. Namely, without R99 subscribers, all the power is

used by the HSDPA subscriber to guarantee rate as high as possible. � According to the requirements on RF optimization of unloaded R99 network, the � CPICH Ec/Io � –9 dB. After HSDPA is introduced, power is dynamically distributed,

and the single HSDPA subscriber at cell edge uses all the power. Meanwhile, the downlink load reaches 90%, and CPICH Ec/Io � 15.5dB.

If operators' requirement on throughput rate at cell edge is not the recommended values as listed in Errore. L'origine riferimento non è stata trovata., search the required value in 0. 0 lists the mapping relationship of HSDPA Catogory12 UE CQI and TB size. The CQIs that is larger than 13 or smaller than 5 are excluded. The rate at MAC-HS layer for the subscriber is (TBsize / 2ms) * (1 – BLER), wherein, the BLER is 10%.

CQI TB Size

5 365

6 365

7 365

8 711

9 711

10 1055

11 1405

12 1742

13 2083

Table 1-3: Mapping relationship of HSDPA Catogory12 UE CQI and TB size

As previously mentioned, to optimize HSDPA is to optimize Ec/Io of target networks. Therefore, in terms of optimization method, the HSDPA and R99 networks are consistent. The following optimization flow will not distinguish HSDPA networks from R99 networks.

2.2 Dividing Clusters

According to the features of UMTS technologies, the coverage and capacity are interactional and the frequency reuse factor is 1. Therefore RF optimization must be performed on a group of or a cluster of NodeBs at the same time instead of performing RF optimization on single site one by one. This ensures that interference from intra-frequency neighbor cells are considered during optimization. Analyze the impact of the adjustment of an index on other sites before adjustment. Dividing clusters involves approval by the operator. The following factors must be considered upon dividing clusters:

� According to experiences, the number of NodeBs in a cluster depends on the actual situation. 15–25 NodeBs in a cluster is recommended. Too many or few NodeBs in a cluster is improper. � A cluster must not cover different areas of test (planning) full coverage services.


Sadok Ben Ali 10

� Refer to the divided clusters for network project maintenance of the operator. � Landform factor : Landforms affect signal propagation. Mountains block signal propagation, so they are natural borders for dividing clusters. Rivers causes a longer propagation distance, so they affect dividing clusters in various aspects. If a river is narrow, the signals along two banks will interact. If the transportation between two banks allows, divide sites along the two banks in the same cluster. If a river is wide, the upstream and downstream will interact. In this situation, the transportation between two banks is inconvenient, dividing clusters by the bank according to actual situation. � A cell-like cluster is much usual than a strip-like cluster. � Administrative areas When the coverage area involves several administrative areas, divide clusters according to administrative areas. This is easily acceptable by the operator. � DT workload: The DT must be performed within a day for a cluster. A DT takes about four hours.

Errore. L'origine riferimento non è stata trovata.2: shows divided clusters in a project.

2.3 Deciding Test Route

Confirm the KPI DT acceptance route with the operator before DT. If the operator already has a decided DT acceptance route, you must consider this upon deciding the KPI DT acceptance route. If the objective factors like network layout cannot fully meet the coverage requirements of decided test route by the operator, you must point this out. The KPI DT acceptance route is the core route of RF optimization test routes. Its optimization is the core of RF optimization. The following tasks, such as parameter optimization and acceptance, are based on KPI DT acceptance route. The KPI DT acceptance route must cover major streets, important location, VIP, and VIC. The DT route should cover all cells as possible. The initial test and final test must cover all cells. If time is enough, cover all streets in the planned area. Use the same DT route in every test to compare performances more accurately. Round-trip DT is performed if possible. Consider actual factors like lanes and left-turn restriction while deciding test route. Before negotiating with the operator, communicate these factors with local drivers for whether the route is acceptable.


Sadok Ben Ali 11

2.4 Preparing Tools and Data

Prepare necessary software (listed in 01), hardware (listed in 02), because the following test and analysis are based on them.

2.4.1 Preparing Software

01 lists the recommended software for RF optimization

No. Software Function

1 Tems investigation Data collection Data collection

2 Tems Rout Analysis, Genex Assistant

Analyzing DT data and checking neighbor cells

3 Actix Analyzing performance, Network KPIs

4 Mapinfo Displaying maps and generating route data

01: Recommended software for RF optimization

2.4.2 Preparing Hardware

02 lists the recommended hardware for RF optimization

No. Device Specification Remarks

1 Scanner DTI Scanner –

2 Test terminal and data line Sonny Ericsson K800i

At least two test terminals. If there is HSDPA request, use the data card E620. U626 does not support HSDPA.

3 Laptop PM1.3G/512M/20G/USB/COM/PRN –

4 Vehicle mounted inverter DC to AC, over 300W –

02: Recommended hardware for RF optimization

2.4.3 Preparing Data

03 lists the data to be collected before optimization


Sadok Ben Ali 12

No. Needed data Whether is necessary Remarks

1 List of engineering parameters Yes –

2 Map Yes By Mapinfo or in paper

3 KPI requirements Yes –

4 Network configuration parameters Yes –

5 Survey report No –

6 Single site verification checklist No –

7 Floor plan of the target buildings Yes For indoor test

03 : Data to be collected before optimization

3 Data Collection

During RF optimization stage, the key is the optimization of radio signals distribution, with the major means of DT and indoor test. Before test, confirm with the customer care engineers the following aspects:

� Whether the target NodeBs, RNCs, and related CN are abnormal due to being disabled, blocked, congested, and transmission alarms. � Whether the alarms have negative impact on the validity of test result data. If the alarms exist, solve the problems before test. DT is a major test. Collect scanner and UE data of radio signals by DT test. The data is applicable in analyzing coverage, handover, and pilot pollution problems. Indoor test involves the following areas: � Indoor coverage areas Indoor coverage areas include inside buildings, department stores, and subways. � Inside areas of important facilities Inside areas of important facilities include gymnasiums and government offices. � Areas required by the operator :Areas required by the operator include VIC and VIP.

Test the previous areas to locate, analyze, and solve the RF problems.

Indoor test also involves in optimizing handover of indoor and outdoor intra-frequency, inter-frequency, and inter-system. The DT and indoor test during RF optimization stage is based on VP service. According to the contract (commercial deployment offices) and planning report (trial offices), if seamless coverage by VP service is impossible in areas, such as, suburban areas and rural areas, the test is based on voice services. For areas with seamless coverage by PS384K service or HSDPA service required by the contract (commercial deployment office) or planning report (trial office), such as office buildings, press centers, and hot spot areas, the test is based on the above services.

3.1 Drive Test

3.1.1 DT Types

According to different full coverage services in the planned areas, DT might be one of the following:


Sadok Ben Ali 13

� 3G ONLY continuous call test by using scanner + unloaded VP According to simulation result and experiences, if the test result meets requirements on VP service coverage, the test result will also meet identical coverage requirements of PS144K, PS128K, and PS64K services. � 3G ONLY continuous call test by using scanner + unloaded voice service � 3G ONLY continuous call test by using scanner + unloaded PS384K � 3G ONLY continuous call test by using scanner + unloaded HSDPA

3.1.2 Setting DT Indexes

The following paragraphs take VP service for example.

Setting DT

Index Meaning

Enable Whether to implement this index. True for implementation. False for non-implementation. The recommended value is True.

Call Number

Called number. Whether the called terminal supports VP must be confirmed.

Setup Time (s)

The maximum time for setting up calls. It ranges from 20–30s. The recommended value is 25s.

Calling Time (s)

The time for a single call from call start to normal end of call. Set it great enough according to actual DT route. The recommended value is 99999s.

Idle Time (s) Call internal time. The recommended value is 10s.

Call Count Total call times. Set it great enough according to actual DT route. The recommended value is 999 times.

01: For setting DT, see the following table.

Collect call data tracing at RNC side while performing drive test. This help to locate and analyze problems.

3.2 Indoor Test

GPS signals are unobtainable in door test. Obtain the plan of the target area before test. Indoor test consists of walking test and vertical test.

Indoor test services are services by seamless coverage required in the contract (commercial deployment office) or planning report (trial office). The method for indoor test and requirements on collecting call tracing data are the same as DT.

3.3 Collecting RNC Configuration Data

During RF optimization stage, collect neighbor cell data of network optimization and other data configured in RNC database. In addition, check whether the configured data is consistent with the previously checked/planned data.


Sadok Ben Ali 14

While checking configured data, feedback the improperly configured data (if found) to product support engineers. During checking, pay special attention to handover reselection parameters and power setting parameters, as listed in 02.

Configured parameters to be checked

Type Content to be checked

Handover reselection parameter

IntraFreqNCell (intra-frequency neighbor cell)

InterFreqNCell (inter-frequency neighbor cell) InterRATNCell (inter-system neighbor cell)

Power configuration parameter

MaxAllowedULTxPower (maximum uplink transmit power of UE)

PCPICHPower (PCPICH transmit power)

HSDPA cell configuration

Whether the HSDPA cell is activated

HS-PDSCH code configuration HS-SCCH configuration

HS-PDSCH and HS-SCCH power configuration

02: RNC parameters


Sadok Ben Ali 15

Chapter 2: Network Problem analyzing

1 Coverage Problem Analysis

Coverage problem analysis is key to RF optimization. It involves signal distribution. The coverage problems to be analyzed include:

� Weak coverage � Cross-cell coverage � Unbalance uplink and downlink � No primary pilot cell

1.1 Coverage Problem Types

1.1.1 Weak coverage

��

Weak coverage refer to that the RSCP of pilot signals in a coverage area is smaller than –95 dBm. It might be in:

� Valley areas � Hillside back � Elevator well � Tunnel � Underground garage � Basement � Areas inside high buildings

If the pilot signals are weaker than that required by full coverage services (such as VP and PS64K), or just meet the requirements, if the PICH Ec/Io cannot meets the lowest requirements on full coverage services due to increased intra-frequency interference, problems like difficult access of full coverage services will occur.

If the RSCP of pilot signals is weaker than that of minimum access threshold in a coverage area, the UE cannot camp on the cell, so the UE drops off the network due to failing in location updating and location registration.

��

For previous problems, use the following methods:

� Increase pilot transmit power, adjust antenna down tilt and azimuth, increase antenna height, use antennas with higher gain to optimize coverage. � If subscribers are abundant in the non-overlapped areas of neighbor NodeBs or the non-overlapped areas are great, construct new NodeBs or expand the coverage range of neighbor NodeBs. This ensures a software handover area with enough great size. Pay attention to that increasing of coverage areas might cause intra-frequency and inter-frequency interference. � Construct new NodeBs or add RRU in valley and hillside back areas with weak coverage to expand coverage range. � Use RRU, indoor distributed system, leakage cable, and directional antenna to solve problems in signal dead zone like elevator well, tunnel, underground garage, basement, areas inside buildings.


Sadok Ben Ali 16

1.1.2 Cross-cell Coverage

��

Cross-cell coverage refers to that the coverage range of some NodeBs is beyond the planned range and discontinuous primary pilot coverage areas form in coverage areas of other NodeBs. For example, if the NodeBs with a height much higher than the average height of adjacent buildings transmit signals along upland or roads over far, a primary pilot coverage area form in the coverage area of other NodeBs, an "island" forms. Therefore, if a call accesses the "island" and the nearby cells of the "island" is not configured as the neighbor cells, call drops once the UE leaves the island. Though the nearby cells of the "island" is configured as the neighbor cells, the "island" is over small, call also drops due to delayed handover.

If the two-side areas along a gulf are improperly planned, cross-cell coverage occurs on these areas due to short distance between two sides of the gulf. Consequently, interference occurs.

��

For the previous problems, use the following methods: � For cross-cell coverage, prevent antennas from transmitting signals straightforward along roads or reduce cross-cell coverage areas by using sheltering effect of adjacent buildings. Meanwhile you must avoid intra-frequency interference to other NodeBs. � For over high NodeBs, change the site. You might have difficulties in finding new sites due to property and equipment installation. In addition, too large mechanism down tilt causes aberration of antenna direction maps. Therefore you can eliminate the "island" effect and reduce NodeB coverage areas by adjusting pilot transmit power and using electric down tilt.

1.1.3 Unbalanced Uplink and Downlink

��

Unbalanced uplink and downlink refers to the following situations in uplink and downlink symmetric services:

� The downlink coverage is good but the uplink coverage is restricted. More specific, the UE transmit power reaches the maximum which still cannot meet uplink BLER requirements. � The downlink coverage is restricted. More specific, the downlink DCH transmit power reaches the maximum which still cannot meet downlink BLER requirements.

If the uplink and downlink are unbalanced, call drops easily. The probable cause is restricted uplink coverage.

��

For the unbalanced uplink and downlink problems, check for interference by monitoring RTWP alarms of NodeB. Other causes may lead to unbalanced uplink and downlink, such as:

� Uplink and downlink gain of repeaters and interference amplifier are faulty. � In an Rx/Tx detach system, the Rx diversity antenna-feeder system is faulty. � NodeB problems, such as power amplifier failure

For previous problems, check the work state whether there are alarms, whether it is normal. Solve the problem by replacing NEs, isolating faulty NEs, and adjust NEs.


Sadok Ben Ali 17

1.1.4 No Primary Pilot

��

No primary pilot areas refer to the areas where no primary pilot is or the primary cell changes frequently. In no primary pilot areas, UE hands over frequently, so the system efficiency is lowered and probability of call drop increases.

��

In no primary pilot areas, you can enhance the coverage by strong signals of a cell (or near cells) and reduce the coverage by weak signals of other cells (or far cells) by adjusting antenna down tilt and azimuth.

1.2 Coverage Analysis Processes

1.2.1 Downlink Coverage Analysis

Downlink coverage analysis involves analyzing CPICH RSCP obtained by drive test.

The quality standard of CPICH RSCP must be combined with the optimization standard. Assume that the optimization standard is as below:

CPICH_RSCP � –95 dBm >= 95% Scanner test result in outdoor unloaded conditions

The corresponding quality standard is:

� Good if CPICH_RSCP � –85 dBm � Fair if –95 dBm � CPICH_RSCP < –85 dBm � Poor if CPICH_RSCP < –95 dBm

Mark the areas with weak coverage or common seamless coverage of large areas for further analysis. Mark the areas with downlink coverage voids, analyze the distance relations with neighbor NodeBs and environments, and check the following:

� Whether the CPICH RSCP of neighbor sites is normal � Whether coverage can be enhanced by adjusting antenna down tilt and azimuth.

During adjusting antennas, avoid new coverage voids while eliminating some coverage voids. If the coverage voids cannot be eliminated by adjusting antennas, construct sites to solve it.

��

Usually, the strongest RSCP received by each scanner in the coverage area must be above –95 dBm.

Start Assistant. Analyze scanner-based RSCP for 1st Best ServiceCell, and you can obtain the distribution of weak coverage area.

In 0, weak coverage areas with RSCP smaller than –95 dBm in the DT route. According to scanner and UE, the pilot RSCP is acceptable. If the scanner antenna is mounted outside the car, and the UE is inside the car, there is a penetration loss of 5 to 7 dB. Use scanner data to avoid incomplete pilot information measured by UE due to missing neighbor cells.


Sadok Ben Ali 18

Figure 2.1: RSCP for 1st Best Service Cell

��

Cell primary pilot analysis is analyzing cell scramble information obtained in DT. The content to be checked include (by Assistant):

� Weak coverage cell Start Assistant. Analyze scanner-based RSCP for SC, and you can obtain the signal distribution of each cell (scramble). According to DT data, if the scramble signals of a cell are not present, probably some sites cannot transmit signals during test. If a cell cannot transmit signals during DT, the DT of relative areas must be re-performed. Very weak coverage might be result of blocked antennas, so you must check the survey report of the site and check installation of on-site antennas. No (poor) coverage cell might be due to that the DT route does not cover the cell coverage area. In this case, reevaluate the DT route for the rationality and perform DT again. � Cross-cell coverage cell Start Assistant. Analyze scanner-based RSCP for SC, and you can obtain the signal distribution of each cell (scramble). If the signals of a cell are widely distributed, even in the neighbor cells and the cells next to its neighbor cells, the signals of the cell is present, the cell encounters a cross-cell coverage which might be due to over high site or improper down tilt of antenna. The cross-cell coverage cells interferes neighbor cells, so the capacity declines. You can solve the problem by increasing the down tilt of antenna or lowering the height of antenna. Avoid forming new weak coverage areas while solving cross-coverage problems. Pay special attention to the adjustment of engineering parameters which might cause coverage voids. Be conservative that cross-cell coverage is better than coverage voids if no other choices are available. � No primary pilot cell Start Assistant. Analyze scanner-based SC for 1st Best ServiceCell, and you can obtain


Sadok Ben Ali 19

the scramble distribution of the best cell. If multiple best cells changes frequently in an cell, the cell is a no primary pilot cell, as shown in 0 No primary pilot cell forms due to the following causes:

� Cross-cell non-seamless coverage due to over high site � Pilot pollution in some areas � Coverage voids at edges of coverage areas

Therefore intra-frequency interferences forms which causes ping-pong handover and affects performances of service coverage.

Figure 2.2: Distribution of pilot SC for the 1st Best Service Cell

��

Missing neighbor cells, improper parameters of soft handover, cell selection and reselection cause the consistent between scanner primary pilot cell and camped cell in idle mode or Best Service Cell in the active set in connection mode of UE. After optimization, the Ec/Io for 1st Best Service Cell of UE and scanner is consistent. In addition, the coverage map of UE is marked by clear bordering lines of Best ServiceCell, as 0.


Sadok Ben Ali 20

Figure 2.3: Analyzing comparison of UE and scanner coverage

1.2.2 Uplink Coverage Analysis

The corresponding quality standard is: � Good if CPICH_RSCP � –85 dBm � Fair if –95 dBm � CPICH_RSCP < –85 dBm

Poor if CPICH_RSCP < –95 dBm

Uplink coverage analysis is analyzing UE transmit power obtained in DT.

The quality standards of UE transmit power must be combined with optimization standards. Assume the optimization indexes of UE transmit power as below:

UE_Tx_Power � 10 dBm >= 95%

The test result of voice service by test handset. Assume the maximum transmit power of UE is 21 dBm.

The defined corresponding quality standards are:

� Good if UE_Tx_Power � 0 dBm � Fair if 0 dBm < UE_Tx_Power � 10 dBm � Poor if UE_Tx_Power > 10 dBm

For areas with poor index, judge whether the increasing of UE transmit power is due to call drop or poor uplink coverage. Geographically displayed on the map, the former is as a point of sudden increment with call drop while the latter is an area with seamless coverage unnecessarily with call drop.

Mark the areas with weak coverage or large common seamless coverage for further analysis. Check whether downlink CPICH RSCP coverage voids exist in the areas with uplink coverage voids. Solve the problem with both uplink and downlink weak coverage by analyzing downlink coverage analysis. If only the uplink coverage is poor without uplink interference (see WCDMA Interference Solution Guide), solve the problems by adjusting down tilt and azimuth of antenna, and adding TMAs.


Sadok Ben Ali 21

��

Check for uplink interference by tracing and analyzing RTWP data. For details, see WCDMA Interference Solution Guide.

� �� !��

The distribution of UE transmit power reflects the distribution of uplink interference and uplink path loss. In 0, UE transmit power is lower than 10 dBm normally. Only when uplink interference and coverage area edge exist will the UE transmit power increase sharply to 21 dBm (Some UEs that support HSDPA, such as E620, with a power class of 3, the maximum transmit power is 24 dBm), and the uplink is restricted. Comparatively restricted uplink coverage occurs much easily in macro cells than in micro cells.

Figure 2.4: Distribution of UE transmit power

1.3 Coverage Problem Cases

1.3.1 Weak Coverage Cases Due to Improper Engineering Parameters

��

In 0, the pilot RSCP is lower than –95 dBm in the marked red area. This belongs to weak coverage, which might cause call drop.


Sadok Ben Ali 22

Figure 2.5: RSCP Coverage

��

In 05, the problem lies sector B mainly covers the marked area but Materials Building sector A partially covers the marked area. Initially engineers consider enhancing the coverage of the marked area by adjusting the two cells. According to the survey report, other buildings opposite Materials Building prevent sector A from transmit signals, so adjusting antenna fails to enhance the coverage of the areas.�

��

Keep the parameter configuration of Materials Building sector A, but adjust the azimuth of the antenna sector B from 170° to 165°, down tilt from 10° to 8°.

1.3.2 Cross-cell Coverage Due to Improper NodeB Location

��

In a trial office, the Erqi Rd. NodeB is 60-meter high, over 20 meters than nearby buildings. This causes cross-cell coverage easily and brings intra-frequency interference to other NodeBs, as shown in Errore. L'origine riferimento non è stata trovata.6.

��

For a high NodeB problem, adjust fixed electric down tilt of antenna from 2° to 6°. Because the Erqi Rd. NodeB is at the edge of network coverage, reduce interferences to other NodeBs by adjusting antenna down tilt and azimuth. In this case, no equipment is removed. Engineers solve the cross-cell coverage by increasing mechanism down tilt and adjusting azimuth.


Sadok Ben Ali 23

��

After adjustment of down tilt to 4°, the most cross-cell coverage areas are eliminated, with only few cross-cell coverage areas, as shown in Errore. L'origine riferimento non è stata trovata.. For similar high NodeBs, you can combine adjustable down tilt of electric antenna and mechanism antenna to better control signal coverage.

1.3.3 Coverage Restriction Due to Improper Installation of Antennas

��

From 06, the antenna of a project is mounted on a roof (10-meter tall).

Figure 2.6: Coverage restriction due to antenna blocked by roof

At the optimization stage after network construction, in front of the traffic lights below antennas, video quality declines due to VP mosaic and PS384K service is reactivated.�

��

In terms of planning, 3G and 2G antennas are mounted in a co-location site. According to coverage test data of 2G antenna, 2G signals does not fluctuate sharply under the site and under the traffic lights. Namely, if the 3G and 2G antennas are in the same location, 3G signals will cover the areas around traffic lights. The problem lies in that the 3G antenna is mounted too close to the wall on the roof and the wall blocks signals so the special installation conditions of antennas are not met. In addition, the 2G antenna and its installation parts affect the pattern of 3G antenna. This changes the radiation pattern of 3G antenna. According to the installation scene, adjusting location of 3G antenna is difficult.

��

According to discussion between 2G and 3G engineers, the minimum adjustment solution without affecting 2G coverage is as below: Connect the 3G and 2G TX/RX feeder to two feeders of outside wideband polarization antenna Connect the 3G and 2G RX feeder to two feeders of inner wideband antenna.


Sadok Ben Ali 24

2 Pilot Pollution Problem Analysis

2.1 Pilot Pollution Definition and Judgment Standards

2.1.1 Definition

The pilot pollution is that excessive strong pilots exist in a point but no primary pilot is strong enough.

2.1.2 Judgment Standards

Pilot pollution exists if all the following conditions are met: � The number of pilots that meet the following condition is more than ThN CPICH_RSCP > ThRSCP_Absolute � (CPICH_RSCP1st - CPICH_RSCP(ThN +1)th)< ThRSCP_Relative Assume that ThRSCP_Absolute = –100 dBm, ThN = 3, and ThRSCP_Relative = 5 dB, and then pilot pollution exists if all the following conditions are met:

� More than three pilots meet the following condition CPICH_RSCP > –100 dBm. � (CPICH_RSCP1st - CPICH_RSCP4th) < 5 dB

2.2 Causes and Influence Analysis

2.2.1 Causes Analysis

Ideally the signals in a cell is restricted within its planned range. However the signals cannot reach the ideal state due to the following factors of radio environment:

� Landform � Building distribution � Street distribution � Waters

Pilot pollution is the result of interaction among multiple NodeBs, so it occurs in urban areas where NodeBs are densely constructed. Normally typical areas where pilot pollution occurs easily include:

� High buildings � Wide streets � Overhead structure � Crossroad � Areas round waters

��

Due to restriction to site location and complex geographic environment, cell distribution might be improper. Improper cell distribution causes weak coverage of some areas and coverage by multiple strong pilots in same areas.

"��#��$��%��#��&��

If a NodeB is constructed in a position higher than around buildings, most areas will be with in the line-of sight range. Therefore signals are widely transmitted. Over high site cause difficult control of cross-cell coverage, which causes pilot pollution.

��

In a network with multiple NodeBs, the antenna azimuth must be adjusted according to the following factors:


Sadok Ben Ali 25

� NodeB distribution of the entire network � Coverage requirements � Traffic volume distribution

The sector azimuth of each antenna is set to cooperate with each other. If the azimuth is improperly set:

� Some factors might cover the same area. This causes excessive pilot pollution. � Weak coverage exist in some areas without primary pilot.

The previous two situations might lead to pilot pollution. Therefore you must adjust the antenna according to actual propagation.

��!� ��

Setting antenna down tilt depends on the following factors:

� Relative height to around environment � Coverage range requirements � Antenna types

If the antenna down tilt is improper, signals are received in the areas which are covered by this site. Therefore interferences to other areas causes pilot pollution. Even worse, interferences might cause call drop.

��#��!��

When the NodeBs are densely distributed with a small planned coverage rang and the PICH power is over high, the pilot covers an area larger than the planned area. This causes pilot pollution.

��'��

The signals cannot reach the planned state due to the following factors of radio environment:

� Landform � Building distribution � Street distribution � Waters

The ambient factors include:

� High buildings or mountains block signals from spreading The signals of a NodeB to cover a target area are blocked by high buildings or mountains, so the target area will have no primary pilot. This causes pilot pollution. � Streets or waters influences signals When the antenna direction is pointing a street, the coverage range is expanded by the street. When the coverage range of a NodeB overlaps with the coverage range of other NodeBs, pilot pollution occurs. � High buildings reflect signals When high glassed buildings stand near a NodeB, they will reflect signals to the coverage range of other NodeBs. This causes pilot pollution.

2.2.2 Influence Analysis

Pilot pollution causes the following network problems.

��(��

Multiple strong pilots interferes useful functional signals, so Io increases, Ec/Io decreases, BLER increases, and network quality declines.


Sadok Ben Ali 26

��#��

More than three strong pilots or no primary pilot exists in multiple pilots, frequent handover occurs among these pilots. This might cause call drop.

��

The interference of the areas with pilot pollution increases, the system capacity declines.

2.3 Solutions to Pilot Pollution

2.3.1 Antenna Adjustment

According to the test, change pilot signal strength of an area with pilot pollution by adjusting antenna down tilt and azimuth. This changes the distribution of pilot signals in the area. The principle for adjustment is enhancing primary pilot and weakening other pilots. To enhance pilot coverage of an area, adjust the antenna azimuth pointing the area. To weakening pilot coverage of an area, adjust the antenna azimuth pointing the opposite direction of the area. Adjusting down tilt is similar. You can increase the cell coverage range by reducing antenna down tilt and reduce cell coverage range by increasing antenna down tilt. Adjusting antennas is restricted to a range. If the down tilt is over small, you might enhance cell coverage but causes cross-cell coverage. If the down tilt is over large, you might weaken cell coverage but you might change the antenna pattern.

07 shows the pilot pollution due to improper antenna azimuth.

Figure 2.7: Pilot pollution due to improper antenna azimuth

In 07, the area marked in black encounters pilot pollution due to improper azimuth of the antenna of SC100 sector (scramble No. is 100). The SC100 sector covers the area with an antenna azimuth of 90°, so the coverage is poor with weak signals and no primary pilot, which cause pilot pollution.

After adjustment of the antenna azimuth from 90° to 170°, the primary pilot signals become stronger and pilot pollution is eliminated.�

08 shows the pilot pollution due to improper antenna down tilt.�


Sadok Ben Ali 27

Figure 2.8: Pilot pollution due to improper antenna down tilt

In 08, the area marked in blacked encounters pilot pollution due to improper antenna down tilt. The down tilt of SC360 cell is 2°, so the coverage area is large, cross-cell coverage is difficult to control, and interferences to other areas form.

After adjustment of antenna down tilt of SC360 cell from 2° to 7°, the cross-cell coverage by SC360 cell is eliminated and pilot pollution is eliminated.

Some areas with pilot pollution is inapplicable to the previous adjustment. You can use the following methods based on actual situation:

� Change the antenna to a different type � Add reflection device or isolation device � Adjust installation position of antenna � Adjust NodeB location

2.3.2 PICH Power Adjustment

Pilot pollution is caused by the coverage by multiple pilots. A direct method to solve the problem is to form a primary pilot by increasing the power of a cell and decreasing the power of other cells. An over large down tilt causes aberration of antenna pattern. To reduce coverage range by pilot, you can decrease PICH power. Over small down tilt causes cross-cell coverage. To increase coverage range by pilot, you can increase PICH power. Adjusting power and adjusting antenna must cooperate.

09 shows the pilot pollution due to improper distribution of cells.


Sadok Ben Ali 28

Figure 2.9: Pilot pollution due to improper distribution of cells

In 09,

� The distance between NodeB A and NodeB B is 1260 meters. � The distance between NodeB A and NodeB C is 2820 meters. � The distance between NodeB B and NodeB C is 2360 meters.

The distances is unbalanced, so the pilot pollution is difficult to eliminate.

The optimization is to reduce weak pilot strength and eliminate pilot pollution, detailed as below: � Ensure seamless coverage between cells by not adjusting transmit power of SC20 and SC30 cells. � Decrease the PICH power of SC10, SC40, and SC50 cells by 3 dB. These cells have little impact on seamless coverage.

2.3.3 Using RRU or Micro Cells

If adjusting power and antenna is not effective to solving pilot pollution, use RRU or micro cells.

Using RRU or micro cells aims to bring a strong-signal coverage in the area with pilot pollution, so the relative strength of other signals decreases.

When a network expansion is necessary or more requirements is on network quality, using RRU or micro cells is recommended. Micro cells are used in traffic hot spot areas, they support multiple carriers. Micro cells are used if large capacity is needed. Compared with using RRU, using micro cells is more expansive. 010 shows pilot pollution due to ambient factors.


Sadok Ben Ali 29

Figure 2.10: Pilot pollution due to ambient factors

The area marked in black encounters pilot pollution due to ambient factors. The area is covered by SC60 cell of NodeB A, SC110 cell or NodeB B, and SC130 cell of NodeB C. However, shown in Errore. L'origine riferimento non è stata trovata., hills prevent NodeB A from transmitting signals, high buildings prevent NodeB B and NodeB C from transmitting signals, so the signals from NodeB A, NodeB B, and NodeB C are weak. On the contrary, SC240 and SC250 cells of NodeB D have good propagation conditions in this direction. Therefore the cross-cell coverage is serious and pilot pollution occurs.�

High buildings or hills blocks the area, so no strong pilot is present in the area. For this problem, adjusting antenna down tilt has little effect on eliminating pilot pollution. Instead adding RRU helps solve the problem.

2.4 Process for Analyzing Pilot Pollution Problem

The process for analyzing pilot pollution problem proceeds as below: Start Assistant. Analyze scanner-based RSCP for 1st Best Service Cell and EcIo for 1st Best ServiceCell. Select the areas with high RSCP and poor EcIo as candidate areas with pilot pollution. Analyze scanner-based Whole PP. Select the areas corresponding to candidate areas as the key areas with pilot pollution. Locate the cells that cause pilot pollution of the key areas. Based on RSCP for 1st Best ServiceCell, judge whether the pilot pollution is caused by existence of multiple strong pilots or lack of a strong pilot. For the former cause, you can solve the problem by weakening other strong pilots. For the latter cause, you can solve the problem by strengthening some strong pilot. Analyze the RSCP and Ec/Io distribution of areas related to pilot pollution and confirm the cells that need eliminating the coverage of an area and that need enhancing the coverage of an area. Based on the actual environment, analyze the specific causes to pilot pollution. For specific causes, provide solutions to pilot pollution. While eliminating pilot pollution in an area, consider the influence to other areas and avoid causing pilot pollution or coverage voids to other areas. Retest after adjustment. Analyze RSCP, Ec/Io and Whole PP. If they cannot meet KPI requirements, re-optimize the network by selecting new key areas until KPI requirements are met.

3 Handover Problem Analysis

During RF optimization stage, the involved handover problem is about neighbor cell optimization and SHO Factor based on DT control.


Sadok Ben Ali 30

Control the size and location of handover areas by adjusting RF parameters. You can eliminate handover call drop due to sharp fluctuation and increase handover success rate.

For other handover problems, see WCDMA Handover and Call Drop Problem Optimization Guide.

3.1 Neighbor Cell Optimization

The neighbor cell optimization includes adding and removing neighbor cells.

Missing neighbor cells causes that a strong-pilot cell cannot be listed into the active set so the interference increases as strong as call drop occurs. For missing neighbor cell, you must add necessary neighbor cells. Redundant neighbor cells causes that the neighbor cell information is excessive and unnecessary signals cost occurs. When the neighbor cell list is fully configured, the needed neighbor cell cannot be listed. For this problem, remove redundant neighbor cells.

During RF optimization stage, missing neighbor cell is a key problem. The methods for adding neighbor cells are listed below.

3.1.1 DT Data Analysis

��

The daemon analysis tools can usually check for missing neighbor cells. The principle is as below:

� Compare the pilots scanned by scanner and the configured pilots of neighbor cell list. � Locate these pilot scrambles that meet the handover conditions and that are not in the neighbor cell list. Output them as a missing neighbor cell report.

The following checks and methods related to missing neighbor cells are based on Assistant.

Type information about NodeB and neighbor cells Decide conditions for judging neighbor cells Change the conditions for judging neighbor cells by selecting Modify Dataset Property. The default configuration is that if the difference between the pilot of candidate cell and the base cell is within 5 dB the candidate cell can be listed as a neighbor cell. The configuration must comply with the actual configuration of system (overall parameters), as shown in 0.

Figure 3.1 Changing conditions for judging neighbor cells

The parameters and meanings are as below (according to default configuration of RNC1.5 (Huawei RNC), you just list the parameters to be changed):


Sadok Ben Ali 31

Parameter Meaning Recommended value

1A Threshold 1A event threshold 3 dB

1A Hysteresis 1A event hysteresis 0 dB

1A Time to Trigger Time to trigger 1A event 0.320s

1B Threshold 1B event threshold 6 dB

1B Hysteresis 1B event hysteresis 0 dB

1C Hysteresis 1C event hysteresis 4 dB

1D hysteresis 1D event hysteresis 4 dB

Count Threshold Times threshold for judging neighbor cells 10

Table 6.1: Handover Parameters

Generate a missing neighbor cell report

Figure 3.2: Generating neighbor cell analysis report by using Assistant


Sadok Ben Ali 32

Figure 3.3 Result of missing neighbor cells

For the missing neighbor cells generated automatically by Assistant, you must check according to the location information of the cell on the map whether to add the missing neighbor cells to the neighbor cell list. For the missing neighbor cells due to cross-cell coverage, the primary task is to solve coverage problem by adjusting RF parameters. If this fails, you can temporarily solve the problem by adding neighbor cells.

��

The daemon analysis tool can seldom analyze UE data automatically and generate missing neighbor cells, so RNO engineers must analyze the missing neighbor cells one by one for confirmation. Missing neighbor cell might cause call drop or access failure or cause Ec/Io to deteriorate for a period. Based on data analysis by scanner, you can easily locate these points with missing neighbor cells, detailed as below: Compare the active set Ec/Io distribution diagram measured by UE and that measured by scanner The spots with missing neighbor cells has a poor Ec/Io measured by UE and a strong Ec/Io scanned by scanner. Locate the areas for further analysis. Check the points with poor Ec/Io and check whether the strongest scramble by scanner is neither in active set nor in monitoring set. If yes, move to the third step for confirmation. If the scramble exists in the monitoring set, the problem is not about missing neighbor cell but about Ec/Io deterioration due to handover (reselection) delay and soft handover failure. Check the latest intra-frequency measurement control whether the neighbor cell list contains the strong scrambles by scanner. You can also directly check the neighbor cells continuation of the base cell under the RNC for deciding missing neighbor cells. The following paragraphs describes a case about call drop due to missing neighbor cell.

Check the Ec/Io coverage information of active set measured by UE, and you can find that the Ec/Io of the active set is weak near call drop point and the signals are as weak as lower than –15 dB. The base cell is SC209 cell.


Sadok Ben Ali 33

Figure 3.4: Variation of active set Ec/Io recorded by UE before call drop

You also need to check data from scanner about the call drop point for the points with poor signals. The signals , from SC128 cell, measured by scanner is strong, as shown in 0.

Figure 3.5: Variation of active set Ec/Io recorded by scanner before call drop

From 0 and 0, SC128 encounters missing neighbor cell. For confirmation, check the message process behind to front for intra-frequency measurement control message. Check whether SC128 exists in the list of intra-frequency neighbor cells. The result is that SC128 is not in the list of intra-frequency neighbor cells, therefore the call drop is caused by missing neighbor cell.

If only UE recorded data in the test without data from scanner, confirm by the following method whether the problem is caused by missing neighbor cell:

Check scrambles of all cells listed in active set measured by UE before call drop Check scramble information of the cell where UE camps again after call drop and check whether the scrambles are in active set and monitoring set before call drop If yes, the call drop might be due to missing neighbor cell. Check the list of neighbor cells


Sadok Ben Ali 34

3.1.2 Removing Redundant Neighbor Cells

According to the protocol, the maximum WCDMA neighbor cells is 32. The base cell itself is also included in the intra-frequency neighbor cell list, so the actual intra-frequency neighbor cell is 31 at most. If there are already 31 or more neighbor cells, adding necessary neighbor cells in optimization is difficult. Therefore, you must remove some redundant neighbor cells.

��

You must be very careful to remove redundant neighbor cells. If the necessary neighbor cells are removed, problems like call drop occur. Therefore follow the principles below:

� Before removing neighbor cells, check the revision records of neighbor cells whether the neighbor cells to be removed are those that were added in previous DT and optimization. � After removing neighbor cells, perform comprehensive test, including DT and call quality test (CQT) in important indoor spots, and check for abnormalities. If there are abnormalities, restore the data configuration.

�� )��

During RF optimization stage, you might remove neighbor cells in the following situations:

� Remove the neighbor cells related to cross-cell coverage on the precondition that the cross-cell coverage problem is solved and no new weak coverage areas are appear. � Remove neighbor cells according to experiences while referring to the network topology structure. This applies to that the original neighbor cell list is full and new neighbor relations must be added. Perform test after removal and confirm that the removal does not cause bigger problems. Otherwise, you must reselect the neighbor cells to be removed.

In the later stages, you can refer to removing traffic measurement statistics. For details, see WCDMA Handover and Call Drop Problem Optimization Guide.

3.2 SHO Factor based on DT Analysis

3.2.1 Definition of SHO Factor based on DT

According to the DT data from scanner, you can obtain the SHO Factor based on DT, defined as below:

DTin points collected-scanner totalofNumber conditionshandover meet the that DTin points collected-scanner ofNumber

RatioHandoverSoft

No subscribers are using the network during RF optimization stage, so UE DT data of entire network in a time is used and geographically averaged by 5 meters. You can obtain the ratio of the points in soft handover state to all DT points. Set the scanner consistent to the system parameters with default configuration, such as 1A and 1B threshold.

3.2.2 General Principles and Methods in Optimization

The SHO Factor based on DT during RF optimization stage must be 5%–10% lower than the KPI target value, because the following optimizations cause SHO Factor based on DT to increase and brings difficulties in ensuring traffic measurement SHO Factor based on DT. At the end of large-scale coverage optimization and pilot pollution optimization, the SHO Factor based on DT will be within or close to the target range. Upon this, no specific optimization on SHO Factor based on DT is necessary, and you can adjust the ratio during parameter optimization. If the SHO Factor based on DT still cannot meet the requirements after large-scale adjustment, you must optimize the SHO Factor based on DT.


Sadok Ben Ali 35

If the SHO Factor based on DT is over large, decrease or change the handover areas by using the following methods for shrinking coverage areas:

� Increase the down tilt � Adjust azimuth � Decrease the antenna height � Decrease the PICH power

The precondition for adjustment is that the adjustment will not cause new coverage voids, coverage blind zone, and more pilot pollution. The adjustment proceeds as below: Start Assistant Analyze scanner-based RSCP for 4th Best Service Cell and for 3rd Best Select candidate cells in the 4th Best Service Cell and 3rd Best Service Cell 0 shows the RSCP for the candidates in 4th Best Service Cell. List the SC136 cell as a candidate cell. At this stage, the pilot pollution comes to an end. RSCP for 3rd Best Service Cell is more useful in terms of reference. Select the sites or cells to which the adjustment is applicable and does not break the preconditions. If the actual SHO Factor based on DT after adjustment is still different from the KPI one, select candidate cells from for 2nd Best Service Cell. The sites are densely distributed in microcell coverage areas, so the SHO Factor based on DT is much higher.

Figure 3.6 RSCP for candidate of 4th Best Service Cell


Sadok Ben Ali 36

Chapter III: Work Process In this chapter we will introduce how we are reporting the detecting and reporting Incidents which founded during the Drive test by making Incidents Reports in which we are giving the incidents with small description and recommendation.

In second step we are delivering KPIs report to the customer in which we are delivering the major WCDMA CS part KPIs. In third step we should produce some recommendations based to our analysis and the Network KPIs in order to increase network performance and reach target KPIs.

1 Incident Report The following figure present a template for the Incident report which is daily report including all incidents detected after one day drive test.


Sadok Ben Ali 37

Figure 3.1: Incident Report Template

2 KPI Analysis

2.1 RSCP

RSCP is an acronym used in UMTS and other CDMA cellular communications systems and stands for Received Signal Code Power. While RSCP can be defined generally for any CDMA system, it is more specifically used in UMTS. Also, while RSCP can be measured in principle on the downlink (DL) as well as on the uplink (UL), it is usually associated only with the downlink and thus it is presumed to be measured by the UE and reported to the Node-B. In brief, the received power on one code after de-spreading measured by UE on the pilot bits on a given channel. The reference point for the RSCP is the antenna connector of the UE. If Tx diversity is applied on the measured channel then the received code power from each antenna is separately measured and summed together in watt to produce a total received code power on the measured channel.

RSCP Analysis:

RSCP (dBm) Sample %

Above -65 34.81%

-75 to -65 35.33%

-85 to -75 24.08%

-95 to -85 5.21%

-100 to -95 0.26%

-100 & Below 0.31%

Total 100%

Table 3.1: RSCP Distribution


Sadok Ben Ali 38

Figure 3.2: RSCP Cumulative Distribution

RSCP Coverage Plot:

Figure 3.3: RSCP Cumulative Distribution


Sadok Ben Ali 39

2.2 Ec/No Analysis:

Ec/No is the received Energy per Chip divided by the Power Density in the band. The CPICH Ec/No is identical to CPICH RSCP/UTRA Carrier RSSI. Measurement shall be performed on the Primary CPICH. The reference point for the CPICH Ec/No shall be the antenna connector of the UE.

Ec/No Values Values range from 0 to -30 dB and correspond to an estimated Bit Error Rate

Ec/No Results:

Legend values Sample %

Below -20 dB 0.07%

-20 dB to -15 dB 0.64%

-15 dB to -10 dB 14.34%

-10 dB to 0 dB 84.95%

Table 3.2: Ec/No Distribution

Figure 3.5: Ec/No Cumulative Distribution


Sadok Ben Ali 40

Ec/No Plot:

Figure 3.6: Ec/No Cumulative Distribution

2.3 UE Tx Transmit Power

In the UMTS standard various power classes are defined; the ones shaded in grey are ones that could be used in a handset and comply with the guidelines on SAR. The maximum (average) power for a class 4 (lowest power) FDD mode terminal is therefore the same as a GSM1800 handset, however, the signal is continuous and not pulsed.

Class Power Class FDD Mode Power (dBm)

1 33(+1/-3)

2 27(+1/-3)

3 24(+1/-3)

4 21(±2)

Table 3.3: UE Classification


Sadok Ben Ali 41

UE Transmit Power values:

UE Tx Power (dBm) Sample %

Below -40 dBm 4.35%

-40 dBm to - 25 dBm 40.41%

-25 dBm to 0 dBm 48.95%

0 dBm to 10 dBm 4.0%

10 dBm to 15 dBm 2.25%

15 dBm to 20 dBm 0.02%

20 dBm and above 0.02%

Table 3.4: UE Tx Transmit Power Distribution

Figure 3.7 UE Transmit Power Distribution


Sadok Ben Ali 42

UE Transmit Power Plot:

Figure 3.8: UE Tx Power Plot

2.4 BLER

The BLER report gives an idea of downlink quality. The BLER result is derived from the radio link control layer’s packet acknowledgement message. BLER is simply the ratio of the number of negatively acknowledged blocks to the total number of blocks acknowledged by the UE: If the target BLER is too large, retransmission will be less probable and delay will increase otherwise retransmission will be less probable and power efficiency will decline. When the power is not restricted by scenarios, lower target BLER to increase effective throughput rate.

BLER Values:

BLER Sample %

Below 5 76.99%

5 to 10 6.47%

10 to 50 14.79%

Above 50 1.75%

Table 3.5: BLER Distribution


Sadok Ben Ali 43

Figure 3.9: BLER Distribution

BLER Coverage Plot:

Figure 3.10: BLER Plot


Sadok Ben Ali 44

2.5 Active Set Size

The set of Cells that the UE is connected to is called its active set (AS). Several radio resource management parameters play a role here, most notably the maximum allowed active set size which is set as 3 in Etisalat UMTS network. If the UE is connected to more than one Active Set, then it affects the radio link quality, this is known as macro diversity in WCDMA and it consumes more radio resources and increases the interference in the system. Having Active set = 2 or 3 should be 30% to 40 % from whole network coverage area. Active set Size Percentage:

Active Set Size Sample %

Active set = 1 44.35%



Table 3.6: Active set size distribution

Figure 3.11: Active Set Size Distribution


Sadok Ben Ali 45

Active Set Size Plot:

Figure 3.12: Active set size plot

2.5 Events

The number of events resulting in NE Zone 1 are as follows:

Events Table

Events Count

Call Total 866

Call Setup Failure 68


Sadok Ben Ali 46

Table 3.7: Events

Events Plot

Figure 3.13: Events Plot

2.6 SC Distribution

Scrambling code distribution is presented below.

Dropped Call 21


Sadok Ben Ali 47

Scrambling code Plot:

Figure 3.14: SC Distribution Plot

2.7 3G – Zone 1 Statistics

Table 3.8: RSCP & Ec/No Statistic

Average Mode Max Min

RSCP (dBm) -69.25 -72 -41 -115

Ec/No (dB) -7.53 -5.4 -3 -22.9

KPI Value

Total calls 937

Call setup Fail 68

Dropped Call 21

Call Setup Success Rate (CSSR) 92.72%


Sadok Ben Ali 48

Table 3.9: KPIs

2.8 Call Setup Time

Minimum Setup Time 0.17

Maximum Setup Time 29.95

Average Setup Time 3.41

Table 3.9: Call setup Time Statistic

Figure 3.15: Call Setup Time Distribution

Call Setup Failure Rate 7.26%

Drop Call Rate (DCR) 2.42%

Call Completion Success Rate (CCSR) 90.50%


Sadok Ben Ali 49

Conclusion In this reports we tried to make a manual for optimizing WCDMA networks which contents many cases of problems in WCDMA RF interface. It was requested by the company which did my internship to keep it as reference for next optimization projects. And I m Introduced a sample of Incident Reports and KPI report which I have been delivering those kind of report to the customer (Etisalat UAE) during this project In which I was involved as WCDMA Optimizer Consultant. This internship was useful for me since I improved my Skills in Mobile network Optimization Filed. And based on that I will be able to find a good job opportunity in wireless planning and optimization filed overall the world.

pw benali imcne06

Documents