Drive Test Analysis a Case Study Libre (1)

Grant Chivandire - Radio Planning and

Optimisation Specialist1

RADIO PLANNING AND OPTIMISATION

Drive Test Log file Analysis- A Case Study

(Presentation/Discussion – 19/06/2009)

By

Grant Chivandire



Objectives

This presentation will cover the following:

Key drive-test log-file aspects

Layer 3 Messages

Call Drops

Handover Issues

Case Analysis 1

Case Analysis 2



Key Drive Test Log-file Aspects

BCCH of serving cell ad neighbours

CGI – MCC, MNC, LAC, CI

RxLev of serving cell and neighbours

BSIC of serving cell and neighbours

Time Slot

Channel type/mode

Hopping Information – HSN, frequencies

Timing Advance

Tx Power

Radio Link Timer

C/I, C/A

RxQual, BER, FER, SQI



Layer 3 Messages

Synchronization Channel Information

Synchronization of MS to BSS

Repeated many times during a call

Contains BCCH, BSIC, Frame number etc

System Information Type 5

Contains BCCH Allocation list

List used to determine the neighbour list

Measurement Report

Contains the created neighbour list

BCCH, BSIC, RxLev and RxQual

System Information Type 6

Contains Location Area Identification information

MCC, MNC, LAC, RLINKT, NCC PERM



Call Drops - Summary

Fishbone Diagram: Call Drops



Call Drops – “uﾏﾏary ふIoﾐt…ぶ

Fishbone Diagram : Call Drops



Handover Issues - Types

HANDOVER TYPES

Power Budget Handover

Difference in RxLev > PBGT margin

Level Handover

DL/UL RxLev < HO Threshold DL/UL

Quality Handover

Quality level < Quality thresholds

Interference Handover

High RxLev but poor quality



Handover Issues – Layer 3 Messages

Layer 3 Messages

Handover Command

Cell description

Handover reference number

Power command

Synchronization indicator

Channel mode/description

Frequency channel sequence

Mobile Allocation

Time stamp

Handover Access

Handover reference number

Handover Complete/Failure

When MS establishes signaling link successfully/unsuccessfully



Handover Issues - Problems

Short Neighbour List

Less than 6 neighbours in neighbour list

Confirm you do not need to define more neighbours

Late Handover

Check handover margins between neighbours

Hierarchical Cell Structure

Power control

Ping pong Handovers

No dominant server

Adjust hysteresis, handover margins and power

May require new site



Handover Issues – ProHleﾏs ふIoﾐt…ぶ

Unnecessary Handover

Increasing number of Handover increases quality problems

Missing Neighbours

No handover to neighbour cell that seems like best server

Sudden appearance of strong cell in neighbour list after handover

Stepping stone effect

Fake Neighbours

Good candidate but handover does not take place leading to a call

drop

Fake neighbour – same BCCH with defined neighbour, itself not

defined

Missing BCCH in measurement list, NCC Perm, Co-BCCH




Overshooting

Island effect

No good candidate to handover to

Timing Advance much greater than half of your inter-site distance

Handover to another BSC

Data configuration should be consistent across the 2 BSCs

CGI, BCCH, BSIC and Power parameters

Target Cell congested

HO signalling performed over FACCH by stealing bits from TCH

Penalties applied to that particular MS

High HO Failure rate and drop due to dragging




Interference

Causes difficulties in decoding corrupt MRs on BCCH

Signalling failures due to high interference on TCH (FACCH)

Too many neighbours

Increase BSC CP load and leads to signalling congestion

Small window period to measure every neighbor may lead to incorrect

decisions

Swapped sectors

Defined neighbours will show poor levels

May exhibit さsteppiﾐg stoﾐeざ effect

May lead to drag and drops



Handover Issues - Summary

Fishbone Diagram : Handover Failure



Case Analysis 1



Case Aﾐalysis ヱふIoﾐt…ぶ

Log file

Central_longCalls_20090514_CELLC_CENTRAL004.dlf

Analysis

High interference at call drop

Handover failure 8seconds before call drop

Serving Cell =>BCCH=1012, BSIC=57, CI=41776, RxLev=-78, RxQual=7

Confirm serving cell details with Signalling messages and other

database – BSCplan, Asset etc

Channel => Hopping TCH

Handover Failure => look at Handover Command and Handover Failure

messages

Check if neighbour is defined properly in database

Measurement Reports => observe target cell is 4th candidate in

neighbour list




Analysis

Confirm neighbour cell information in database

Notice the proximity of the MS to nearest neighbour cells

Formulate ideal behaviour – why did the MS try to handover to the

fourth candidate leaving the better 3?

Suspect fake neighbour problem

In Actix map, show CI properties on cells

In database, look for the BCCH of the 3 candidates and identify

defined cell and CI

Notice that the 2nd sector of serving site, a defined neighbour, has the

same BCCH as the first candidate

Missing neighbor has been detected




Analysis

Notice the distance of serving cell compared to nearby cells

Check terrain profile using Asset, clutter using Google Earth and

neighbour candidates levels to see if acceptable

Site is overshooting, consider down tilting



Case Analysis 2



Case Aﾐalysis ヲふIoﾐt…ぶ

Log file

Central_longCalls_20090514_CELLC_CENTRAL004.dlf

Analysis

High interference at call drop

Handover failure 15seconds before call drop

Repeated inter-cell handover failures following a series of intra-cell

handover

Neighbour defined correctly

Handover failure due to timer expired

Deduce that high interference corrupted signalling information

Notice intra-cell handover did not help interference problem

Review hopping parameters



Case Analysis 2

Analysis

From the map, notice that 21626 is facing the location but not in

neighbour list

Check from database if it is defined as a neighbour

Perform site audit and do coverage check

Check for swapped sectors

Perform scans with drive-test tool and determine best server

Observe the cell reselection after call drop – CI=21962

Check if defined in database

Notice that the defined cell with same BCCH is CI=31622

This is a fake neighbour

Explain that after failing to HO to 31626, an attempt to 21962 was

expected

Make corrections and propose new site to enhance coverage

THANK YOU !

Drive Test Analysis a Case Study Libre (1)

Documents