Title: Date: Page Number:

Drive Test Analysis 05/07/2006 1/30

Created by: Approved by: Doc Ref.:
Alexandre Silva UMTS-RN0-0005

Drive Test Analysis

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
 Contents

1 Measurements on a drive test............................................................................ 3

1.1 Radio measurements ......................................................................................... 4
1.1.1 Spectral measurements on UTRA carrier ........................................... 4
1.1.2 Measurements on common channels .................................................. 5
1.1.3 Measurements on dedicated channels ................................................. 8
1.2 Higher-layer measurements ............................................................................ 10
1.2.1 Transport channels ............................................................................ 10
1.2.2 User-data measurements ................................................................... 11
1.3 Load simulation .............................................................................................. 12
1.3.1 Uplink radio load 12
1.3.2 Downlink radio load.......................................................................... 13
2 Radio optimization based on drive tests ......................................................... 14
2.1 Call set-up failure............................................................................................ 15
2.1.1 Coverage problem ............................................................................. 16
2.1.2 Admission Control problem.............................................................. 17
2.1.3 Interference problem ......................................................................... 18
2.1.4 Active Set Management .................................................................... 19
2.2 Call drop 19
2.2.1 Coverage problem ............................................................................. 19
2.2.2 Interference problem ......................................................................... 22
2.2.3 Active Set Management problem...................................................... 22
1.1.4 RLC problem .................................................................................... 24
1.1.5 RL problem ....................................................................................... 25
1.3 Higher-layer performances on User plane ...................................................... 25
1.3.1 Offered bit-rate.................................................................................. 26
1.3.2 Transport Channel BLER.................................................................. 27
1.3.3 Performances for RLC AM............................................................... 27
1.4 Higher-layer performances on Control plane ................................................. 27
Abbreviations .............................................................................................................. 29

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
1. Measurements on a drive test

In this section, we are listing the different measurements that could be taken on a drive test,
while using either a 3G scanner or a trace mobile or both.
The purpose of this document is not to tackle with implementation-specific measurements
for commercial 3G scanners or trace mobiles.
On a drive test, one tester can take following measurements:
■ Radio measurements
■ Higher-layer measurements (RLC and over)

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
 1.1 Radio measurements
The 3GPP25.215 and the 3GPP25.133 are specifying the UE
measurement capabilities required in UMTS FDD.

3G scanner versus trace mobile

The main advantage of a 3G scanner is that the user can specify the
UTRA carrier to monitor, whereas a trace mobile will only measure the
various UTRA carriers at the cell selection without reporting this
measurement to the system.
Then, the user can specify the cells to monitor on a 3G scanner, whereas
a trace mobile will only measure the various cells at the cell selection in
idle mode and the active cells together with the declared neighboring
cells in connected mode.
A 3G scanner may also give some measurements onto the P-SCH and S-
SCH.

1.1.1 Spectral measurements on UTRA carrier

UTRA carrier RSSI

The UTRA carrier RSSI is the key measurement for the DL interference
seen by the UE.
From the 3GPP TS25.215, the Received Signal Strength Indicator is the
wide-band received power within the relevant channel bandwidth. In
wide-band systems, the spreading of the power by code channels do not
ensure a perfectly homogeneous power level over the spectrum. The
PAR (Peak to Average Ratio) gives an indication of the homogeneity of
the power level over the 5MHz bandwidth. The higher the number of
codes mixed the lower the PAR.
From 3GPP TS25.133, the reporting range for UTRA carrier RSSI is
from –100dBm to -25 dBm.
The UTRA carrier RSSI mainly encompasses the intra- and inter-cell
load on the UTRA carrier. But, interference due to adjacent channels or
other radio transmitters are also included.
The intra-cell contribution Iintra to the UTRA carrier RSSI can be
extracted from the Transmitted Carrier Power measured at the Node B
antenna connector and the path-loss measured by the drive test chain on
the CPICH as:
Iintra = Transmitted Carrier Power cell, Node B- Pathloss cell, UE

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
 In the same way, the contributions of the neighboring cells for which
the CPICH can be demodulated can be analyzed.

1.1.2 Measurements on common channels

Cell detection
On one UMTS carrier, a cell is identified by its PSC (Primary
Scrambling Code). There are 512 scrambling codes grouped into 64
groups of each 8 different scrambling codes.
To detect a cell, the UE searches sequentially for:
● the universal 256-chip primary synchronization code, being
identical for all cells and repeated at the beginning of each slot
of the P-SCH. Once the peak is detected, the slot boundary is
known.
● the largest peak from the secondary synchronization code
word based on the slot boundary. The UE needs to check the
64 possible scrambling code group for the secondary
synchronization code word, beginning at each of the 15
positions, since only the slot boundary is known and not the
frame boundary.
● the current PSC in the identified scrambling code group by
scanning the CPICH: the correlation peak obtained while
descrambling the CPICH with the current PSC allows the
detection of the PSC.

RAKE measurements
The RAKE receiver performs the multi-path combining at the mobile
side. The 3GPP definitions for measurements are always after RAKE
recombining. It does not make sense to analyze each path for radio
optimization for following reasons:
● The RAKE coefficients are not configurable
● The RAKE algorithm is proprietary
● Multi-path profile changes from one test to the other

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
CPICH measurements
The CPICH is emitted continuously with a fixed power level, so that it
serves for the cell evaluation and downlink channel estimation at the
UE.
The reference point for the CPICH measurements is the antenna
connector of the UE, even though the RAKE receiver combines
measurements.
The CPICH measurements are the key radio measurements for cell
optimization.

CPICH RSCP
The CPICH RSCP is the key measurement for DL coverage. The
CPICH RSCP is an RXLEV measurement. This measurement is for:
● Cell re-/selection
● UL open loop power control
From the 3GPP TS25.215, the CPICH RSCP (Received Signal Code
Power) is the received power on one code measured on CPICH.
From the 3GPP TS25.133, the reporting range is for CPICH RSCP is
from –115dBm to -25 dBm.
As a rule of thumb, one can say (for dense urban):
► -108dBm ≤ CPICH RSCP < -105dBm
Uncertain CS64 coverage on unloaded network
► -105dBm ≤ CPICH RSCP < -98dBm
Uncertain CS64 coverage on 50%UL-loaded network
► -98dBm ≤ CPICH RSCP < -85dBm
CS64 coverage on 50%UL-loaded network for outdoors
► -85dBm ≤ CPICH RSCP < -70dBm
CS64 coverage on 50%UL-loaded network for in-car
► -70dBm ≤ CPICH RSCP < -62dBm
CS64 coverage on 50%UL-loaded network for indoor daylight
► -62dBm ≤ CPICH RSCP
CS64 coverage on 50%UL-loaded network for indoor first-wall

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
Downlink path-loss
The path-loss measured on the CPICH is defined as:
Pathloss = CPICH TX Power - CPICH RSCP

With a typical value of 33dBm for the CPICH TX power, the range for
path loss is then 58dB to 148dB.
One should be careful with the figures for the path-loss, since this
definition does not correspond to the pure radio loss.
3GPP Pathloss = CPICH TX power @ NB antenna connector -
CPICH RSCP @ UE antenna connector

Pure Radio Loss = CPICH TX power @ NB antenna output -

CPICH Rx power @ UE antenna input

3GPP Pathloss = Pure Radio Loss + feeder losses - NB antenna

gain + UE antenna gain

With typical values of feeder losses (3dB), NB antenna gain (17dBi)

and UE antenna gain (0dBi), 3GPP Pathloss = Pure Radio Loss – 14dB
and the range for radio loss is then 72dB to 162dB.

CPICH EC/I0
The CPICH EC/I0 is the key measurement for radio optimization for
CDMA. The CPICH EC/I0 measures the soft radio capacity. This
measurement is for:
● Cell re-/selection
● Radio admission control
● DL open-loop power control
● Soft HO and inter-frequency Hard HO

From the 3GPP TS25.215, the CPICH EC/I0 is the received energy per
chip divided by the power density in the band. The EC/I0 is identical to
RSCP/RSSI.
From the 3GPP TS25.133, the reporting range for CPICH EC/I0 is from
–24dB to 0dB.
The typical distribution of the CPICH EC/I0 highly depends on the cell
load. Requirements in term of CPICH EC/I0 depends on:
► The non-limitation in UL and DL coverage
► The number of cells that can be recombined

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
 ► The mobile sensitivity
As a rule of thumb, one can say:
► -15dB ≥ CPICH EC/I0 ≥ -18dB
Preferred situation is more than 1 active RL
► -12dB ≥ CPICH EC/I0 ≥ -15dB
Good value for 60%-loaded network
► -8dB ≥ CPICH EC/I0 ≥ -12dB
Preferred situation is not more than 1 active RL
► CPICH EC/I0 ≥ -8dB
Only 1 active RL for high-quality best server

1.1.3 Measurements on dedicated channels

Per definition, a 3G scanner does not perform measurements on
dedicated channels.

Uplink

UE TX power
The UE TX power is the key measurement to analyze UL coverage.
It is the only uplink measurement that is available at the UE side. It is
defined in the 3GPP TS25.215 as the total UE transmitted power on one
carrier at the UE antenna connector.
From the 3GPP TS25.133, the reporting range for UE transmitted power
is from -50 dBm to 33 dBm. But, the Power Class of the UE bounds the
upper limit (defined in 3GPP TS25.101).
Power Class Maximum TX power
1 +33 dBm
2 +27 dBm
3 +24 dBm
4 +21 dBm

Test mobiles in R1 and R2 were Power class 3 only.

UE Tx power = PDPCCH,UL+ PDPDCH,UL

PDPCCH,UL = (βD/βC)2 x PDPDCH,UL

βC and βD are parameterized, βD being fixed at 15

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
 As a rule of thumb, one can say:
► 16dBm ≤ UE TX power < 24dBm (shadowing and fading)
► 9dBm ≤ UE TX power < 16dBm (in-car penetration)
► 0dBm ≤ UE TX power < 9dBm (indoor day-light)
► -10dBm ≤ UE TX power < 0dBm (indoor first-wall)
► UE TX power < -10dBm(deep indoor)

Downlink

Signal-to-Interference Ratio
The Signal-to-Interference ratio is not standardized for the UE.
This internal measurement should be used in the DL outer-loop power
control. The SIR target for DL is defined through UE-proprietary
algorithms to achieve the BLER Quality Value assigned by the UTRAN
(see Downlink Transport Channel BLER).
Therefore, one should be cautious when analyzing such measurements
in a mobile trace, since various definitions are possible.

Metrics for synchronization

The UE Rx-Tx time difference, the CFN-SFN observed time difference
and the SFN-SFN observed time difference are concerning
synchronization issue at the UE side over the radio interface.
● The SFN-SFN observed time difference is for identifying time
difference between two cells.
● The CFN-SFN observed time difference is for handover timing
purposes to identify active cell and neighbor cell time
difference.
● The UE RX-TX time difference is used for RL set up purposes
to compensate propagation delay of DL and UL in order to get
TPC commands on DL with the right timing for the UL power
control.

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
 1.2 Higher-layer measurements

1.2.1 Transport channels

Downlink BLER
The Transport Channel block error rate (BLER) is based on evaluating
the CRC of each transport block associated with the measured transport
channel after RL combination (performed by the Rake receiver). It is
computed as the ratio between the number of received transport blocks
resulting in a CRC error and the number of received transport blocks
over the measurement period.
Transport channel BLER value shall be calculated from a time window
with following size:
► if periodical reporting mode is specified by the UTRAN, the size
should be equal to the IE Reporting interval in mentioned SIB11 or
12 (see 3GPP TS 25.331)
► otherwise, this is an internal measurement at the UE side and the
window size is freely designed
The Transport channel BLER reporting range is from 0 to 1.
The BLER Quality value is a target indicated by the UTRAN at the
RRC CONNECTION SETUP and at the RADIO BEARER SETUP for
the initiated service. Therefore, the follow-up of the BLER helps
assessing whether the UE is performing according to the quality
requirements or not.
► If the BLER is lower than the targeted Quality value, the QoS
offered by the UTRAN is sufficient, as far as the data integrity is
concerned.
► If the BLER is higher than the targeted Quality value, the radio
interface is not able to satisfy the required QoS. This could be due
to:
● Shortage of power to be dedicated to the UE on DL
● Power limitations due to radio parameter settings
● Power limitations due to high traffic on DL
● High interference level on DL
● Poor performances of the DL outer-loop power control
Logically, the Transport channel BLER and the SIR are bound through
the DL outer-loop power control, which is a mobile-proprietary
algorithm.

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
 Poor performances could be following:
● Bad or slow convergence

RLC statistics
Three RLC modes are available for Radio Bearers: Transparent,
Unacknowledged and Acknowledged.
For all RLC modes, statistics can be done on traffic volume and data
rate.
In the RLC Acknowledged Mode, several algorithms manage packet
retransmission, mainly:
● Transmission and reception windows
● Polling and Status signaling
In the RLC Acknowledged Mode, following measurements can be done:
● Transmitted but not acknowledged PDU (UL)
● Retransmitted PDU (UL)
● Received erroneous PDU (DL)
● Window follow-up (UL and DL)

1.2.2 User-data measurements

TCP statistics
Mostly used assessments are the FTP throughput and the PING round-
trip time.

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
1.3 Load
We are not dealing with stress of network elements to determine critical
load in terms of processing capacity. Radio load consists only in
interference on radio interface. Therefore, no signaling or data
processing is generated. Radio load is critical for W-CDMA systems.

1.3.1 Uplink radio load

The main concern on uplink is to figure out the increase of the
interference level due to the traffic available on the cell. The
interference level can be expressed as the noise rise in dB and depends
on the cell load in % of the pole capacity.
Cell load, XUL , is in term of % of pole capacity and impacts the noise
rise as follows:
NoiseRise = -10log (1- XUL)

Following reference values can be used:

Cell load Noise Rise
50% 3dB
75% 6dB

N oise rise a s a function of the Cell loa d

45
40
No 35
ise
Ris 30
e
in 25
dB
20
15
10
5
0
0.00% 20.00% 40.00% 60.00% 80.00% 100.00%
Cell Loa d

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
1.3.2 Downlink radio load
The limiting factor on downlink will be the maximum TX power for the
Node B. Therefore, the power usage on DL is expressed as the
percentage of the maximum TX power for the cell.

Maximum output power

Others traffic
Other mobiles

Trace Mobile
traffic
Trace mobile

OD settings dBm dB related Activity factor

to CPICH (% time)

CPICH 33,0 1
P-SCH 28,0 -5 0,05
S-SCH 28,0 -5 0,05
PCCPCH (BCH) 31,0 -2 0,9
SCCPCH (FACH, PCH) 31,0 -2 1
AICH 24,0 -9 1
PICH 28,0 -5 1
Total power (common channels) 37,3
MAX cell power 43,0
(total power for 60% cell load) 40,8
(OCNS = 60%cell load - CCH) 38,2

dBm Total cell power

MAX DL power for AMR 31,0 41,2
MAX DL power for CS64 37,0 42,3
MAX DL power for PS128 39,0 43,0

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
2 Radio optimization based on drive tests

The analysis of drive test measurements allows the detection and eventually the localization
of radio problems over the coverage area. Generally, radio problems are sorted into:
■ Coverage issues
■ Mobility-management issues
■ Best-server issues
■ Interference issues

In this section, we are analyzing following aspects:

■ Call set-up failure
■ Call drop
■ Higher-layer performances

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
2.1 Call set-up failure
Here, we only consider call set-up failures that are due to the radio.
Following call set-up failures are not considered:
► Invalid USIM
► Unavailability of the network (CN, UTRAN, transport)
Investigations on call set-up problems should go through the following
list:
● Coverage problem
● Admission Control problem
● Interference problem
● Active Set Management problem
Following questions should drive the investigations:
► Are the CPICH measurements for the selected cell at normal values
with regard to the path loss?
► What is the load status of this cell at the Call set-up?
► At which stage of the call set-up did the procedure fail?

UE Node B RNC CN

RRC connection request

DCCH Radio Link set-up
Initiate AAL2 Iub data
connection for DCCH
RRC connection setup
RRC connection setup complete

Initial NAS message transfer

Authentication Procedure
Security mode command
RAB assignment request
Synchronised Radio Link reconfiguration
Initiate AAL2 Iub data
connection for DTCH

RB setup request Iu UP initialisation

RB setup complete

RAB Assignment Response

Uu Iub Iu

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
 2.1.1 Coverage problem
Call set-up may fail because of Coverage limitation when:
► The RACH message can not be sent: the RRC CONNECTION
REQUEST can not be addressed (UL, PRACH power settings)
► The UE can not serve the power required after the UL Open-Loop
Power Control
► The UE does not receive the Paging message for MTC (DL, PICH
and S-CCPCH power settings)
In case of repeated errors, further investigations must be driven,
respectively on:
► RRC CONNECTION REQUEST sending can be checked at the
mobile side
► Path loss on DL can be assessed based on scanner or mobile
measurements (CPICH RSCP). This should be done on the UE-
selected cell. Then, the UL Open-loop power result can be checked.
► Reception of PAGING notification can be checked at the mobile
side
► UL and DL MAPL for call set-up
In case of errors repeated over an area, further investigations must be
driven, especially on:
● Feeder installation for RX diversity for UL limitation
● Feeder losses
● Antenna azimuth
● PRACH , PICH, S-CCPCH power setting
● RACH parameters

UL MAPL for call set-up

When establishing the DPCCH, the proprietary SRNC calculation for
initial DPCCH power will require from the UE a given power level for
the DL DPCCH, which should be lower than the maximum Tx power of
the UE power class.
DPCCH_Initial_power = (CPICH_TX_power - CPICH_RSCP) +
(UL_SIR_target – 10log(UL_SF=256)) – 97

AND

DPCCH_Initial_power < UE_Max_Tx_Power

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
 Following default values are given as indications for R2:
● UL_SIR_target (on a per-service basis, DCCH: 2dB)
● UL_SF (on a per-service basis, DCCH: 256)
● UE_MaxTxPower (on a per-mobile basis: 24dBm)
● CPICH_TX_power (on a per-cell basis: 33dBm)

DL MAPL for call set-up

When establishing the DL DPCCH, the proprietary SRNC calculation
for initial DPCCH power will require from the Node B a given power
level for the DL DPCCH, which should be lower than the maximum
allowed Tx power on DPCCH.

DL_SIRTarget m 1
DPCCH _ initial _ power = ⋅ ⋅ Pt CPICH ,i ⋅ 0.4 ⋅ 10 Margin / 10
CPICH_Ec Iom,i SFm
AND
DPCCH_Initial_power < MaximumDLpower+CPICH_Tx_power

Following default values are given as indications for R2:

● UE_MaxTxPower (on a per-mobile basis:24dBm)
● DL_SIR_target (on a per-service basis,DCCH:5dB)
● DL_SF (on a per-service basis,DCCH:256)
● Margin (on a per-service basis,0dB)
● MaxDLpower (on a per-service basis,DCCH:-2dB)

2.1.2 Admission Control problem

Call set-up may fail because of Admission Control limitation when:
► DL radio resources are overloaded (RAC issue)
► UL radio resources are overloaded (RAC issue)
► Node B processing resources are overloaded (RAC issue)
► Transport resources are overloaded (CAC issue - tbc)
► RNC resources are overloaded (CAC issue - tbc)
RAC issue can be assessed by monitoring NBAP messages for RADIO
LINK SETUP FAILURE with the associated cause.
UL and DL radio overload can be due to:
► High cell load due to user traffic for UL a/o DL

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
 ► OCNS activation for DL
► Interference problem for UL a/o DL

2.1.3 Interference problem

Interference problems may be due to:
► Surrounding cells of the operating UTRAN
► Other radio sources than the operating UTRAN
Call set-up failure may be due to Interference problem when the RAC
rejects the call, while the effective cell load is not at maximum:
► The RTWP on UL in Node B measurements are high but with a
small number of users, i.e. not only cell users are contributing to
the noise rise.
► The RSSI on DL is higher than the Transmitted Carrier Power on
DL of the detected cells decreased by the Path Loss seen on CPICH
for each detected cell.

Intra-network interference
Intra-network interference can be analyzed with the help of the scanner
on DL.
See Pilot pollution
See Missing adjacencies in the neighboring cell list

External interference
Specific investigations with a spectrum analyzer have to be performed
for interference coming from other radio sources than the operating
UTRAN.
This becomes a Spectrum Clearance issue.
Two kinds of problems may then be distinguished:
► Interference that is due to out-of-band emissions from other
transmitters. In this case, filters must be set at the interfering
transmitter.
► Interference that is due parasite transmitters that should not
transmit in this band. In this case, parasite emissions must be
stopped in the UMTS band.

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
2.1.4 Active Set Management
The mobile can not synchronize on the Radio Link due to high pollution

2.2 Call drop

Here, we only consider call drops that are due to the radio. Following
call drops are not considered:
► Mobile auto-reset, power-off, etc.
► CN-generated reasons (to be analyzed in RANAP IU RELEASE)
► Failure in transport network
Possible reasons for radio call-drop are:
► Coverage problem
► Interference problem
► Active Set Management problem
► RLC unrecoverable error
► RL failure
Following questions should drive the investigations:
► Are the CPICH measurements for the Active Cells and the UE TX
power at normal values with regard to the path loss seen on the
cells?
► What are the last messages that the UE sent or received? (CELL
UPDATE, MEASUREMENT REPORT, ACTIVE SET UPDATE)
► What are the last BLER measurements from the UE?

2.2.1 Coverage problem

Call drop may be due to Coverage limitation when:
● The UE TX power reached the maximum value at the end of
the call (UL limitation)
● The UE SIR or the UE Rx power or the UE BLER on DL did
not reach its target value at the end of the call (DL limitation)
See Call set-up failure, Coverage problem

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
 On unloaded networks
On unloaded networks, coverage problem on UL a/o DL reveals RNP
defects a/o tradeoffs (propagation model, database accuracy, calibration
settings, etc.) or errors in RNP implementation.
Path loss on DL can be assessed based on scanner measurements
(CPICH RSCP). This should be done on the cells of the Active Set only.
In case of errors repeated over an area, further investigations must be
driven, especially on:
● Feeder installation for RX diversity for UL limitation
● Feeder losses
● Antenna azimuth
● CPICH power setting

On radio-loaded networks
On loaded networks, coverage problems may be linked to radio
limitations on DL. Indeed, the cell may not be able to serve the power
required by the Radio Link if the cell is already emitting at full power.
Cell TX power can be assessed based on Node B measurements on DL
Transmitted carrier power.

Drop Calls examples with explanation

As all sites are co-channel, if the RF is present, it will be used (for better
or worse). Therefore, it is imperative that each sector is carefully RF
controlled. Ideally each serves its intended area but not beyond.
Controlling and managing the RF environment is central to
optimization.

Problem Maps Solution

COVERAGE: Best - Best RSCP Scanner/Mobile Check antennas and feeders are OK, with nothing obscuring
server CPICH RSCP - Tx Power Mobile (note that high mobile view of antenna, like block building, rooftop effect or any
below required level power can mean either coverage or radiating object nearby the antenna pattern radiation,
interference problem) consider reducing the antenna down-tilt to improve
- Best SC regarding CPICH Ec/Io coverage. As a last resort, increase pilot power.

QUALITY: Best - Best CPICH Ec/Io Scanner/Mobile If Ec/Io lower than -9 db (network without load), the best-
CPICH Ec/Io below - RSSI CPICH from Scanner server RSCP is at expected levels, and the RSSI is high, need
required level - Best SC regarding CPICH Ec/Io to reduce Interference from other cells.
Scanner/Mobile
- Best RSCP 2nd, 3rd, 4th, 5th, (…) Scanner Check Best RSCP 2nd, 3rd, 4th, 5th, (…) in order to decide
- Tx Power Mobile which cells to remove from the interference area.

Can the azimuth of other cells be changed to reduce number

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
 of servers in that area? Can neighbouring cells be down-
tilted? Avoid facing sectors shooting against each other.
QUALITY/CAPACI - Number of link in the AS Mobile (Generally this will correspond also to areas with low Ec/Io,
TY: Active set too - Number of link in the AS simulated but not always). Check that the cell best server areas are
large, Soft handovers Scanner confined, not excessive, but big enough to guarantee a small
area too high and too - Best SC regarding CPICH Ec/Io overlap between cells in order to allow the soft handover.
many handovers along Scanner/Mobile
routes - Number of link in the AS Excel Graph Can the azimuth of neighbouring cells be changed? Balance
(desired values for Urban environments coverage levels with other cells using down-tilts?
will be 10 to 20% of soft HO areas)
Optimizing capacity will be important when traffic grows.
AS < 4 pilots, for 5dB window (ReportingRange) relative to
the best Ec/Io should be used to calculate the number of cell
in the active set.

Based on IS-95 experience, the best performing networks are

those that are designed for three-way hand-off or less.
Site with coverage far - Best SC regarding CPICH Ec/Io Scanner Typical of high sites. Ensure electrical down-tilt is sufficient.
beyond desired area - Best RSCP SC Scanner Increase combined down-tilt, but not below 10 to 12o
(depending on antenna type). Be aware that once off the
Cell appears in - Best SC regarding CPICH Ec/Io Scanner main lobe of the antenna, increasing tilting will be unlikely
neighbour list of cells - Best RSCP SC Scanner to reduce interference, hence consider maintaining tilt to
beyond the area cover desired area.
desired
Can antennas be lowered? As a last resort, reduce pilot
power; however be aware that this will also reduce in-
building coverage of the cell: hence reduce power just
sufficiently to keep the cell out of active sets or raise Ec/No
values.
CAPACITY: Cells - Number of link in the AS Mobile Check azimuths between co-site cells: where possible
from the same site - Number of link in the AS simulated maintain the azimuth between cells to 100-120o for 3 sectors
appear at similar Scanner site. Check antennas have unobstructed views, including at
coverage levels - Best SC regarding CPICH Ec/Io Scanner roof edges to minimise scatter.

Call Performance: If the correct cell is not used in soft Tuning neighbour lists is a major aspect of optimisation.
Neighbour Lists handovers, the Ec/No and BLER will Check ACTIX options to find quickly missing neighbour
degrade and Ue_Tx_Power will increase cells between Mobile and Scanner measurements.
in Mobile Plots.
Check correct values for parameters:
Compare also: ReportingRangeEvent1A_1B_1C and
- Best SC regarding CPICH Ec/Io Scanner HysteresisEvent1A_1B_1C.
- Best SC regarding CPICH Ec/Io Mobile
Check Active Set Management problems, mainly: ►The UE
is sending repetitive MEASUREMENT REPORT without
getting an ACTIVE SET UPDATE; ►The UE can not
manage as many radio links as present in the Active Set (e.g.
in case of Rake limitations)-PILOT POLLUTION.

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
Call Performance: For AMR, any areas where the voice Check BLER Exit Criteria’s:
Voice/PS performance quality degrades or the call drops happen - Voice AMR 12.2: BLER<2%
should be noted and analysed. - PS 64/384: DL- BLER<2% and Throughput>345 kb/s
UL- BLER<2% and Throughput>57 kb/s
For PS, it should note areas where the
throughput of the application reduces and Check Call Drops Reasons and related parameters. For radio
identify any parameter or RF related issues optimization solutions please refer to Quality problems.
that may be causing this.

Analyse also:
- BLER Mobile
- BLER Statistics

2.2.2 Interference problem

Call drop may be due to Interference problem when:
● The UE TX power reached the maximum value at the end of
the call (UL limitation) but the path loss should allow the
service coverage
● The UE SIR or the UE Rx power or the UE BLER on DL did
not reach its target value at the end of the call (DL limitation)
but the path loss should allow the service coverage
See Call set-up failure, Interference problem

2.2.3 Active Set Management problem

Call drop may be due to Active Set Management problem when:
► The UE is polluted by cells that could enter the Active Set but that
are not monitored
► The UE is sending repetitive MEASUREMENT REPORT without
getting an ACTIVE SET UPDATE
► The UE is adding a cell to the Active Set that is too low (e.g. in
case of radio resurgence)
► The UE can not manage as many radio links as present in the
Active Set (e.g. in case of Rake limitations)
► The UE has lost the synchronization between UL and DL (event 6f,
6g on UE RX-TX time difference)
In case of repeated errors, further investigations must be driven,
respectively on:
► Pilot pollution and Missing adjacencies in the neighboring cell list

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
 ► Reason for the non-reception of ACTIVE SET UPDATE from the
RNC
● MEASUREMENT REPORT are not received at the RNC
● ACTIVE SET UPDATE are not received at the UE
● Maximum numbers of RL a/o Node B in AS are reached
● RAC on target cell rejects the RL establishment
► Time-to-trigger and Hysteresis, Antenna tilt to cancel resurgence
► UE radio capabilities and maximum AS size

Missing adjacencies in a cell neighborhood

The UE is evaluating the radio quality of neighboring cells to detect
potential cells to add to the Active Set. The UE does not report
measurements on other cells that may be detectable but only on the list
of neighboring cells signaled by the S-RNC.
The list of neighboring cells is compiled at the S-RNC and transmitted
to the UE in RRC MEASUREMENT CONTROL messages. The
compilation of this list is proprietary. It mainly consists in assembling
the different sets of neighboring-cell lists for all the cells of the Active
Set.
The operator parameterizes the neighboring-cell list for each cell. The
first computed list is inherited from the RNP, which runs proprietary
algorithms to list neighboring cells. These lists may be not optimized
due to used algorithms or due to inappropriate methods for
neighborhood follow-up in deployment/densification phases.
Drive tests may help detecting missing adjacencies on a live network.
► Criteria for Event 1A
10Log(MNew) ≥ Wx10Log(MBest)+(1-W)x10Log(∑MI) -(R1a+H1a)

MNew : measurement on the cell entering the reporting range

Mi : measurement on a cell in the active set
NA: number of cells in the current active set
MBest: measurement result of the strongest cell in the active set
R1a : reporting range for the event 1a
H1a : hysteresis parameter for the event 1a

► Criteria for Event 1B

10Log(MOld) ≤ Wx10Log(MBest)+(1-W)x10Log(∑MI)-(R1b +H1b)

MOld : measurement on the cell leaving the reporting range

Mi : measurement on a cell in the active set

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
 NA: number of cells in the current active set
MBest: measurement result of the strongest cell in the active set
R1b : reporting range for the event 1b
H1b : hysteresis parameter for the event 1b

► Criteria for Event 1C

10Log(MNew) ≥ 10Log(MI) + H1c

MNew : measurement on the cell leaving the reporting range

Mi : measurement on a cell in the active set
H1c : hysteresis parameter for the event 1c

Pilot pollution
Soft handover is working well, when:
● Radio links have a non-negligible life time and no Ping-Pong
overloads the signaling,
● Radio links can be recombined in an effective manner at the
Rake receiver.
A polluter transmitter is a transmitter that checks all criteria to enter in
active-set but which is not admitted due to the active-set limit size.
One can distinguish different levels of criticality for pilot pollution:
● A non-potential candidate cell for soft handover
● A potential candidate cell for soft handover but with poor
performance
● A potential candidate cell for soft handover but not declared in
the neighboring cell list
To limit pilot pollution, further investigations must be driven, especially
on:
● Antenna azimuth and tilt
● CPICH power setting
● Cell maximum TX power
● Neighboring-cell list and Monitored Set

2.2.4 RLC problem

The Radio Link Control protocol can engender call drops, when errors
in this protocol can not be recovered. Errors do only concern the
Acknowledged Mode. In an error case, the UE drops the call and sends

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
 a RRC CELL UPDATE message on the RACH with the cause “RLC
unrecoverable error”.
Possible deadlocks in RLC AM are:
► Not acknowledged PDU in the Sending Window (UL)
► Not received error-free PDU in the Receiving Window (DL)
► Not received Status
► Not received Polling result

2.2.5 RL problem
The Radio Link maintenance can engender call drops, when
unrecoverable RL failures occur. In a failure case on DL, the UE drops
the call and sends a RRC CELL UPDATE message on the RACH with
the cause “RL failure”.
RL failures are described in 3GPP TS25.331. They consist in L1
synchronization loss. In CELL_DCH State, after receiving N313
consecutive "out of sync" indications from L1 for the established
DPCCH physical channel, the UE starts T313. If N315 successive "in
sync" indications from L1 are received before T313 expiry, the RL
failure is recovered and T313 is stopped. Otherwise, this is a RL failure.
L1 reports "out of sync" if one of these criteria is fulfilled:
► The UE estimates the DPCCH quality over the previous 160ms
period to be worse than a threshold Qout
► The 20 most recently received transport blocks with a non-zero
length CRC have been received with incorrect CRC. In addition,
over the previous 160ms, all transport blocks with a non-zero
length CRC have been received with incorrect CRC.

2.3 Higher-layer performances on User plane

The offered end-user QoS should also be monitored upon drive-tests.
However, one should separate Application Optimization from Radio
Optimization. Indeed, the assessment of the QoS offered by the UTRAN
should only restrict to QoS that the UTRAN can actually offer, this
mean RAB QoS.
Optimization of end-user QoS should be driven next and if the
subjective end-user QoS is not sufficient, following changes could
apply:
► Modify the Application settings for a better use of the RAB

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
 ► If it not sufficient, higher the RAB QoS requirements. This will
trigger further Radio Optimization with the new RAB
implementation
The RAB QoS is described in 3GPP TS23.107 with following
attributes:
► Maximum bit-rate (kbps)
► Guaranteed bit-rate (kbps)
► SDU error ratio a/o Residual bit error ratio
► Delivery order (y/n)
► Maximum SDU size (octets)
► Delivery of erroneous SDU (y/n/-)
► Transfer delay (ms)
► Traffic handling priority
► Allocation/Retention Priority
Following attributes are taken into account for the choice of the Radio
Bearer:
► Maximum bit-rate for PS and Guaranteed bit-rate for CS
► SDU error ratio a/o Residual bit error ratio for RB BLER for CS
► RLC mode is Transparent for CS and Acknowledged for PS (in-
order delivery, max SDU size, delivery of erroneous SDU)
Therefore, the monitoring of the QoS offered by the UTRAN relies on:
► Offered bit-rate
► Transport Channel BLER
► Performances for RLC AM (Acknowledged Mode)

2.3.1 Offered bit-rate

For CS services, the offered bit-rate can only be the requested one. The
bit-rate is guaranteed.
For PS services, the UTRAN can apply service downgrade, if the RAC
can not admit the requested bit-rate. The downgrade is then reconsider
at each Cell-FACH→Cell-DCH transition. Once a bit-rate is attributed,
a channel with the corresponding bandwidth is allocated over the radio
interface.

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
2.3.2 Transport Channel BLER
For CS services, the Transport Channel BLER is the Frame Error Rate
for the service, since no retransmission is allowed.
For PS services, the Transport Channel BLER is a tradeoff between the
effort in terms of bandwidth to support retransmission and the effort in
terms of radio to limit errors.

2.3.3 Performances for RLC AM

For PS services, the RLC performances in terms of retransmission rate
and delay impact the higher-layer QoS.

2.4 Higher-layer performances on Control plane

For the proper working of the Telecom procedures, minimal
performances on Control plane are required, especially in terms of
delay. But, the delay introduced for the Telecom procedures definitely
impacts the end-user perception of the network quality, especially at the
Call set-up.
Performances in terms of delay on the Control Plane depend on:
► The buffering and the processing times in the nodes
► The quality offered by the Transport Network
► The quality offered by the Radio Links for signaling
This is on this last item that the Radio Optimization can bring
improvement. The quality offered by the Radio Links can be mainly
monitored through:
► Access to common channels
► Offered bit-rate
► Transport Channel BLER
► Performances for RLC AM (Acknowledged Mode)
Radio optimization is done in the same way as for traffic channels.

End-user assessment will rely on Procedure times, such as:

► Attach time
► Paging time
► Call set-up time (MOC, MTC, MTM)
► Time for PDP context establishment

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
 Other Procedure Times, such as the time for SHO completion, may only
impact call set-up failure rate and call drop rate from an end-user point
of view.
Other Procedure Times, such as the Detach time, may only impact the
network performances from an operator point of view.
Events for the definition of the time interval are often specific for each
operator and related to commitments.

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
 Title: Date: Page Number:
Drive Test Analysis 05/07/2006 29/30
Created by: Approved by: Doc Ref.:
Alexandre Silva UMTS-RN0-0005

Abbreviations
AM Acknowledged Mode
AS Active Set
BLER Block Error Rate
CAC Connection Admission Control
CFN Connection Frame Number
CN Core Network
CPICH Common Pilot Channel
CRC Cyclic Redundancy Check
DL Downlink
DPCCH Dedicated Physical Control Channel
DPDCH Dedicated Physical Data Channel
FTP File Transfer Protocol
HO Handover
MAPL Maximum Allowable Path Loss
MOC Mobile Originating Call
MTC Mobile Terminating Call
MTM Mobile To Mobile
NB Node B
NBAP Node B Application Protocol
OCNS Orthogonal Channel Noise Simulation
PDU Protocol Data Unit
PSC Primary Scrambling Code
QoS Quality of Service
RAB Radio Access Bearer
RAC Radio Admission Control
RANAP RAN Application Protocol
RLC Radio Link Control
RNP Radio Network Planning
RRC Radio Resource Control
RSCP Received Signal Code Power
RSSI Received Signal Strength Indicator
RTWP Received Total Wide-band Power
SCH Synchronization Channel (Primary: P-SCH, Secondary: S-SCH)
SFN Sequence Frame Number
SIR Signal to Interference Ratio
SRNC Serving RNC
TPC Transmit Power Command
TX Transmitted
UE User Equipment
UL Uplink
UTRA UMTS Terrestrial Radio Access

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document
 END OF DOCUMENT

Drivetel – Serviços e Projectos de Telecomunicações Lda

Confidential Document