Beruflich Dokumente
Kultur Dokumente
UM TS N e t w or k
Actix Education Services
April 2006
All contents of this document are the property of Actix and are provided for information
purposes only. The information in this document is subject to change without notice.
Actix will not be held liable for technical or editorial omissions made herein, and will not
be held liable for incidental, consequential or other similar damages resulting from the
use of its products.
Copyright Actix 2006. All Rights Reserved
Con t e n t s
INTRODUCTION ........................................................................................................................................ 5
SCOPE OF THE COURSE ................................................................................................................................. 5
WHO SHOULD ATTEND THIS TRAINING COURSE?.......................................................................................... 6
PREREQUISITES ............................................................................................................................................ 6
TOPICS TO BE COVERED ............................................................................................................................... 7
CHAPTER 1.................................................................................................................................................. 9
INTRODUCTION TO NETWORK DEPLOYMENT ............................................................................................... 9
CHAPTER 2................................................................................................................................................ 11
SCANNER THRESHOLDS, PREFERENCE & ATTRIBUTES .............................................................................. 11
UE THRESHOLDS, PREFERENCE & ATTRIBUTES ........................................................................................ 15
ATTRIBUTES .............................................................................................................................................. 23
EXERCISE 1: EXAM YOUR OWN DATA (ATTRIBUTES) ............................................................................... 31
CHAPTER 3................................................................................................................................................ 32
SITE AND CLUSTER INTEGRATION ............................................................................................................. 32
INTEGRATION PRE-REQUISITES ................................................................................................................. 33
CHAPTER 4................................................................................................................................................ 36
SITE INTEGRATION .................................................................................................................................... 36
POST PROCESSING AND ANALYSIS USING ACTIX SOFTWARE .................................................................... 38
EXAMPLE: HOW TO DETERMINE HARDWARE PROBLEM OR CROSSFEEDER DURING SITE INTEGRATION? . 39
EXERCISE 2: SITE INTEGRATION ................................................................................................................ 46
CHAPTER 5................................................................................................................................................ 47
CLUSTER INTEGRATION ............................................................................................................................. 47
FIRST DRIVE TEST: SC SCAN AND DEDICATED MODE ................................................................................ 49
IMPORTANT CRITERIA FOR CLUSTER INTEGRATION .................................................................................. 52
MISSING NEIGHBOURS ............................................................................................................................... 58
MISSING NEIGHBOURS ATTRIBUTES .......................................................................................................... 69
SCANNER BASED NETWORK ROLLOUT ANALYSES: NEIGHBOUR LIST ANALYSIS ..................................... 71
INTERFERENCE FACTOR (F FACTOR) ......................................................................................................... 74
F FACTOR REPORT ..................................................................................................................................... 77
EXAMPLE: COVERAGE ANALYSIS .............................................................................................................. 78
EXAMPLE: CPICH POLLUTION ANALYSIS ................................................................................................. 80
SECOND DRIVE TEST: DEDICATED MODE WITH SC SCANNING ................................................................. 83
TYPES OF DROPS (RADIO AND NON-RADIO RELATED):............................................................................. 86
EXAMPLE: DROPPED CALL ANALYSIS ....................................................................................................... 89
EXAMPLE: CALL SETUP FAILURE ANALYSIS ............................................................................................. 91
EXIT CRITERION DRIVE TEST .................................................................................................................... 93
EXERCISE 3: CLUSTER INTEGRATION ......................................................................................................... 96
SITE INTEGRATION IN INTEGRATED CLUSTER............................................................................................. 97
CHAPTER 6................................................................................................................................................ 98
CELL RESELECTION ................................................................................................................................... 98
RESELECTION CRITERIA ............................................................................................................................ 99
MEASUREMENT RULES ............................................................................................................................ 100
INTRA FREQUENCY SOFT HANDOVER ...................................................................................................... 101
HANDOVER ATTRIBUTES AND RELATED EVENTS .................................................................................... 105
Introduction
Scope of the course
Troubleshooting and Optimizing UMTS Network workshop provides a
comprehensive overview of the integration and radio optimisation procedure
for WCDMA networks making use of the Actix Analyzer. Our goal is to give
new users an understanding of what Analyzer can do and how it can help
them improve the quality of service for their network subscribers. Advanced
users will be introduced to Analyzer features they may not know about and
to new network applications that can be performed using Analyzer. The
instructors for this course are experts in optimising 3G networks and are
familiar with the challenges faced by engineers using measurement and postprocessing software.
Class size is limited to ensure that everyone receives extensive one on one
instruction. Participants are given ample opportunity to ask questions
relevant to their particular optimization needs. Upon completing the course,
participants will be able to use what they learned in class to identify and
solve real network problems.
Troubleshooting and Optimizing UMTS Network workshop begins with a highlevel presentation of the Analyzer interface and its capabilities. We will then
cover how Analyzer can be customized to tailor existing optimization
processes to the specific needs of the participants. We will try to establish
procedures and create Analyzer customized solutions for these procedures
Prerequisites
Familiarity with network infrastructure and operation of the air
interface.
A general understanding of drive test and call trace collection
and post-processing methodology and advantages/limitations
with each dataset.
Access to Actix software.
Basic computer literacy, including familiarity with Windows and
MS Office applications.
Optimal
Performance
Minimal
Performance
Processor:
P4 2.6 GHz
1 GB
256 MB
60+ GB Available
1 GB Available
1280x1024
64k colour resolution
1024x768
256 colour resolution
Memory (RAM)
Disk Space
Monitor
CD-ROM drive
One parallel port
Yes
Topics to be covered
Configuring your workspace, including map, cell site, and
parameter configuration in order to display your geographic
area and network cell plan.
Viewing data in a variety of ways to gain insight into possible
network problems.
Diagnosing and solving quality of service problems with drive
test and switch data.
Using pre-defined and custom analyses to analyze data and
devise tailored optimization solutions.
Chapter 1
Introduction to Network Deployment
The common practice for Network Deployment is shown below:
Link budget
establishment and
Nominal design
Deployment activities
(site acquisition, civil
works, installation etc)
Design Validation
Ongoing optimization
= Ongoing Optimization of
Growing and mature network
Site Integration
Site integration is performed to test how well Node B is functioning,
hardware (e.g. installation of aerials and cabling), and software
(e.g. correct parameters downloaded and connection with RNC
functions). A drive test can be conducted where scanner and trace
mobile are used to verify if node B is functioning well.
Cluster Integration
Cluster integration is performed to maximize the coverage in the
cluster with the available sites. When performing on an unloaded
network, cluster integration is coverage based. After the initial
tuning (cluster integration), ongoing optimization tasks, such as
Dedicated Drive Test (dedicated mode with SC scanning) will be
performed. The network deployment activities will deal with real
traffic KPI s, and trouble tickets will be raised if necessary.
Ongoing Optimization
The ongoing optimization is performed on the growing and mature
network to assure best quality of the network. Tracking and
defining the optimum KPIs will be the key tasks in fine tuning and
optimizing the network. If a new site or cluster is required, the site
and cluster integration phases will be repeated.
10
Chapter 2
Scanner Thresholds, Preference & Attributes
Thresholds are user-definable values that can be used as constants in event
detection and queries. For example: Uu_Scan_Poor_RSCP_Threshold
defines the minimum value which the RSCP signal is considered to be poor.
This threshold has a default value of -95 dBm, but users can change this to
any other value. Using a threshold (rather than a hard-coded value) in
queries and event diagrams makes it is easy to adjust them for individual
requirements.
Tools
Display Thresholds
UMTS
Scan_Coverage
Scanner Thresholds
Uu_Scan_PilotPollutionThreshold (Scanner Pilot Pollution)
Recommended value is -15 dB and value should vary between -10 and
-18 dB. By changing this threshold value, it will directly affect the
Uu_Scan_PilotPollution attribute.
Uu_Scan_TooManyServersThreshold (Scanner Too Many Server
Event)
Recommended value is 5 dB and value should vary between 1 and 10
dB. By changing this threshold value, it will directly affect the
Uu_Scan_TooManyServers attribute.
11
12
A sample set of scanner data for three SCs with individual colour and
vertical lines indicating transitions of pilots into and out of the Active
Set is shown below:
Scanner Preferences
The WCDMA options within the General Settings group in the Tool
Preferences dialog control the simulated active set feature. When this
feature is selected, Analyzer calculates a UMTS simulated active set from
the scanner data, as it envisages the handset would see it. You can
configure the simulated active set and specify the maximum size of the
simulated active set in the range 1 8. The simulated active set feature was
useful in the early days of WCDMA technology, when logging devices were
not readily available. Now that they are available, you may want to turn off
the feature.
13
The parameters depicted in the picture are based on the WCDMA events
defined in the 3GPP TR 25.922 V4.1.0 (2001- 09) Radio Resource
Management Strategies specification.
Max. Active Set Size (up to 8) specifies the maximum size of the Active
Set (1-8).
Disable active set simulation.
Select this check box to turn off the active set simulation feature. This will
usually give faster file loading times and will mean that the lines to cells
feature in the map will be based on the measured SC (if it's available)
rather than the simulated active set. This option is off (deselected) by
default.
14
Tools
Display Thresholds
UMTS
UE_Coverage
15
16
17
18
19
20
21
22
Attributes
It is important to differentiate between the scanner and UE attributes. Upon
loading a UMTS data file, several data groups are created for each device
contained in the file.
23
Example
Example below shows that measurements obtained from an uplink
measurement report might not be always correct. A measurement report
gives measurement information on all detected SC of the combined
neighbour list.
Screenshot below shows that the attribute is not giving the expected data.
On the measurement point the active set count is 6 (impossible).
24
Downlink Measurements
This group contains Uu interface
measurements reported by the mobile.
Pilot Measurements
This group contains signal
strength measurements for
individual SCs detected within the
logfile.
25
Uu_RRC
Radio Resource Control is the protocol
layer governing air interface
communication within the UMTS
network. Parameters associated with
protocol messaging can be found in this
group.
Event Data
Call events triggered by the current state
of the UMTS mobile. If an event is not
present in the tree, it did not occur in the
file. This group also contains events with
user-definable thresholds, such as
Uu_PilotPollution and
Uu_CoverageLimited.
Statistics Data
This group contains statistics calculated based on UMTS protocol messaging,
including Call Duration and timing components of Call Setup Time.
Vendor Specific
This group contains parameters
unique to a specific type of data
collection equipment.
Rake Finger Measurements This
group (Qualcomm chipset handsets
only) provides individual finger
measurements, including multipath measurements for the same
SC.
26
Other
This group contains scanner
measurements concerning RSSI, Chip
Offset, Carrier Error and the Scan
Code Group.
Nth Best
This node contains ranked
measurements for EcNo, RSCP, SC
and PathLoss. Within each group, the
0 element contains the best
performing data. For example,
CPICH_Scan_EcIo_SortedBy_EcNo_0
gives the strongest EcNo of all SCs
scanned.
Ec/Io
This group contains the interferenceto-chip-energy measurements
organized by various dimensions.
RSCP
This group contains the Received Signal Code Power organized by various
dimensions.
Delay
Find the CPICH Scan Delay Spread for each SC in this group.
27
Pilot Measurements
Find the CPICH PathLoss for each SC in
this group.
Active Set
This group contains simulated Active
Set measurements based on scanner
data and the user-defined WCDMA
settings in the
Tools
Preferences dialog.
Monitored Set
This group contains simulated Monitored Set measurements based on
scanner data and the user-defined WCDMA settings in the
Tools
Preferences dialog.
Event Data
Events in this group are based on the simulated Active Set and Monitored Set
SCs based on scanner data and user defined thresholds set under WCDMA in
the Tools
Preferences dialog.
28
GPS Data
This group contains mobile longitude, latitude,
distance traveled, and speed.
Message Info
The date and time for the start of the data stream can
be found in this group. This information is useful
when building report templates.
29
File Info
In this node find label and timestamp
information for the logfile under investigation.
Device Info
This group contains settings for the mobile
device on which data is logged.
Vendor Specific
Vendor Specific measurements may be grouped here or under the UMTS
node. In either case, this group provides measurements specific to the
particular collection device used.
30
31
Chapter 3
Detailed Radio
Design using
Planning Tool
Software
Performance
Engineering
and Ongoing
Optimization
Site Integration
Site integration is performed to test the well functioning of the Node B,
hardware (e.g. installation of aerials and cabling), and software (e.g.
correct parameters downloaded, connection with RNC functioning).
Cluster Integration
Cluster integration is performed to maximize the coverage in the cluster
with the available sites.
Cluster integration is coverage based and as such it is performed in an
unloaded network.
After cluster integration (initial tuning), ongoing optimization tasks will be
performed. These actions are based upon (real) traffic KPI s and eventually
trouble tickets.
If a new site is needed, the steps in site integration will be repeated.
32
a) Cluster definition
Cluster definition is performed at the stage of network design and according
to the rules and thresholds coming from the link budget analysis. In general
clusters will contain between 15 and 20 sites.
33
c) Data build
The RF part of the data build can be prepared just after the installation and
measurement of the radio chain of the site. In this stage the following data
needs to be prepared for each cell:
Scrambling Code (from planning tool )
Cell Id (from local rule)
Local Cell Id (can be equal to Cell Id)
SAC (service area code) (can be equal to Cell Id)
LAC (location area code)
RAC (routing area code)
CPICH power (set to 30dBm at TMA level per default)
Cable loss (from installation measurements)
34
35
Chapter 4
Site Integration
Purpose of Drive Test
When a single site or a new site is integrated, it is important to
complete the I&C tests (on-line commissioning tests mainly). This
verification shall be done by performing some drive tests around each
site clockwise and anti-clockwise at 25-35 percent of the expected cell
range (50m 100m in build up area). A scanner and a trace mobile are
used to verify if Node B is functioning well.
Drive Tests
Traces are collected from the scanner and the mobiles using any
measurement data collection tool.
First analysis is done live on the measurement trace tool. Post
processing is done with Analyzer by an optimization engineer.
The purpose of the site verification is NOT to propose optimization
changes for the aerial system or to propose new parameter settings.
At this stage, the neighbour planning is not complete and the cluster is
not completely build.
36
37
38
Example:
How to Determine Hardware Problem or Crossfeeder
During Site Integration?
The following example demonstrates how to determine hardware problem
or crossfeeder during site integration. After taking the measurements from
a site integration drive test, the optimization engineer will check the
measured data for the following conditions:
1. RF Condition
a) Active Set
By examining the Active Set attributes, you can find out the answers
of the following questions:
Are the correct Scrambling codes implemented on each cell
compared with the data fill values?
Are all intra-site SHO working, i.e. active set updates in both
directions: addition and deletion?
39
Good Practice: Screen layouts are very useful to have an immediate view of
how well a site is functioning. The screen shots from previous page shows:
Information to look for
Attribute
Uu_SimulatedActiveSet_EcIo_0;
Uu_SimulatedActiveSet_Count
(if you want to know the count)
Uu_ActiveSet_EcN0
UL Tx power
UE_TxPow
40
Attribute
Serving SC from UE
Uu_ActiveSet_SC_0
CPICH_Scan_RSCP_For_SC_xxx
CPICH_Scan_EcIo_SortedBy_EcIo_0
CPICH_Scan_RSCP_SortedBy_EcIo_0
Things to be considered:
Is the CPICH RSCP a normal value?
Should it be better than -70 dBm, close to the site (< 150m)
and outdoor?
Is the CPICH Ec/Io a normal value?
Should it be better than -8 dB, close to the site (< 150m) and
outdoor?
d) UE TX power
41
2. Call events
By displaying the Call Events attributes, such as outgoing call, drop call, and
call setup failures, you can obtain the following information:
Call Setup Status
PS Throughput
CrossFeeder
If issues occur, it is possibly come from the Node B (hardware) or from the
CS core side.
Information to look for
Attribute
Serving SC from UE
Uu_ActiveSet_SC_0
Uu_OutgoingCall_OK
Uu_CallSetup_Failure
Drop Calls
Uu_CallDropped
42
To check if a call was successfully done on each sector and soft handover
from one sector to another one was good, this is done using the message
logging information, checking if call setup and active set update were
successful in CS and in PS.
b) PS throughput
c) Crossed feeders
A typical problem that can be, and must be detected at this stage is the
crossedfeeder problem.
A SC (scrambling code) plot with the designated SC per cell indicated on
the antenna is sufficient to detect the problem.
43
3. MHA malfunction
A MHA (TMA) malfunction will cause a very high uplink UE Tx Pwr on one cell.
If the high TX power occurs on all cells of a Node B, it is more likely the
Digital radio modulation (CE) unit has a malfunction. In the example below
you can see this last phenomenon.
A simple verification on the UL and DL coupling loss gives you an indication
of the correctly applied UL Tx Pwr:
Formulae:
Coupling Loss Downlink (CL DL)
CL DL = CPICH Tx Pwr at AE (30 dBm)
CPICH RSCP
|CL UL
For most measurement points over the service area of the cell, the
MHA functionality needs to be checked. This criteria can be checked
with the Analyzer tool.
44
Example
The map below shows you the result of UE TxPower = 0 based on the
criteria mentioned in the previous page. This means there is no problem.
45
46
Chapter 5
Cluster Integration
Procedure
47
48
Drive Test 1:
Unloaded scan test
& Dedicated CS
49
50
DU Area
Urban
SU Deep
Indoor
SU Indoor
Daylight
-82 dBm
-82 dBm
-83 dBm
-90 dBm
-8 dB
-8 dB
-8 dB
-8 dB
4 best CPICH
RSCP <
Best CPICH
RSCP -8dB
Best CPICH
RSCP -8dB
best CPICH
RSCP -8dB
best CPICH
RSCP -8dB
-71 dBm
-71 dBm
-72 dBm
-79 dBm
th
51
52
The map above is a combination of different drive tests in the same cluster.
The attribute CPICH_Scan_RSCP_Sortedby_EcIo_0 is dragged and
dropped different times on the map view.
Best practice:
The better solution is to create a super stream of all files from one cluster
and perform the analysis on the super stream.
In this map, we only have problems of UL coverage outside the cluster and in
the centre of the 3 sites. The next necessary step is to find out where we
have the dominant servers. We can do this by displaying the
CPICH_Scan_SC_Sortedby_EcIo_0 attribute on the map to find out the
dominance areas.
53
54
Not much problem for the example above, only some degraded downlink
quality at the edge of the cluster. This is due to edge of coverage.
55
2 x SH O_ w in dow
All measurement points fulfilling the criterion above suffer from pilot pollution
and optimization changes need to be undertaken. The possible solution for
pilot pollution can be:
Up tilt the antenna providing the best server in the area in order to
improve the CPICH RSCP
Down tilt the antennas from interfering cells and as such reduce the
RSSI.
Change azimuth from antennas in order to create a dominant server in
the polluted area.
In the following graph you can detect an area of pilot pollution. The next step
in this case would be to detect which of the surrounding sites is covering the
area and to try to limit the amount of servers through up- and down tilting.
That comes down to creating 1 or 2 dominant servers.
56
In the example above, there is only some coverage overlap on area were the
sites are built very close to each other. By applying the appropriate tilt and
azimuth, you can avoid this problem.
57
M issing neighbours
Missing Neighbour Analysis is a fundamental aspect to WCDMA network
optimisation. It should be performed as part of the rollout phase of a network
to construct and/or optimise the adjacency lists of each cell; and throughout
the life of the network as changes to traffic and tilt affect the coverage
pattern and interference interaction between the cells.
Missing Neighbour Analysis in Actix RVS solution has been designed to allow
engineers to generate a list of suggested Missing Neighbours and allow
engineers to visualise the suggested cells on a map. Full control of the
suggestion criteria is included to allow for changes in clutter, terrain and
rollout phase of the network. These suggestion criteria are described as
follows:
58
Criteria Default
Value
Description
59
S
At point X , the strongest signal is SC 103 is at -5dBm.
Applying the Reporting Range of 5dB creates a short-list of SC 202,
302, 203, 402 and 301 (greater than -10dBm). The signals from SC
101 and 303 are too weak to be considered.
Note: The scanner reports SC numbers, so Actix RVS selects the closest cell to the drive
route at that time, with a matching SC.
1
60
Example of angle to site threshold for validation of SC 203, 302 and 302
61
Our final list of Missing Neighbours is then 301 and 302. Due to the fact that
SC 302 has a stronger EcIo value than SC 301 at point X , the Missing
Neighbours at point X are as follows:
Missing_Neighbours_0 = 302
Missing_Neighbours_1 = 301
..relative to CPICH_Scan_SC_SortedBy_EcIo_0 = 103 at point X .
This is repeated for all points in the scanned drivetest route, creating an
array of SC s at each point, sorted by their EcIo value, which satisfy the
selection criteria and are not currently defined as a neighbour of the
Nth_Best_0 cell at that time.
If the Only use Nth Best cell option is disabled, not only the Nth_Best cell will
be considered. Using the principle that if cell A and cell B are within 5dB, and
cell A and cell C are within 5dB, the relationship between cell B and cell C
should be analysed too. This obviously increases where processing is needed
at each point, and is recommended for neighbour list optimisation of more
mature networks.
Once the final list of suggested Missing Neighbours has been compiled, the
results can be viewed in two ways: through the Accelerated Network
Rollout Analysis pack report (Neighbour List Recommendations), and
also by visualising them on the map/chart/table.
62
63
The lists are sorted by signal strength and contain the SC values of the cells
in the list at each point in the scanned drivetest. The strongest Missing
Neighbour (Missing_Neighbour_0) should be analysed first, as it could
potentially be the strongest pollution source at that time.
Drag the parameter onto the map and open the Top 10 Scan Measurements
stateform (optionally dock it to under the Workspace Explorer). The legend of
the map gives you the list of SCs that were the strongest Missing Neighbour
at any point in the drive. The most frequent SC should be targeted first, and
by selecting any point in the map, the stateform will synchronise to show the
top 10 scanned measurements giving the Nth_Best SC at the top of the list2.
A decision needs to be made whether to add this suggested Missing
Neighbour SC into the adjacency list of the Nth_Best SC.
In the case of repeated SC values, the closest cell with a matching SC to the Nth_Best_0 cell
will be selected.
64
Once all the strongest Missing Neighbours have been analysed, the next
strongest signal should be analysed. Following this process with ensure that
the most important Missing Neighbours are analysed first.
The same SC could appear as a Missing Neighbour for multiple Nth_Best
cells, so it may be easier to create a Crosstab query with two dimensions:
Missing_Neighbours_0 and Nth_Best_0, to give all occurrences of the
combinations. This list can then be filtered from the Statistics Explorer to
filter the map view for single SC combinations individually.
65
66
This requires the scanner and UE traces to be collected at the same time,
and for the resulting collection files to be superstreamed together (and
correcting any time offset between the collection devices). See the online
Help for instructions on superstreaming using Time Offsets. Once the files
have been combined, parameters such as the UTRA_UE_CarrierRSSI and
the Uu_CallDropped event can be dragged onto the map. By opening the
UE Missing Neighbours stateform for the superstream and synchronising it
with dropped call, the top 10 scanned measurements can be visually
compared to the Active and Monitored sets for the UE at the time.
67
Screenshot showing how dropped calls from the handset can be viewed with
the Missing Neighbour attributes when the scanner and handset are
synchronised. Here the Active Set and Monitored Set SC s are at very low
RSCP values, and poor EcNo too. The scanner detected 3 other SCs, and a
missing neighbour just before the drop.
This allows the engineer to understand whether the UE had dropped a call
due to a missing neighbour not being defined in the cell s adjacency list, or
whether it was a handset problem, resulting in a neighbour that had already
been defined not being added to the Monitored Set, and therefore never
being able to be added to the Active Set.
68
69
70
71
C
Neighbour 2
In the figure above, Cell A is the best server by CPICH Ec/Io. Cells B and C
are within a user-specified threshold of Cell A's are Ec/Io, and so are counted
as potential neighbors of A. Cell D is not within the required threshold and
so is not counted as a prospective neighbor, nor is Cell E which did not have
a measurable signal contribution at this point in the drive test.
Here is a sample symmetric prospective neighbor array using sector IDs A, B,
C and D:
A
B
C
D
N/A
10
2
15
10
N/A
40
0
2
40
N/A
12
15
0
12
N/A
72
UMTS Accelerated
73
2 3 .1
2 6 .7
3 3 .0
40
H6 5 V 6 T0
H6 5 V 6 T2
2 .6
2 .4
2 .0
3 .4
3 .1
2 .6
4 .3
4 .0
3 .5
5 .7
5 .1
4 .5
7 .2
6 .7
6 .0
10
6 .6
6 .1
4 .6
H6 5 V 6 T4
9 .8
9 .6
8 .9
20
8 .2
7 .9
7 .3
1 6 .7
1 6 .5
1 6 .5
30
1 2 .2
1 1 .9
1 1 .0
F_DL cov e r ag e p e r ce n t ag e
50
>
2
a)
=
tr
<
L
in
(i
te
in
te
ii
ii
tr
a)
<
a)
tr
FD
(i
e
g
ra
ve
<
Co
1,
g
1,
=
1,
4
n
ii
/
r
te
L
FD
(i
L
FD
<
ra
ve
ra
g
Co
ve
=
a)
<
=
<
a)
tr
in
te
ii
tr
ii
/
r
te
in
1,
<
FD
(i
in
L
FD
<
4
Co
Co
ve
ra
1,
2
1,
e
1,
1,
=
<
a)
n
tr
n
ii
/
r
te
(i
in
(i
L
<
1
e
g
Co
ve
ra
ra
ve
1,
8
<
a)
tr
n
ii
/
in
(i
L
FD
FD
<
8
0,
Co
0,
=
<
=
tr
n
ii
/
r
te
r
te
in
(i
L
e
g
ra
ve
Co
a)
<
a)
tr
n
ii
/
r
FD
<
6
0,
e
g
Co
ve
ra
ra
ve
Co
te
n
(i
i
L
FD
4<
0,
e
g
g
ra
0,
4
0,
2
0,
=
<
=
<
a)
a)
tr
n
tr
ii
n
/
ii
/
te
in
te
(i
in
L
(i
FD
L
FD
2<
0,
0<
e
g
ra
ve
ve
Co
Co
The graph above shows the F factor on a map and indicates areas
where we have too much interference. In this example, F factor is
smaller than 2 covers 95% of the area.
74
75
6.7541x
3
2
2
F
Expon. (F)
1
1
0
0.40
0.60
0.80
1.00
1.20
Fraction DL PL
76
F Factor Report
7
-5
-10
EcIo
F factor
4
-15
3
2
-20
1
-25
-88.5
-91.5
-94.5
-97.5
-100.5
-103.5
-106.5
-109.5
-112.5
-115.5
-118.5
-121.5
-124.5
-127.5
-130.5
-133.5
-136.5
-139.5
-142.5
-145.5
-148.5
-151.5
-154.5
-157.5
-160.5
-163.5
-166.5
-169.5
-172.5
-175.5
-178.5
F Factor Report
F_Factor Report.xls
77
Attributes
Uu_ActiveSet_RSCP_0
Uu_CallDropped
Uu_IncomingCallSetupFail
Uu_OutgoingCallSetupFail
Uu_Trch_DownlinkBlerAgg
CPICH_AllSets_UE_RSCP_For_SC
Stateforms
UMTS UE Active Set + Monitor Set
Queries
Poor Mobile Receive Power, or
Low Mobile Receive Power
(Condition: Uu_ActiveSet_RSCP_0 < -95dBm)
Screen Layout
UMTS Coverage Analysis
Flirst click on the stream name, then click on
Layouts UMTS
UMTS Coverage Analysis
78
Action
Superstream logfiles.
Find and investigate further area with poor coverage and drop
calls.
79
80
A
Active Set
In the example above, Cells A, B and C are part of the Active Set, as
determined by the Simulated Active Set module. Cell D has a CPICH Ec/Io
within a user-specified pollution threshold of the Active Set s best server
Ec/Io, and so is counted as a contributer to pilot pollution at this point in the
drive test. Cell E has a CPICH Ec/Io that is not within this threshold and so is
not a pollution source.
Shown below is a sample pollution array indicating the number of points at
which each sector caused pilot pollution for sector IDs A, B, C and D:
Sector ID
A
Pollution
Count
0
150
45
12
81
Perform the following steps to access the Pilot Pollution Analysis report:
Step
Action
1
UMTS Accelerated
As shown below, the Pilot Pollution Analysis Report indicates the worst
interferers sorted by Scrambling Code:
82
Drive Test 2:
Dedicated Mode with
SC Scanning
83
-2 dBm
-2 dBm
-2 dBm
1%
1%
1%
-2 dBm
-2 dBm
-2 dBm
1%
1%
1%
Apart from the coverage criteria, the dropped calls will be analyzed.
The cluster integration is successful when no dropped calls are due
to radio quality. Note that this is only applicable when the cluster is
fully built. If radio drops are a consequence of a missing site (and the
site is in the final design), they cannot be a blocking point for cluster
integration at this point.
Analyzing dropped calls to determine the cause can be done using
Analyzer, which enables to see decoded RRC messages. A dropped call
usually occurs during an abnormal behaviour in one of the following
procedures:
call setup CS and PS
call clearing
soft handoff
If a call drops the messaging before the drop should be checked and
compared with the normal messaging sequences.
84
In order to find the cause for a dropped call, we should approach the
problem in a structured way and always look at the different possible
causes in the same way. Actix Software has many ways to endorse
this procedure, and one of them is displaying the results on the
Statistic Explorer. The sequence to check were the cause for dropping
is can be as follows:
Import and run Drop Analysis Crosstab Query to obtain the
information mention below:
If Yes, check if
Out of coverage DL?
High F factor?
Due to load
Failed handover?
Missing neighbour?
Out of coverage UL
b) DL BLER
Information to look for
If Yes
DL channel lost
c) RRC Messages
Information to look for
If Yes
Check cause
Measurement reports
Radio synchronization
85
86
The example shows a drop due to low CPICH RSCP. Further analysis
will show that on this point, the situation could be avoided by
optimising the cell with SC100.
In this case there is no other way to optimize this area than adding a
new site, as no other antenna change will improve the situation. Else
low RSCP can be improve by up tilt, and low Ec/Io by giving only one
sector dominant in the area where the drop occurs.
3. UE TX power
No example from pre-launch. This will hardly ever happen since the
measurements are taken out door, while our design is done for indoor
coverage. The only possible occurrence of this problem is where we
drive outside 3G coverage areas. This will only happen during the exit
criterion drive test.
87
The reason of call drop in the example above is not very clear, but the BLER
goes up to 100% and the serving cells are not the obvious ones. A downlink
power limitation is probably the reason for this.
1. Mobile Issue
CM service abort, pegged as a drop, but coming from a cancellation on
mobile side
2. Core Network Issue
Disconnect received as there are a temporary unavailability from the
CS side.
88
Attributes
Uu_ActiveSet_SC_0
Uu_IncomingCallOK
Uu_OutgoingCallOK
Uu_CallDropped
CPICH_AllSets_UE_RSCP_For_SC
Stateform
UMTS UE Active + Monitor Set
Queries
UMTS Call Statistics query shows which File/Sector/Cell/Call ID that
contains the drop call. Which one should we look at?
Dropped Call Analysis query shows event and what causes the drop
89
Action
Superstream logfiles.
Display Drop Analysis crosstab query or drop call related query
on Statistic Explorer to obtain the following information:
Drop Call
RSSI
BLER
Exam and further investigate the result (i.e: Check out which File
contains the drop call).
If necessary, select a data point shortly before the start of the call,
and use the Replay Tool to step toward the drop.
direction.
10
11
12
90
Attributes
Uu_IncomingCallSetupFail
Uu_OutgoingCallSetupFail
Uu_ActiveSet_SC_0
Uu_ActiveSet_EcNo_0
UE_TxPow
Stateform
N/A
Queries
UMTS Call Setup Statistic, or
UMTS Call Setup Failure Analysis
Application Packs
UMTS Call Setup Analysis
91
Action
Superstream logfiles.
Exam and further investigate the result (i.e: Check out which File
contains the call setup failures).
In the same Application Pack, run the Call Setup Status report to
find out the reasons why the call failed.
Display Uu_InComingCallSetupFail,
Uu_OutgoingCallSetupFail, Uu_ActiveSet_RSCP_0, and
Uu_ActiveSet_EcNo_0 on the chart.
10
11
12
13
92
93
Drive Tests
For best practise, it is recommended to perform two different drive
tests using the same pre-defined route:
Drive test 1:
Within 3G coverage, focus on CR (setup, maintain and clear)
Within the 3G coverage area, with existing scenario (in-car,
route density etc) suitable
o
o
Metrics
o UE1 - DL Throughput (average), for successful calls
o DL CR (call success rate, = Call setup, maintained and
cleared)
o UE2 - CR (call success rate, = Call setup, maintained and
cleared)
Targets
o
o
Drive test 2
o UE 1: CS voice short call: 10 sec call and release with 5
sec guard
o UE 2: PS file download: 10 sec call and release with 5 sec
guard
o The following statistics are taken from the drive tests:
CS call setup success rate
PS call setup success rate
94
95
96
97
Chapter 6
Cell Reselection
What is cell reselection?
Cell reselection is the procedure where the UE reselects the cell from
which he reads the BCCH. This happens in idle mode, but also in cell
FACH and cell/URA PCH mode.
In idle mode, the cell reselection can be detected through the decoding
of a new BCCH. In cell PCH and cell FACH state, the reselection can be
noticed by an RRC cell update message. The transitions from those
states to idle mode and back can be detected by the following RRC
messages:
Physical channel reconfiguration
Transport channel reconfiguration
RRC connection release
98
Reselection Criteria
Cell is suitable if this condition is fulfilled:
S- Criterion for FDD cells:
Srxlev > 0 AND Squal > 0
Where:
Description
Cell Selection quality value (dB)
Applicable only for FDD cells.
Srxlev
Qqualmeas
Qrxlevmeas
Qqualmin
Qrxlevmin
Pcompensation
max(UE_TXPWR_MAX_RACH
UE_TXPWR_MAX_RA_CH
P_MAX
P_MAX, 0) (dB)
99
In order to perform cell reselection UE, measures and ranks the neighbour
cells. For each type of neighbour cells (Intra freq, inter-freq, Inter-RAT)
thresholds are definable. Measurements of neighbour cells will be triggered if
these thresholds are reached.
(Re-) Selected cell is a suitable cell (fulfils the S criterion) and is the best
ranked cell (has the highest R). The UE shall however reselect the new cell,
only if the following conditions are met:
The new cell is better ranked than the serving cell during a time
interval Treselection
More than 1 second has elapsed since the UE has camped on the
current serving cell
In the below mentioned ranking criteria the following naming convention
applies:
s indicates the serving cell
n indicates the candidate cell
Measurement Rules
If Squal > Sintrasearch,
UE need not perform intra-frequency measurements.
If Squal <= Sintrasearch,
perform intra-frequency measurements.
If S intrasearch,
is not sent for serving cell, perform intra-frequency measurements
Reselect highest ranked if access not restricted: R- Criterion
R s = Qmeas,s + Qhysts
R n = Qmeas,n Qoffsets,n
100
You can obtain the Intra-Freq handover information by displaying the UMTS
Handover Analysis (Intra Freq) screen layout from
Layouts
UMTS
UMTS Handover Analysis (Intra Freq)
101
H andover Events
The following events and thresholds are in use in the WCDMA handover
algorithm:
Radio Link Addition
Event 1A
If Pilot Ec/Io > Best Pilot Ec/Io reporting range + hysteresis 1A for a
period of t and the active set is not full, the cell is added to the active
set.
Radio Link Removal
Event 1B
If Pilot Ec/Io < Best Pilot Ec/Io reporting range hysteresis 1B for a
period of t, then the cell is removed from the active set.
Combined Radio Link Addition and Removal
Event 1C
If the active set is full and Best candidate Pilot Ec/Io > Worst Pilot Ec/Io +
hysteresis 1C for a period of t, then the weakest cell in the active cell is
replaced by the strongest candidate cell.
These are the basic, standardised handover events and they are UE triggered
event. As for other, more complicated handover algorithms will be vendor
dependant and you can make use of the available measurement report data
to obtain the information you need.
102
UE Measurement reports
Depending on the handover type (MEHO or NEHO), different measurement
reporting criteria can be used. These measurement reports give the
necessary information to the RNC (NEHO) or UE (MEHO) to trigger a
handover, but they also allow the RF engineer to understand the behaviour of
the UE/network in certain conditions (load, location). Measurements done by
the UE are collectable during drive tests, whereas Node B measurements
have to be found in the Iub traces.
In a UE report, the following Ies are included to identify the neighbours; PCPICH information, UTRA RF channel number (for IF measurement reports),
BCCH freq (for GSM IS measurement report).
Inter- Frequency
They are always periodic and the way of using the measurements in
handover algorithms is not part of the standard.
103
UE Internal measurements
The UE internal measurements can be divided in two groups:
measurements to indicate the status of the mobile TX power
measurements to indicate the UE TX RX time difference
The first group of measurements can be used by the RNC to trigger interRNC or system handover. The second group is used to adjust the downlink
DPCH air interface timing, when the difference ion time between the UE
uplink DPCCH/DPDCH frame transmission and the first significant path of the
downlink DPCH frame from a measured active set cell becomes too large.
Event 6A: UE TX power becomes larger than an absolute threshold
Event 6B: UE TX power becomes less than an absolute threshold
Event 6C: UE TX power reaches its minimum value (-50 dBm)
Event 6D: UE TX power reaches its maximum value (21 dBm)
Event 6E: UE RSSI reaches the UE s dynamic receiver range
Event 6F: UE Rx Tx time difference for a radio link included in the
active set becomes larger than an absolute threshold
Event 6G: UE Rx Tx time difference for a radio link included in the
active set becomes less than an absolute threshold
104
105
Uu_IRAT_Reselection3G2G, Uu_IRAT_Reselection2G3G
(IRAT Reselection)
If a device is determined to be idle, and system information from the
other technology is seen then there will be an event to indicate there has
been a technology change.
Uu_CellReselection (Cell Reselection)
If a device is determined to be in Cell Fach and there is a server cell
change, then there will be a event to indicate there has been a Cell
reselection.
106
107
Handoff Analysis
The Handoff Analysis module uses the Simulated Active Set to determine the
handoff state at each point along a drive test. Statistics on handoff state
may then be calculated and presented in a report format. Excessive handoff
state reduces capacity and increases infrastructure costs for a given traffic
level. Please see the Simulated Active Set section for more details on how
the Active Set is estimated based on WCDMA scanner measurements.
The handoff state algorithm has the following components:
The Active Set of pilots is determined using the Simulated Active Set
module.
Using the geographic information in the log file and the SC, the
network element database is searched to identify the Sector and Cell
IDs of the SC.
Handoff state is calculated by determining the configuration of the
sectors in the Active Set.
All bins in the log file are then processed into the handoff state
matrix.
Reports showing the percentage of handoff state for each sector and for the
total drive test may then be calculated.
108
The Handoff State Analysis examines Sector IDs involved in call at a given
drive test point and determines which of the above states applies, based on
UMTS scanner data:
UMTS Accelerated
109
110
Exercise 4 : Handover
Exam Your Own Data
Load logfile and select handover scenario
Display on graph the quality indicators per cell
Display handover triggers
How can they be influenced (which parameters)
Study missing neighbours
Do we automatically add missing neighbours to the neighbour list?
111
Chapter 7
Report
Creating Reports
You may want to create high-level management reports based on your
collected data. You can print from map and chart windows, but the Workbook
function (using Microsoft Excel ) allows you to collate these views and
perform other data analysis functions to produce an integrated, polished
report.
Actix makes the creation of user-definable report templates effortless, by
allowing users to have a direct access to MS Excel. The easiest way to get
data into reports is to use the right-mouse selection Display on Workbook,
which automatically displays three types of datasets in Excel: series data,
statistical data, and histogram data.
Note: Series data always takes the longest to re-generate as a template.
112
Procedure
Step
Action
113
Step
4
Action
Keep the workbook open. Switch back to Actix Software and
display the Uu_CallDropped (from UMTS
Event Data
Call
data group) on the existing workbook (active workbook).
Notice how all three views from these two attributes have been added
to the workbook (Series, Histogram and S.tatistic Formatted Data
tabs).
5
Save your query into the workspace and check the results are
accurate.
10
11
12
These tabs will contain the results from whatever file you run the
report template on. Open Sheet 1 and create your KPI statistics
here as follows (see next page):
114
From Statistics
results
From Crosstab
results
Step
Action
13
Now go back into the all other sheets and delete the raw data
from row 2 downwards. This cleans the report template so that it
can be run on any other file, of varying size.
14
15
Open another log file and run the report template on this file as
you did in step 10. Compare the results.
115
Key
Word
Search
Result
116
Attribute
Definition
Selected
Attribute
Highlighted Item in the Workspace and the Corresponding Entry in Attribute Help
117
Technical Support
Technical support is available by phone, email, and on the Actix Web site.
Technical personnel are located in London UK, USA, and Singapore to provide
assistance. In the EMEA technical support can be reached at:
Phone
: +44 (02) 8735 6300
Email
: EMEAsupport@actix.com or support@actix.com
Web (Extranet): www.myactix.com
MyActix (Extranet)
If you would like to obtain the latest information of the software release,
problem reporting, support issues, access to our self-study modules or
Knowledge Base, please register yourself at http://www.myActix.com
118
Appendix A
Event Data
Call Events
Triggering Sequence
(1)
Uu_OutgoingCallOK
(1)
(*)
(*)
(*)
(*)
(*)
(*)
(*)
(*)
(*)
(1)
(*)
The messages with the symbol star are usually present but not
mandatory
Uu_IncomingCallOK
119
Triggering Sequence
*TerminatingConversationalCall
*TerminatingStreamingCall
*TerminatingInteractiveCall
*TerminatingBackgroundCall
Uu_RRC_MsgType == RRC Connection Setup
(2)
(2)
(*)
(*)
(*)
(*)
(*)
(*)
(*)
(*)
(*)
(1)
(*)
The messages with the symbol star are usually present but not
mandatory
Uu_OutgoingCallSetupFail
120
Triggering Sequence
Uu_RRC_MsgType == PagingType1 with
Uu_RRC_PagingRecord_cn_Identity_pagingCause equals any of
the following:
*TerminatingConversationalCall
*TerminatingStreamingCall
*TerminatingInteractiveCall
*TerminatingBackgroundCall
Uu_RRC_MsgType == RRC Connection Request with
Uu_RRC_RRCConnectionRequest_establishmentCause equals
any of the following:
*TerminatingConversationalCall
*TerminatingStreamingCall
*TerminatingInteractiveCall
*TerminatingBackgroundCall
Then any of the following options:
*Uu_RRC_MsgType == RRC Connection Reject or
*Uu_RRC_MsgType == RRC Connection Setup
*Uu_RRC_MsgType == RRC Connection Release or
*Any BCCH messages during the call setup
Uu_CallCompleted
When in Call (Outgoing Call Ok or Incoming Call Ok), you get one
of the following messages:
*GSM_Um_Msg_Type == CC Disconnect or
*GSM_Um_Msg_Type == CC Release Complete or
*GSM_Um_Msg_Type == CC Release
And any of the above messages with a normal cause for ending
the call (CauseCodeCC is equal or less than 31)
Uu_CallDropped
When in Call (Outgoing Call Ok or Incoming Call Ok), you get any
of the following options:
Any BCCH Message or
Uu_RRC_MsgType == RRC Connection Release AND Release
Cause is not Normal or
One of the following messages:
*(GSM_Um_Msg_Type == CC Disconnect) OR
*(GSM_Um_Msg_Type == CC Release Complete) OR
*(GSM_Um_Msg_Type == CC Release)
*AND any of the above messages with NOT a normal cause for
ending the call (CauseCodeCC is greater than 31)
Uu_LocationUpdateOK
Uu_LocationUpdateFail
121
Triggering Sequence
GSM_Um_Msg_Type == MM Location Updating Reject
Event Data
RRC Events
Triggering Sequence
Uu_RRC_MsgType == RRC Connection Request with
Uu_RRC_RRCConnectionRequest_establishmentCause equals
any of the following:
*RRC_OriginatingConversationalCall
*RRC_OriginatingStreamingCall
*RRC_OriginatingInteractiveCall
*RRC_OriginatingBackgroundCall
*RRC_OriginatingSubscribedTrafficCall
*RRC_Registration
Uu_RRC_MsgType == RRC Connection Setup
Uu_RRC_MsgType == RRC Connection Setup Complete
Uu_IncomingRRC_ConnectionOK
122
Event Data
RAB Events
Triggering Sequence
Uu_RadioBearerSetupOK
Uu_RadioBearerSetupFail
Uu_RadioBearerReleaseOK
Uu_RadioBearerReleaseFail
Event Data
Coverage Events
All Coverage Events which make use of thresholds may be customized by the
user in the Tools
Display Thresholds menu.
Coverage Event Name
Triggering Condition(s)
Uu_PilotPollution
Uu_TooManyServers
Uu_System_Interference
Uu_Coverage_Limited
Uu_PoorDL_Coverage
Uu_PoorUL_Coverage
123
Triggering Condition(s)
Uu_HandoverProblem
This event monitors the Ec/No of Monitored Set SCs with the Ec/No of
Active Set SCs in an 8 second window prior to each dropped call. The
Uu_HandoverProblem event is pegged if the Monitored Set SCs are
stronger than the Active Set SCs more often than the Active Set SCs are
stronger than the Monited Set SCs.
Uu_MissingNeighbor
SC in Active Set after drop is not in Active Set before drop and
SC in Active Set after drop is not in Neighbor List before drop
Event Data
Handoff Events
Triggering Sequence
ActiveSetUpdate message (Uu_RRC_MsgType == ActiveSetUpdate)
ActiveSetUpdateComplete message (Uu_RRC_MsgType ==
ActiveSetUpdateComplete)
Uu_HandoffFail
124