LTE Optimization Analyze

Contents

• Part 1 Basic Cell Parameter Planning

• Part 2 LTE Call Drop Optimization
• Part 3 LTE Cell Reselection Optimization
• Part 4 LTE Coverage Optimization
• Part 5 LTE Handover Optimization
Access Problem due to PRACH not Planned
(1)
Symptom:
In a site test, the UE under one cell can make a call in the areas opposite to the antenna,
but it is difficult to access or even unable to access the eNodeB in the areas that other
cells can also cover. The RSRP and CINR meet the access requirements.

Analysis:
1. After the other two cells of the eNodeB are disabled, the UE can access the eNodeB if
the signals meet the access requirements.
2. After all three cells are enabled, access is difficult or unable in the overlapped
coverage areas of the three cells. The UE sends the RRC Connection Request message,
but the system does not respond.
3. Signaling on the eNodeB side shows the system receives the message and responds,
but the response is not received by the UE.
4. Check results of the access parameter configuration show the parameter “Logical
root sequence start number used to generate prach preamble” of the three cells is set
to 0, and the parameter “NCS used to generate prach preamble” is set to 11. This
means the access preambles of the three cells sent to the UE are the same. When the UE
is in the overlapped coverage area, the three cells receive the message from the UE and
returns the MSG4 message at the same time. The UE receive two MSG4 messages and
cannot demodulate them, so the access attempt fails.
Access Problem due to PRACH not Planned
(2)
Solution:
Reset the PRACH parameters, including “Logical root sequence start
number used to generate prach preamble”, and “NCS used to generate
prach preamble”. After the resetting, the UE can access the system in any
coverage area of the eNodeB so long as the signals are of good quality.

Summary:
Prior to the kickoff of a project, it is required to plan access-related
parameters in addition to PCI and neighbor cells.
UE Unable to Access due to Incorrect
Parameter Configuration (1)
Symptom:
In the pre-preliminary acceptance test of PGP, it is found under the cell whose
PCI is 50, UEs cannot access the SMV eNodeB and several access failures take
place in one minute. However, the radio signal quality is found good (RSRP = -
104 dBm, CINR = 102 dB).
UE Unable to Access due to Incorrect
Parameter Configuration (2)
Analysis:
The UE should report the Status packet to the eNodeB after it receives the LTE DL
AM All PDU message.
However, the analysis of UE logs shows the eNodeB does not receive the Status
packet. As a result, the eNodeB repeats sending the message in the RLC layer.
On the UE side, the UE has received the message 4 times, while on the eNodeB side,
the RNLU sends the RnluRnlcErrInd message to the RNLC, and the RNLC returns the
RrcConnectionRelease message, causing the UE’s failure to access.
By default, our network parameter setting allows SRB signaling to be repeatedly
transmitted for up to 16 times.
Check the parameter setting and find the maximum RLC AM retransmission time
was incorrectly set to 4.
Meanwhile, the SRB RLC Type of the eNodeB is also incorrectly configured as UM
(Unacknowledged Mode), which should be AM (Acknowledged Mode).

Solution:
Modify the SRB RLC Type to be AM and the problem is solved.
Access Failure due to a UE Fault (1)
Symptom:
In the pre-preliminary acceptance test of PGP, it is found Samsung UEs fail
to work from time to time, leading to access failure.
Access Failure due to a UE Fault (2)
Analysis:
First, the UE logs are analyzed. The radio environment is good as the RSRP
is -82 dBm and the CINR is 19 dB. But after the UE receives the “Activate
default EPS bearer context request” message from the MME, it rejects
the request through the “ulInformationTransfer” message on the uplink
and returns the reason “PTI unmatched”.
However, the comparison of the signaling elements in the normal
procedure with those in the abnormal procedure shows the PTI
(Procedure transaction identity) sent by the MME is the same as the PTI
carried in the attach request message of the UE.
Therefore, it is concluded that the problem is caused by Samsung UE’s
failing to maintain the PTI properly.

Solution:
Update the HW/SW version of Samsung UE or replace Samsung UE with
other UEs, for example, Qualcomm UEs do not report such a problem.
Contents

• Part 1 Basic Cell Parameter Planning

• Part 2 LTE Call Drop Optimization
• Part 3 LTE Cell Reselection Optimization
• Part 4 LTE Coverage Optimization
• Part 5 LTE Handover Optimization
Call Establishment and Call Drop

• NAS and AS
• NAS: Non-Access Stratum
• AS: Access Stratum

2. Idle/ EPC
1. Power-off Attached 3. Activated
Registered Connected

Evolved Mobility Deregistered Registered

Management (EMM)
Evolved Connection
Idle Connected
Management (ECM)
Radio Resource Control
(RRC) Idle Connected Idle Connected
Call Establishment and Call Drop

• Connection Release
UE EUTRAN

RRCConnectionRelease

• Call Drop – Definition

• A call will be dropped when an UE exits the RRC_CONNECTED state.
Call Establishment and Call Drop

• Call Drop – Common Causes

1. Connection Reestablishment Reject (Signaling)
(1). The UE sends the rrcConnectionReestablishmentRequest; Cause = otherFailure
message through the UL-CCCH.
(2). The eNodeB returns the rrcConnectionReestablishmentReject message through the
DL-CCCH.
(3). When experiencing a call drop problem, the UE starts to receive the broadcast
message that contains SIB1 through the BCCH-SCH until the UE originates the next call.
Call Establishment and Call Drop

• Call Drop – Common Causes

• 2. Poor Signal over the Air Interface (Signaling)
• Incomplete signaling: Without receiving the Connection Release message, the UE is
transitioned from the RRC-CONNECTED state to the RRC-IDLE state.
• A typical symptom of this call drop problem: The UE sends the
RRCConnectionReestablishmentRequest message but does not received the
RRCConnectionReestablishment message, and even the UE does not send the
RRCConnectionReestablishmentComplete message.
Call Establishment and Call Drop

• Call Drop – Common Causes

• 3. Other Causes
• Narrowly speaking, a call drop problem occurs only when the UE sends the
rrcConnectionReestablishmentRequest message, that is to say, the RRC connection is
broken.
• If the RRC connection is reestablished due to a handover or any other problem, the RRC
reestablishment is usually successful. In a real project, you need to pay special attention
to such problem. Whenever necessary, you must reach an agreement with the customer
on whether regarding such a problem as a call drop problem.
Cause Analysis 1 – Poor Coverage

• Problem Symptoms
Poor
Coverage

20

10
Serving Ce
0 ll CINR

-10 Drop

-70

-90
Serving
Cell RSR
-110 P

-130
Cause Analysis 1 – Poor Coverage

• Cause Analysis: Analyzing the Drive Test Data

• Step 1. Collect the drive test data.
• Step 2. Analyze the drive test data using ZXPOS CNA or TEMS Discovery.
• Step 3. Check whether there is any of the following symptoms through the
geographical chart or Table View when a call drop problem occurs:
• Low RSRP (serving cell): RSRP < -105dBm
• Low CINR (serving cell): CINR < 0dB
• Poor signal strength (neighboring cell): Signal strength > -105dBm
Cause Analysis 1 – Poor Coverage

• Optimization Solutions
• 1. Determine which sectors are serving current poor coverage areas.
• 2. Determine the most suitable sectors based on the network topology and
radio environment, and strengthen the coverage of this area.
• Troubleshoot the hardware faults from the baseband, RF, and antenna system, VSWR
alarms.
• Increase the RS power of the serving cell.
• Increase the power of the serving sector.
• Adjust the antenna tilt of the serving sector.
• Adjust the antenna azimuth of the serving sector.
• Add some sites plus adjusting the antenna tilt and azimuth of the surrounding base
stations.
• 3. Implement the Coverage & Capacity Optimization (SON-CCO) function
(pending).
Case Study 1 – Poor Coverage

• Poor Coverage

RSRP < -120 dBm

Serving Cell CINR < -2 dB
Cause Analysis 2 – Handover Failure

• Problem Symptoms
• Before a call drop problem occurs, the UE has ever sent the measurement
report and received the RRCConnectionReconfiguration message from the
eNodeB.
• The UE sends the rrcConnectionReestablishmentRequest; Cause =
handoverFailure message immediately when receiving the broadcast
message from the target cell.
• The UE usually sends the rrcConnectionReestablishmentRequest message to
the source cell after experiencing a handover failure.
Cause Analysis 2 – Handover Failure

• Cause Analysis: Analyzing the Signaling Data

• Obtain the signaling data collected from the drive test.
• Check whether there is any of the following symptoms when a call drop
problem occurs:
• Before a call drop problem occurs, the UE has ever sent the measurementReport
message.
• The UE can receive the rrcConnectionReconfiguration message that contains the
MobilityControlInfo field from the eNodeB.
• When handover to the target cell that is carried by the rrcConnectionReconfiguration
message, the UE can receive the systemInformationBlockType1 message through the
BCCH-SCH.
• When receiving the systemInformationBlockType1 message, the UE sends the
rrcConnectionReestablishmentRequest Cause=handoverFailure message.
• Usually, the UE can successfully reestablish the connection within 200 ms, and handover
back to the source cell.
Cause Analysis 2 – Handover Failure

• Optimization Solutions
• Check whether the neighboring cells are properly configured by comparing the
Neighbor Cell worksheet of the source cell to the Serving Cell worksheet of the
target cell.
• Check whether the target cell is normally running.
• No bit errors during the transmission
• Normal power output
• No handover failure due to the cell overload
• Check whether the software versions of both source and target cells are correct.
• Familiarize yourself with common handover failures.
• A handover failure occurs because the X2 interface is improperly configured.
• A handover failure occurs because traffic load is burdened in a single cell.
• A handover failure occurs because the handover success rate is too low in this cell.
• A handover failure occurs because a new site is deployed in the surrounding area.
• A handover failure occurs because the UE is located at the edges of different MMEs.
• A handover failure occurs because the UE is located at the borders of base stations in different
frequencies.
Case Study 2 – Handover Failure

• Inter-Frequency Handover Unsupported

Cause Analysis 3 – Missing Neighbor
Missing
• Problem Symptoms
Neighbor

20

10
Servin
g Cell R
0 CINR N1 CIN
Drop
-10

-70

-90 Serving C
e N1 RSRP
ll RSRP

-110

-130
Cause Analysis 3 – Missing Neighbor

• Cause Analysis ： Analyzing the Signaling Data

• Obtain the signaling data collected from the drive test.
• Check whether there is any of the following symptoms when a call drop
problem occurs:
• Before a call drop problem occurs, the RSRP of the serving cell is continually decreased.
• Before a call drop problem occurs, the UE sends the measurementReport message
consecutively. The purpose is to check whether the signals of neighboring cells are
strong enough. The signal of the serving cell becomes poorer and poorer because no
neighboring cell is configured, and eventually this causes a call drop problem.
• The measurementReport message sent by the UE contains the target neighboring cells
that satisfy the A3 or A5 event, depending on system setup.
• The system (neighboring cell) message sent by the serving cell does not contain the
target neighboring cell of the measurementReport message sent by the UE.
• After sending the measurementReport message, the UE does not receive the
rrcConnectionReconfigurationRequest message with a handover instruction from the
eNodeB.
Cause Analysis 3 – Missing Neighbor

• Optimization Solutions
• Using the configuration tool or batch-import function in the OMC, add the
missing neighbor cells to the serving cell list prior to the call drop.
• Implement the ANR function to optimize the configuration of neighboring
cells. (pending)
Case Study 3 – Missing neighboring Cell

• The UE sends the measurementReport message for several times

before a call drop problem occurs.
Cause Analysis 4 - Overshooting

• Problem Symptoms
Overshooting
(Pilot Pollution)

20

10

0
Serving Serving
Cell CINR Cell CINR
-10 Drop Drop
N1 CINR

-70

-90
Serving Serving Servin Serving
Cell g
-110 1 RSRP Cell 2 Cell 2 Cell 1
RSRP RS RP RSRP
-130
Cause Analysis 4 - Overshooting

• Cause Analysis: Analyzing the Signaling Data

• Obtain the signaling data collected from the drive test.
• Check whether there is any of the following symptoms when a call drop
problem occurs:
• The signals in a coverage area where a call drop problem occurs, serving cell, or
searchable neighboring cell contains some overshooting signals (overshooting 3 or more
layers).
• Determine whether the coverage area where a call drop problem occurs is in nature a
pilot polluted area, which is served by more than three cells (RSRP > -110 dBm, CINR < 0
dB).
• Determine whether any neighboring cell is unconfigured. Check whether the
neighboring cell list of the coverage area contains any overshooting cell.
Cause Analysis 4 - Overshooting

• Optimization Solutions
• The general principle is: on the condition that the coverage area is provided
with stable and proper signals, try as best as possible to control the signals of
the overshooting area.
• Decreasing the power of the overshooting area
• Increasing the antenna downtilt of the overshooting area
• By taking the border coverage of the overshooting area as well as the network topology
into consideration, carefully adjust the antenna azimuth of the overshooting area.
• Pilot Pollution Caused by Overshooting Coverage
• Increase the power of the primary sector
• Adjust the antenna downtilt of the primary sector
• Adjust the antenna azimuth of the primary sector
• Control the polluted signals from other areas
• Neighbor Cell Unconfigured: Optimizing neighbor cells
Cause Analysis 5 – Equipment Fault

• Problem Symptoms
• Equipment Faults
• Handover abnormalities
• A call drop problem occurs and can reoccur when a service is processed at a fixed time
or place.
• A call drop problem occurs and can reoccur in one or more specific sectors of the
eNodeB.
• A call drop problem occurs and can reoccur in an inter-MME, inter-TA, or any other
special area.
Cause Analysis 5 – Equipment Fault

• Cause Analysis: Analyzing the Drive Test Data and OMC Statistics Data
• Step 1. Collect the required data.
• Step 2. Analyze the data before and after the call drop problem occurs.
• Radio environment (GE): poor coverage, fast fading, shadowing
• Coverage of the serving cell (RSRP and SINR)
• Neighboring cell configuration and handover statistics
• Signaling flow abnormalities
• Step 3. Eliminate the causes one by one and sum up some helpful principles
• Whole or part of a site?
• S1 or X2 interface?
• Affected by any critical operations?
Cause Analysis 5 – Equipment Fault

• Optimization Solution
• If a call drop problem is caused by any equipment fault, you need to submit it
to an R&D engineer.
• When necessary, you must capture the required information for further
analyzing the call drop problem, in coordination with the R&D engineer.
• Follow up the problem troubleshooting procedure.
• Eventually, verify that the problem is properly eliminated.
Cause Analysis 6 – Interference

• ProblemInterference
Symptoms (Uplink Interference)
(Uplink)

20 UE Tx Power
Drop
10 Serving Cell CINR

-10

-70 eNodeB RSSI

-90
Serving Cell RSRP
-110

-130
Cause Analysis 6 – Interference

• Problem Symptoms (Downlink Interference)

Interference
(Downlink)

20

10

0
Drop
Serving Cell and Neighbor Cell CINR
-10

-70
Serving Cell & Neighbor Cell RSRP
-90

-110

-130
Cause Analysis 6 – Interference

• Cause Analysis ： Drive Test Data Analysis + OMC Dynamic Data

Analysis
• Step 1. Collect the drive test data and dynamically observe the RSSI data in
the OMC when a call drop problem occurs.
• Step 2. Analyze the data features when a call drop problem occurs:
• Check whether the RSSI of the BTS is too high (such as -85 dBm or higher). If yes, you can
infer that uplink interference is present.
• Check whether the transmit power of the UE is greater than 20 dBm within a few
minutes prior to the call drop occurrence. At this time, the UE is not located in any
weakly covered area. In this case, you can infer that uplink interference is present.
• Check whether the RSRP is -90 dBm or better but the CINR is smaller than 0 dB in the
serving cell (even including neighbor cells) that is being tested. If yes, you can infer that
downlink interference is present.
• Step 3. Determine which type of interference is present, and then handle it
accordingly.
Cause Analysis 6 – Interference

• Optimization Solution
• Determine and troubleshoot uplink interference:
• Specify the rough range of the interference (which cells experiencing interference and
whether these cells are included in a certain cluster)
• Using a spectrum scanner such as YBT250 plus Yagi antenna to pinpoint the interference
source.
• Determine and troubleshoot downlink interference:
• Confirm that downlink interference does not come from any internal system (This
requires eliminating the overshooting coverage and unconfigured neighbor cell
problems).
• If downlink interference comes from an external system, you need to use a spectrum
scanner such as YBT250 plus Yagi antenna to pinpoint the interference source.
Cause Analysis 7 – Traffic Congestion

• Problem Symptoms
• There are relatively too many real-time activated subscribers in the cell.
• The cell starts to reject the UEs during the admission control procedure.
• The transmit power of the cell is becoming almost saturated.
• The call setup success rate and the call drop rate are becoming worse.
Cause Analysis 7 – Traffic Congestion

• Cause Analysis: OMC Performance Statistics Analysis

• Step 1. Collect performance statistics in busy hours from the OMC, including
call setup, handover, and call release.
• Step 2. Query the number of subscribers as well as the traffic load when a
call drop problem occurs, and then check whether the cell is heavily loaded.
• Step 3. Query the cell setup success rate, handover success rate, and call
drop rate; and locate their potential causes.
• Step 4. When a cell starts to reject the UEs during the admission control
procedure because of resource insufficiency, you can infer that the call drop
problem is caused by traffic congestion.
Cause Analysis 7 – Traffic Congestion

• Optimization Solutions
• Increase the system capacity.
• Increase the cell power.
• Compress the overhead channel power.
• Increase the resource blocks (RBs).
• Add some base stations.
• Change the network topology and balance the traffic load.
• Shrink the coverage and adjust the RF for mitigating the power overload
• Implement the SON-CCO function (pending).
Contents

• Part 1 Basic Cell Parameter Planning

• Part 2 LTE Call Drop Optimization
• Part 3 LTE Cell Reselection Optimization
• Part 4 LTE Coverage Optimization
• Part 5 LTE Handover Optimization
The S-criterion for Cell Selection (Srxlev >
0)
• Srxlev = Qrxlevmeas – (Qrxlevmin + Qrxlevminoffset) – Pcompensation
• Qrxlevmeas is the reference signal received power (RSRP) of the candidate cell.
• Qrxlevmin and Qrxlevminoffset are the minimum received levels for cell
camping configured in the SIB (SIB1). The Qrxlevminoffset is used only in the
situation when the UE camps on a VPLMN cell.
• Pcompensation uses a larger value among 0 and the value of PEMAX (the peak
uplink transmit power of the UE) deducting PUMAX (the maximum RF output
power of the UE), that is MAX （ PEMAX － PUMAX, 0). The unit is dB.
• Actually, it is the compensation of UE’s transmit power and the maximum allowed transmit
power to the cell selection. (If the transmit power of UE is larger than the maximum allowed
transmit power, the Pcompensation is 0; otherwise it is necessary to use the transmit power of
UE to compensate the access threshold.)
Cell Reselection Process

• The process of cell reselection

includes these steps: conducting
measurement as per
measurement rules, reselecting
a cell as per cell reselection
rules, and cell access
verification.
• After the UE has camped on a cell,
it will conduct measurement to
determine whether there is a
better cell so as to conduct cell
reselection.
• The priority-based cell reselection
is a new mechanism in the E-
UTRAN.
Cell Reselection Measurement Rules

• To reduce power consumption, less measurement operations are preferred. The

3GPP Specifications define two thresholds for triggering the UE’s measurement.
They are s-IntraSearch and s-NonIntraSearch, which are all configured in SIB3.
• If the quality of the serving cell is better than the s-IntraSearch, the UE can conduct no intra-
frequency measurement.
• If the quality of the serving cell is better than the s-NonIntraSearch but worse than the s-
IntraSearch, the UE will conduct the intra-frequency measurement, but not the inter-frequency or
inter-RAT measurement.
• If the quality of the serving cell is worse than the s-NonIntraSearch, the UE will conduct the intra-
frequency, inter-frequency, and inter-RAT measurements.
• During the early stage of a commercial network, the UEs are basically data cards,
and there is no need to consider the power consumption. As a result, the
configuration for s-IntraSearch and s-NonIntraSearch are not available. The UEs
are required to conduct the intra-frequency, inter-frequency, and inter-RAT
measurements.
Evaluation and Cell Sequencing 1

• According to the priorities of neighbor cells and the serving cell, the
cell The
reselection takes place in three
cell with lower situations (see next
Serving page):
The cell with higher
priority frequency Cell priority frequency
Evaluation and Cell Sequencing 2

• The priority of the neighbor cell is higher than that of the serving cell.
• The signal quality (SnonServingCell,x) of the measured neighbor is higher
than ThreshX,High in the time of Treselection.
• The priority of the neighbor cell is lower than that of the serving cell.
• If the above two situations are not satisfied, the cell is reselected when the
quality of the serving cell is lower than Threshserving, low, and the
measurement quality SnonServingCell,x of a neighbor cell with lower priority
is higher than Threshx, low.
Evaluation and Cell Sequencing 2

• The priority of the neighbor cell is equal to that of the serving cell.
• All cells (including the serving cell) are put in a queue. If a cell ranks highest in the queue and
lasts longer than the time set in Treselection, the cell is selected. The R-criterion is shown
below:
Rs = Qmeas,s + QHyst
Rn = Qmeas,n - Qoffset

• Rs is the quality score of serving cell (ranking criterion for serving cell); and Rn is the quality
score of the neighbor cell.
• Qmeas,s and Qmeas,n are the RSRP values of the serving cell and the neighbor cell
respectively.
• Q_Hyst is the parameter to control the hysteresis of the serving cell in the sequence.
• Q_offset defines the offset for a neighbor cell. For the intra-frequency cells, it is the cell-
based offset value. For the inter-frequency cells with the same priority, it includes two parts:
the cell-based offset value and the frequency-based offset value.
Accessibility Verification for Cell
Reselection
• If the best cell in the LTE frequency range is barred or reserved, it is
required that this cell is removed from the cell reselection candidate
list on the UE.
• The UE can try other cell of the same frequency, unless the cell indicates that
the UE can not select any intra-frequency cell in a period of time.
• If the UE evaluates the best target cell and finds that it is not suitable
due to some reasons.
• For examples, the tracking area prohibits roaming, or the cell is not a cell in
the registered PLMN.
• In a period of time (which would not be longer than 300s) that follows, the
UE will not take any cell in this frequency as the target cell for cell
reselection.
Cell Reselection Speed Factor 1

• By calculating the frequency of cell reselections, the UE determines

which mobility state it is in. In SIB3, the mobility state parameters are
configured, including n-CellChangeMedium, n-CellChangeHigh, t-
HystNormal and t-Evaluation.
• In the period of t-Evaluation, if the count of cell reselections is larger than n-
CellChangeHigh, the UE determines it is in a high speed mobility state.
• If the count of cell reselections is larger than n-CellChangeMedium, the UE
determines it is in a medium speed mobility state.
• In the period of t-HystNormal (t-HystNormal must be larger than t-
Evaluation), the UE will return to the normal mobility state if it does not
detect any condition that meets the medium speed mobility state or high
speed mobility state.
Cell Reselection Speed Factor 2

• In the high speed mobility state and medium speed mobility state,
the UE needs to obtain and uses the sf-Medium and sf-High
configured in t-ReselectionEUTRA-SF and q-HystSF from SIB3. Of
which, sf-Medium is the factor used for medium speed mobility state,
and sf-High is the factor used for high speed mobility state.
• Multiply the corresponding factors of Treselection and t-ReselectionEUTRA-
SF, and use it as the new timer for cell reselection.
• Add the sf-Medium or sf-High in q-HystSF to Qhyst. The Qhyst is used in the
R-criterion.
Black / White List, Access Restriction, Cell
States
• The UE should bypass the cells on the black list when it conducts cell selection,
measurement and cell reselection.
• Blacklisted cells are not considered in event evaluation or measurement reporting.

• For the CSG cells and Hybrid cells, the UE should obtain the CSG IDs from the
system information. If the CSG ID is stored on the CSG whitelist of the UE, the UE
should put the cell in the candidate list and conduct measurement and
sequencing
• Each UE belongs to an access class (AC) in the range 0 - 9.
• Some UEs belongs to one or more high-priority ACs in the range 11-15, which are reserved for
specific uses. AC10 is used for emergency access.
• The UE considers access to be barred if access is barred for all its applicable ACs.

• When the cellBarred state of a cell in SIB1 is barred, the UE can not select this
cell in cell selection nor cell reselection, even in emergency.
Late Measurement

• Measurement threshold: s-IntraSearch and s-NonIntraSearch. Due to

improper configuration (too small value), the cell reselection
measurement is started late. The conditions to start measurement
are ready, but no measurement is started.
• The serving cell does not provide configurations for s-IntraSearch (affecting
intra-frequency measurement) or s-NonIntraSearch (affecting inter-frequency
or inter-RAT measurement). The UE conducts the intra-frequency, inter-
frequency and inter-RAT measurements.
• Increase the value of s-IntraSearch or s-NonIntraSearch so that the cell
reselection measurement is started as soon as the quality of the serving cell
worsens.
Ping-Pong Reselection

• If ping-pong cell reselections occur, you can increase the cell

reselection hysteresis, and raise the priority of LTE cell reselection.
• If the above method does not work, you can decrease the value of
Threshserving, low for the serving cell.
• For the same operator, the priority settings for UTRAN and E-UTRAN
in the UMTS and LTE networks should be the same. The cell
reselection parameter configurations on the two sides must match
each other so as to prevent the ping-pong handover.
Late Cell Reselection

• If the cell reselection is conducted too late, the UE may originate calls
on an unsuitable cell, which easily causes access failure or call drops.
• Take an intra-frequency cell reselection for example. To speed up the
cell reselection, the following methods can be used:
• Set the value of cell reselection Treselection as small as possible. At present,
we recommend setting it to 1s.
• Decrease the value of Qhyst.
• Decrease the value of Qoffset.
Call Congestion

• If subscribers suffer call congestions in the overlapped coverage area,

you should decrease the threshold for cell reselection to other cells,
no matter such call congestions occur in which network, so as to
properly distribute traffic.
 Case Study 1
Improper measurement control threshold setting leading to access failure
Fault Description
During a commissioning test, we found that the UE failed to originate calls in a
cell, but it could originate calls in a neighbor cell with better quality soon
(hundreds of milliseconds).

Time FD Frame KeyInfo Appearance PCI RSRP CINR RSRQ

16:07:35.421 FD12 LTE RRC Signaling UL-CCCH: rrcConnectionRequest; Cause = mo-Signalling Fail
16:07:36.500 FD12 LTE RRC Signaling DL-CCCH: rrcConnectionSetup
16:07:36.515 FD12 LTE NAS Signaling UL EPS MM: Attach request
16:07:36.515 FD12 LTE RRC Signaling UL-DCCH: rrcConnectionSetupComplete
16:07:36.515 FD12 LTE NAS Signaling UL EPS SM: PDN connectivity request 307 -109 -5 -22.2
16:07:36.515 FD12 LTE RRC Signaling DL-CCCH: rrcConnectionSetup
16:07:36.640 FD12 LTE NAS Signaling DL EPS MM: Authentication request
16:07:36.640 FD12 LTE RRC Signaling DL-DCCH: dlInformationTransfer
16:07:36.984 FD12 LTE NAS Signaling UL EPS MM: Authentication response
16:07:36.984 FD12 LTE RRC Signaling UL-DCCH: ulInformationTransfer
16:07:38.375 FD12 LTE RRC Signaling BCCH-SCH: systemInformationBlockType1
16:07:38.375 FD12 LTE RRC Signaling UL-CCCH: rrcConnectionRequest; Cause = mo-Signalling Success
16:07:38.406 FD12 LTE RRC Signaling UL EPS MM: Attach request
16:07:38.437 FD12 LTE NAS Signaling DL-CCCH: rrcConnectionSetup
16:07:38.437 FD12 LTE RRC Signaling UL EPS SM: PDN connectivity request
16:07:38.437 FD12 LTE NAS Signaling UL-DCCH: rrcConnectionSetupComplete 397 -87 13
16:07:46.421 FD12 LTE RRC Signaling DL EPS MM: Identity request
16:07:56.515 FD12 LTE RRC Signaling DL-DCCH: dlInformationTransfer
Case Study 1
Fault Analysis
We checked the UE log and found that the radio environment of the cell
that the UE accesses for the first time was poor (PCI=307, RSRP=-109dBm,
CINR=-5dB). There was a neighbor cell with good signal quality (PCI=397,
RSRP=-87dBm), but the UE did not select it. The situation was that the UE
failed to access the cell with PCI=307, but it soon successfully accessed
the cell with PCI=397. We checked the SIB3 messages from the network,
and found s-IntraSearch=7 in the intraFreqCellReselectionInfo section.
According the calculation formula, the serving cell originate intra-
frequency measurement when the RSRP value is smaller than -130+14=-
116dBm. Then, it is too late.

Solution
In current network environment, the terminals
are data cards, which have no requirement on
power consumption. So, we can set the
parameter bySIntraSrchPre to false, ensuring that
the terminal conducts RSRP measurement for
intra-frequency neighbor cells all the time.
Case Study
Improper 2 parameter setting leading to call drops
cell reselection

Fault Description
During a preliminary acceptance test for PGP, the signal quality of the serving cell
worsened gradually in the process of UE access, leading to RRC call drop. The cell
reselection parameters were improperly configured, and the UE originated calls on
the cell with poor quality signals, leading to RRC call drops.
Case Study 2

Fault Analysis
We checked the UE log and found that the radio environment of the cell
that the UE accesses was poor (RSRP=-92dBm, CINR=1dB). There was a
neighbor cell with good signal quality (PCI=251 ， RSRP=-87.6dBm), but
the UE did not select it. During the access process, the UE could not be
handed over to a better cell, the signal quality of the serving cell became
worse and worse, and at last the call dropped. When the time requirement
(1s, the minimum value) for intra-frequency cell reselection is met, the cell
reselection process can be initiated according to the R-criterion.
Furthermore, decreasing the value of Qoffset or Qhyst can accelerate cell
reselection.

Solution
Modify the cell reselection parameter Qhyst (3dB -> 1dB) to speed up cell
reselection so that the UE can select the cell with better quality.
Contents

• Part 1 Basic Cell Parameter Planning

• Part 2 LTE Call Drop Optimization
• Part 3 LTE Cell Reselection Optimization
• Part 4 LTE Coverage Optimization
• Part 5 LTE Handover Optimization
Trans-Regional Coverage

• The trans-regional coverage usually refers to the coverage of some

base stations exceeds the planned scope. Such problems are usually
the following solution:
1. Reduce the power of cross-district coverage area;
2. Reduce the antenna down tilt;
3. Adjusting the antenna angle;
4. Reduce the antenna height;
5. Replace the antenna. Use small-gain antenna. Mechanical downtilt
antenna replaced by electronic downtilt antenna. Lobe width is narrower
beam antenna replacement antenna lobe;
6. If the site is too high causing more area coverage, is invalid in the case of
other means, you can consider adjusting the network topology, the
relocation site is too high.
No dominant cell

• Such an area is no dominant or leading cell, replace cell too often.

This will result in frequent handover, which reduces system efficiency,
increasing the likelihood of dropped calls.
• Area for lead-free areas should be under the antenna by adjusting
the angle of inclination and direction of the other methods, and
enhance a strong signal area (or close-quarters) of coverage, weaken
the other weak signal area (or remote area) coverage.
Interference Analysis

1. Downlink interference analysis

• Scanner DT test by analyzing the received SINR to target.
• If RSRP good coverage SINR falls below a certain threshold, but there may be
downward interference. Deterioration of regional identity out of the SINR,
check the deterioration of the region downstream RSRP coverage. If the
difference is down RSRP coverage also identified as coverage issues, coverage
issues in the analysis to be addressed. SINR for the RSRP good and bad
situation recognized as the downlink interference problems because of
interference and to resolve them.
2. Uplink interference
• Uplink interference problem judges by checking noise floor of site. If the
noise floor of a cell is too high, and there is no equivalent high-traffic is
present, confirm the presence of uplink interference problems, analysis
interference and resolve.
Analysis of handover

• Neighbor list optimization focus on neighbor list missing. Neighbor

list missing will lead to dropped calls. Through the drive test data
analysis software and statistical analysis, for each district to provide
neighbor list added, deleted, reservations recommended.
• By adjusting azimuth and downtilt of the antenna, we change
position of handover area and signal distribution. If the handover
area is too small, can solved by reducing downtilt or adjust the
azimuth. If you handover zone signal changes too frequently,
appropriate azimuth and downtilt adjustments could be considered
to ensure a smooth change of a single cell.
RF common method of optimization

• Antenna azimuth adjustment

• Antenna downtilt adjustment
• Neighbor list / PCI adjustment
• Basic wireless parameters verification
• Adjust the antenna height
• Adjust the antenna location
• Adjust the antenna connector
• to use particular antenna
• Adjust the accessories, such as the tower amplifier
• Modified downlink power
 A ntenna downtilt calculation formula

• Antennas downtilt formula 1

θ = atan (2H / L) * 360 / (2 * p) + b / 2 - e_γ
• Antennas downtilt formula 2
θ = atan (H / L) * 360 / (2 * p) - e_γ
Where: θ indicates initial antenna mechanical downtilt; H indicates site effective hei
ght; L indicates that the distance between antenna and cell coverage edge; b repres
ents the vertical lobe angle; e_γ indicates electronic downtilt.
A ntenna downtilt calculation formula in
struction
1. Formula 1 main scenarios: Under urban dense sites, for the most energy
of the antenna radiation in the coverage area can reduce the interference
of neighboring cells, when setting the initial downtilt, the antenna's half-
power point above the main lobe aim at coverage edge (defined as L / 2).
Generally not recommended to set initial downtilt according to the formu
la, avoid the initial inclination may be set too high, which lead to network
coverage problems, and more as reference of optimization.
2. Formula 2 is the general formula, the main scenarios: in the suburbs, villa
ges, roads, sea and so far as to cover as much as possible, can reduce the
initial downtilt, so that a maximum gain of the antenna main lobe is align
ed on cell edge.
A ntenna downtilt calculation formula in
struction
3. The actual wireless network optimization, optimization of the antenna se
t up downtilt depends mainly on the analysis of drive test data to comple
te.
• According to the SINR of the pilot overlay can draw the coverage of each sector, the sector fo
r more area coverage, consider increasing the antenna down tilt, as required under the angle
to increase the amount of formula 1 can be used to refer to As RF optimization of empirical s
trong, many people in this setting is also under the angle and the surrounding environment.
• For the more serious areas covered sectors, optimizing the antenna set may be greater than
under the dip calculated under the formula of a much larger angle.
• For the coverage is insufficient, or contaminated areas to be used in the pilot frequency opti
mal use of the leading sectors do, as long as more areas will not be the optimal coverage, the
antenna is less than under the dip can be calculated by the formula 1 under the angle, Even l
ess than calculated by the formula 2 under the angle.
Downlink power allocation principle

• power allocation principle defined by 3GPP 36.213

• For the OFDM symbol does not contain the RS,
• Definite = -PDSCH-to-RS
 power offset  PA
EPRE ratio
A
= [dB]
 power -offset
• Among them, the only effective multi-user MIMO, and the re
maining values are 0dB format. PA parameter configured
UE-specific
by the RRC , its range is [3,2,1,0, -1.77, -3, -4.77, -6] dB.
Downlink power allocation principle
B B
• For the OFDM symbol that contains RS
• Definition = PDSCH-to-RS EPRE ratio, B /  A
PB the value of which provide
d by the following table , PB and the value of the number of ant
enna port. is high-level configuration parameter, is cell specific.

B /  A

PB Two and Four Antenna

One Antenna Port
Ports
0 1 5/4
1 4/5 1
2 3/5 3/4
3 2/5 1/2
 Downlink power allocation principle

• Within the OFDM in the presence of RS, and does not contain within
the OFDM symbol, respectively, the following equation holds.

Single antenna port (  B /  A * 5) and two, four antenna port (  B /  A *

4) .The two values are equal, so the value of P has nothing to
CRS _ RE

do with the antenna ports.

 Downlink power allocation principle
• For example, 20M-bandwidth, coupled with the maximum symbol power 20w assumpti
on (43dBm). We can estimate the range of reference signal power. Calculated on the ba
B /  A
sis of above rule, the combination of the various possible values ,
and is,

PA PCRS _ RE
（ dBm ）
(dB) Type B
Type A

3 9.586 10.457 11.549 13.010 9.208

2 10.494 11.343 12.400 13.799 10.218
1 11.370 12.193 13.208 14.537 11.218
0 12.218 13.010 13.979 15.228 12.218
-1.77 13.638 14.363 15.234 16.325 13.988
-3 14.559 15.228 16.020 16.990 15.228
-4.77 15.739 16.319 16.989 17.781 16.988
-6 16.478 16.989 17.570 18.239 18.239
Average() 13.010 13.737 14.618 15.738 13.413
RS power parameter configuration

• RS (Cell-specific Reference Signals Power): This parameter indicates the cell

reference signal power (absolute value). Reference signal is used for the cell
search, downlink channel estimation, channel detection, directly impact on
cell coverage. This parameter is broadcast by SIB2 notice UE, and in the
whole system bandwidth and the downlink sub-frames is all constant, unless
SIB2 message updates (such as the RS-power enhancement).
• RS parameters configuration instructions: downlink power settings are based
on reference signal power, so the reference signal power settings and
change, affecting the entire downlink power setting. RS power too large will
cause pilot pollution, and inter-cell interference; too small will cause cell
selection or re-selection failure, the data channel cannot demodulation.
Cell Transmit Power parameter configuration

• Cell Transmit Power parameter configuration

• Cell actually used transmission power (Cell Transmit Power): This
parameter indicates the actual use of cell transmit power.
• Cell Transmit Power parameter configuration instructions: depends
on network planning and coverage is expected to determine the
maximum cell transmit power, both to ensure effective coverage and
avoid cross-district coverage. This parameter is the cell total multi-
antenna transmit power. 43dBm corresponds to 20W, 46dBm
corresponds to 40W.
Contents

• Part 1 Basic Cell Parameter Planning

• Part 2 LTE Call Drop Optimization
• Part 3 LTE Cell Reselection Optimization
• Part 4 LTE Coverage Optimization
• Part 5 LTE Handover Optimization
 A3 事件的进入条
件为：

Radio Parameters – A3-triggered Reporting

Event A3 entering condition: Mn  Ofn  Ocn  Hys  Ms  Ofs  Ocs  Off

• Event A3 related parameters
leaving condition: Mn  Ofn  Ocn  Hys  Ms  Ofs  Ocs  Off
Of which,
Mn: The measurement result of the neighbouring cell, not taking into account any offsets
Ofn: The frequency specific offset of the frequency of the neighbour cell
Ocn: The cell specific offset of the neighbour cell, which is set to zero if not configured for the
neighbour cell
Hys: The hysteresis parameter for this event
Ms: The measurement result of the serving cell, not taking into account any offsets
Ofs: The frequency specific offset of the serving frequency
Ocs: The cell specific offset of the serving cell, which is set to zero if not configured for the
serving cell
Off: The offset parameter for this event
If the values of Ofs and Ocs of the servicing cell are configured the same as the values of
Ofn and Ocn of the neighbor cell during the network planning, the entering condition can
simplified as
 A3 事件的进入条
件为：

A3 事件的离开条

Radio Parameters – A3 Threshold 件为：

• a3-Offset
• Off in the entering/leaving condition is a3-Offset
• Its default value is 3 dB
• The recommend value is 2, 3 or 4 dB. A greater or smaller value is not
recommended because it may lead to delayed handover and too many
handovers at a time
• This offset is applicable to all neighbor cells
 A3 事件的进入条
件为：

A3 事件的离开条

Radio Parameters – TimeToTrigger 件为：

• TimeToTrigger (TTT)
• This parameter specifies the value range used for time to trigger parameter,
which concerns the time during which specific criteria for the event needs to
be met in order to trigger a measurement report.
• Value range: enumerated (0, 40, 64, 80, 100, 128, 160, 256, 320, 480, 512,
640, 1024, 1280, 2560, 5120)
• Default value: 320 ms 。
• A high TTT can put off measurement reporting and reduce handovers.
• Recommended values: 256, 320, 480, 512 or 640 ms.
 A3 事件的进入条
件为：

A3 事件的离开条

Radio Parameters – Hysteresis 件为：

• Hysteresis
• Hysteresis, used to adjust the entering/leaving threshold of event-triggered
reporting, determines the period the UE stays at Event A3.

• Default value is 0 。
• If ReportOnLeave is not enabled ， the recommended value is 0 。
 A3 事件的进入条
件为：

A3 事件的离开条

Radio Parameters – ReportOnLeave 件为：

• ReportOnLeave
• This parameter indicates whether or not the UE shall initiate the
measurement reporting procedure when the leaving condition is met.
Mn  Ofn  Ocn  Hys  Ms  Ofs  Ocs  Off

• Default value is False (no reporting)

 A3 事件的进入条
件为：

A3 事件的离开条

Radio Parameters – CIO 件为：

• Cell Individual Offset (CIO)

Event 
•Mn A3Ofn condiction
entering Ocn  Hys  Ms  Ofs  Ocs  Off

• Ocn in the formula is CIO

• Enumerated (-24,-22,-20, -18, -16, -14, -12, -10, -8, -6, -5, -4, -3, -2, -1, 0, 1, 2,
3, 4, 5, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24), unit: dB
• To quicken handover from the serving cell to one neighbor cell or to lower
the handover threshold, you can increase the CIO of the neighbor cell by 1 to
3 dB.
 A3 事件的进入条
件为：

Radio Parameters – ReportInterval and

ReportAmount
• ReportInterval and reportAmount
• Event triggering: reportInterval shall be used with reportAmount. When
reportAmount is greater than 1, reportInterval is valid.
• Periodical reporting: ReportInterval is the period to report the measurement
• Value range of ReportInterval: enumerated (120, 240, 480, 640, 1024, 2048,
5120, 10240, 60000, 360000, 720000, 1800000, 3600000), unit: ms
• Value range of ReportAmount: enumerated (1, 2, 4, 8, 16, 32, 64, Infinity),
unit: times
• Recommend value for ReportInterval is 480ms, and that for ReportAmount is
4.
 A3 事件的进入条
件为：

A3 事件的离开条

Radio Parameters – T304 件为：

• T304
• Timer waiting for handover success
• Value range: enumerated (50, 100, 150, 200, 500, 1000, 2000), unit: ms
• Default value: 1000 ms
Case Study

• Handover failure due to improper radio parameter settings

• Problem description
• The cluster KPI analysis report shows there is about 5% call drop because of MR
(measurement report) not processed.
• In a cluster, the call drops because of MR (measurement report) not processed even
account for over 50% of the call drops excluding those arisen from poor coverage.
• Handover configuration: S1-based; the handover request is triggered when the neighbor
cell’s RSRP is 3.5 dB higher than the serving cell’s.
Case Study

• Problem Description
Case Study

• Problem description
• After receiving a handover
request from an eNB, the MME
processes the request.
• If the MME receives another
handover request from the eNB
within 1s, it will ignore this
second handover request.
Case Study

• Problem analysis
• Current parameter configuration
• HO Prepare Timer = 1 s
• (Upon receiving one UE’s MR request, the source eNB sends the “handoverRequired”
messages to MME at an interval of 1 s.)
• Intra-frequency Reporting Interval for Event = 1024 ms
• Intra-frequency Amount of Reporting for Event = 1
• (If the target cell meets the MR reporting criterion, the UE reports MR every 1024ms
and the total times is 1.)
• If the UE performs handover to two neighbor cells one by one in 1 second,
and the amount of reporting is only 1, the second handover request message
will be bypassed by the 1s timer on the core network side. As a result, the
handover fails and call drop occurs.
Case Study

• Problem analysis: handover time

HO prepare sequence
timer time-out,
handover process is
Receive MR Receive cancelled; trigger a 2nd
Trigger handover 2nd MR handover based on 3rd MR

eNB

950 ms

UE
0 480 ms 960 ms 1440 ms 1920 ms 2400 ms 2880 ms

Trigger 1st MR Trigger 3rd MR

Trigger 2nd MR Trigger 4th MR

 Case Study

• Problem solution Setting

Parameter Explanation
before after
In an S1-based handover, the source eNB
sends the HandoverRequired message to the
core network, and starts this timer; After
HO Prepare Timer 1s 950 ms
receiving handover command from the core
network, it stops this timer. If the timer reports a
timeout, it means an abnormality.
Intra-frequency
The parameter indicates the interval to report the
Reporting Interval for 1024 ms 480 ms
MRs interval after event triggering.
Event

Intra-frequency
The parameter indicates the maximum report
Amount of Reporting 1 4
times after event triggering.
for Event