UMTS CS Call Drop Analysis Guide zte

Internal▲

UMTS CS Call Drop Analysis Guide

Version: V2.0

ZTE UMTS Radio Network Planning & Optimization Department

内部公开 ▲

UMTS Network Planning & Optimization Guidebook Release Notes: Version

Date

Author

V1.0

2009/4/23

Qin Jianhan

V2.0

2009/11/25

Zhou Changjing

Reviewed by None Yin Jianhua, Gan Yi, Xu Zhexian

Revision History Draft 2nd edition

内部公开 ▲

Key words: CS, Call drop

Abstract: This document introduces ways to evaluate, test, analyze and solve the call drop problem.

Abbreviation: None

Reference: None

Table of Contents Scenario one that may cause the same PSC problem 7.................................................................III Scenario two that may cause the same PSC problem 8.................................................................III Scenario three that may cause the same PSC problem 8...............................................................III Measurement of antenna power on PMS 9.....................................................................................III Flow chart to test call drops by DT 13.............................................................................................III Flow chart for top cell selection 16..................................................................................................III PMS cell performance measurement figure 18..............................................................................III RSCP and Ec/Io threshold for different services 4........................................................................IV Common E1 faults and handling suggestions 10..........................................................................IV Timer and counter related to the UE 23.........................................................................................IV 1 Introduction .......................................................................................................................................1 2 Definition.............................................................................................................................................2 2.1 Definition of Call Drop from Drive Test Aspect ................................................................................2 2.2 Definition of Call Drop at OMC Side .................................................................................................2 3 Call Drop Analysis .............................................................................................................................4 3.1 Call Drop Reasons ...............................................................................................................................4 3.1.1 Call Drops Caused by Poor Coverage ......................................................................................4 3.1.2 Call Drop Caused by Neighbor Cells .......................................................................................5 3.1.3 Call Drop Caused by Interference ............................................................................................6 3.1.4 Call Failure Caused by Two Cells Using the Same PSC...........................................................7 3.1.5 Call Drops Caused by Engineering Causes ..............................................................................9 3.1.6 Call Drops Caused by 2G/3G Interoperability .......................................................................12 3.1.7 Call Drops Caused by the System ..........................................................................................12 3.2 Analyzing Call Drops by DT ............................................................................................................13 This document contains proprietary information of ZTE Corporation and is not to be disclosed or used except in accordance with applicable agreements.

I

Internal▲ ....................................

3.3 Analyzing Call Drops by Traffic Statistics .......................................................................................14 3.3.1 Procedure of KPI Analysis ......................................................................................................15 3.3.2 Basic Methods to Analyze KPIs .............................................................................................16 3.3.3 KPI Analysis Tools .................................................................................................................18 3.4 Radio Parameters Involved During Optimization ............................................................................20 3.4.1 Radio Parameters Related with CS Call Drops ......................................................................20 3.4.2 Timer and Counter Related with Call Drop ............................................................................23

-II-

Internal▲

Figures Scenario one that may cause the same PSC problem..........................................................................7 Scenario two that may cause the same PSC problem..........................................................................8 Scenario three that may cause the same PSC problem.......................................................................8 Measurement of antenna power on PMS..............................................................................................9 Flow chart to test call drops by DT.....................................................................................................13 Flow chart for top cell selection...........................................................................................................16 PMS cell performance measurement figure.......................................................................................18

This document contains proprietary information of ZTE Corporation and is not to be disclosed or used except in accordance with applicable agreements.

III

Internal▲ ....................................

Tables RSCP and Ec/Io threshold for different services ................................................................................4 Common E1 faults and handling suggestions ...................................................................................10 Timer and counter related to the UE ..................................................................................................23


IV

Internal▲

1 Introduction This document is compiled to guide the network optimization engineers to solve the call drop problem, to reduce the call drop rate, and to improve the quality of the network. It also introduces ways to evaluate, test, analyze and solve the call drop problem. In addition, it also includes some typical cases. In the actual network optimization activities, handover and call drop are strongly related. In most cases, handover failure would lead to call drops. For this kind of call drops, you may refer to the guidebook for call drops caused by handover. This document mainly focuses on call drops which are not caused by handover failures.


1

Internal▲

2 Definition 2.1 Definition of Call Drop from Drive Test Aspect Air interface signaling at the UE side: Call drops refer to call releases caused by Not Normal Clearing, Not Normal, or Unspecified when the message on the air interface satisfying any of the following three conditions: 

The UE receives any BCH information (system information).



The call is released for Not Normal and the UE receives the RRC

Release information. 

The UE receives CC Disconnect, CC Release Complete, and CC

Release information. Signaling recorded at the RNC side: Call drops refer to call releases when the RNC has sent the Iu Release Request to the CN through the Iu interface, or when the RNC has sent the RAB Release Request information to the CN through the user panel.

2.2 Definition of Call Drop at OMC Side The definition of call drop in a broad sense contains the call drop rates at both the CN and UTRAN sides. Since the network optimization focuses on the call drop rate at the UTRAN side, this document only focuses on the KPI analysis at the UTRAN side. The KPIs at the UTRAN side refers to the number of released RABs of different services triggered by the RNC. Two aspects are involved: (1) After the RAB is established, the RNC sends the RAB RELEASE REQUEST information to the CN. (2) After the RAB is established, the RNC sends the IU RELEASE REQUEST to the CN, and then it receives the IU RELEASE COMMAND from the CN. The statistics can be collected based on specific services. Meanwhile the traffic statistics also imply reasons that the RNC triggers the release of the RABs of different services. The call drop rate can be calculated by the following formula:


2

Internal▲ CS _ CDR =

∑ CSRabrelTriggedByRNC *100% ∑ CSRABSetupSuccess

CSRabrelTriggedByRNC contains the number of RABs included in RAB RELEASE REQUEST for CS services and that included in IU RELEASE REQUEST for CS services.

PS _ CDR =

∑ PSRabrelTriggedByRNC *100% ∑ PSRABSetupSuccess

RabrelTriggedByRNC contains the number of RABs included in RAB RELEASE REQUEST for PS services and that included in IU RELEASE REQUEST for PS services. It should be specified that the RNC traffic statistics calculates the times of call drops through the signaling at the Iu interface, and counts the number of RAB RELEASE REQUEST and the number of IU RELEASE REQUEST initiated by the RNC. While call drops in the drive test aspect emphasizes the information at the air interface and non-access stratum and their cause value. It is different from call drops at the OMC side.


3

Internal▲

3 Call Drop Analysis Many reasons may lead to the call drop problem, and call drop is an expression of the deep network problems. This chapter focuses on the call drop reasons, commonly-used call drop analysis methods, and main call-drop optimization instruments.

3.1 Call Drop Reasons 3.1.1 Call Drops Caused by Poor Coverage In the definition of network coverage, the requirements of effective coverage for a certain sampling point is that its RSCP and Ec/Io should be better than the specified threshold. In this section, bad coverage is represented by poor RSCP value. Note that coverage at cell edges is a special case. Coverage at cell edges would have bad RSCP value and excellent Ec/Io owing to little cell number, but still the coverage in these cell edges is defined as bad coverage. In UMTS network, initiation and maintenance of different services would have different requirements on coverage. Table 1 lists the reference values. Table 1 RSCP and Ec/Io threshold for different services Service Type

RSCP [dBm]

Ec/Io [dB]

AMR12.2K

-105

-13

CS64K

-100

-11

PS384K

-95

-10

HSDPA

-90

-8

The coverage condition at the UL and DL of the network can be estimated through the power of the dedicated channels for the UL and DL before call drops, which can be performed through the following methods. If the UL TX power before the call drop has reached the maximum value and the UL BLER is bad, or it is found out through the single user tracing record at the RNC that the NodeB has reported RL failure, then the call drop is caused by bad UL coverage. If the DL TX power before the call drop has reached the maximum value and the DL BLER is bad, then the call drop is caused by bad DL coverage. This document contains proprietary information of ZTE Corporation and is not to be disclosed or used except in accordance with applicable agreements.

4

Internal▲ For the coverage optimization method, see the WCDMA Radio Network Optimization Guide.

3.1.2 Call Drop Caused by Neighbor Cells 1.

Missed neighbor cell

Neighbor cell optimization is an important link of radio network optimization. If certain cells should be included but excluded from the neighbor cell list of one cell, then call drop would happen and the interference in the network would also increase and system capacity would be impacted. Therefore, neighbor cell optimization is an important part of the engineering optimization. It is easy to estimate whether the cell is configured with any neighbor cell, and you can playback the call drop data, perform NCOS analysis, and analyze the scanner data. 

Use ZTE CNA to playback the call drop data. If the blue pillar

(representing the detected set) in the histogram of the pilot signals is the longest, then the missed neighbor cell problem exists. 

Use the automatic analysis tool of ZTE NCOS, it would study the

handover data of the network, and automatically add the missed neighbor cell. For details, see the operation guide of NCOS. 

During the drive test, the UE would acquire the neighbor cell list from

the NodeB, and the scanner would scan the 512 PSCs and record the Ec/Io. If one of the PSCS is not included in the neighbor cell list, and its pilot strength is stronger than the threshold, and the phenomenon lasts for a few seconds, then the missed neighbor cell problem exists. 2.

Removal of key neighbor cells caused by combination of macro

diversity Assign the priority of the neighbor cell when performing the initial neighbor cell planning, then optimize the priority and number of neighbor cells periodically with NCOS as the traffic volume increases. 3.

Untimely update of the external cell information

Check the external cells of the RNC periodically, and ensure the cells in the list are correct. This document contains proprietary information of ZTE Corporation and is not to be disclosed or used except in accordance with applicable agreements.

5

Internal▲ 3.1.3 Call Drop Caused by Interference Distinguish the UL and DL interferences. The interferences from the UL and DL would both lead to call drop. Generally, when the CPICH RSCP of the active set is large than -85dBm, and the comprehensive Ec/Io is lower than -13dB, call drop occurs, then the call drop is caused by the DL interference. Note that when the handover is not timely, the RSCP of the serving cell may be good, but the Ec/Io is bad. However, the RSCP and Ec/Io of the monitored set are both excellent under this condition. When the UL RTWP is 10dB higher than the normal value, which is -107~-105, and the interference duration is 2s or 3s longer, call drop may happen and the problem must be solved. Two reasons may cause DL interferences, which are pilot pollution and missed neighbor cell. The missed neighbor cell has already been discussed in the above part and would not be repeated here. In the pilot pollution area, signals of multiple cells exist, the RSCP of these cells is good, while Ec/Io is bad, the UE would frequently reselect the neighbor cell or perform the handover, and the incoming and outcoming of calls can hardly reach the UE. Generally, three factors would lead to pilot pollution in the network. 

Overshooting of sites at a high location



NodeBs in ring-shaped distribution



Wave-guide effect, large reflectors, and some other effects that may

cause the distortion of signals. The typical feature of DL call drops is that the RNC sends the Active Set Update message, while the UE cannot receive it, then the call is dropped for RL Failure. You can judge whether the UL interferences exist by the Average RTWP and Max RTWP on the OMC-R. For an idle cell, the Average RTWP is about -105dBm; for a cell carrying 50% of UL load, the Average RTWP is around -102dBm. If the Average RTWP of an idle cell exceeds -100dBm, we can believe that UL interferences exist. The UL interferences make the UL TX power of the cell in connected mode increase, and then an excessively high BLER is generated. Then call drop happens. During handover, the newly-added link is out of sync for UL interferences, which further leads to failed handovers and call drops. The UL interference may be intra-RAT or inter-RAT interferences. In most cases, the UL interferences are inter-RAT interferences. This document contains proprietary information of ZTE Corporation and is not to be disclosed or used except in accordance with applicable agreements.

6

Internal▲ When DL interference exists, the UL TX power is very small or the UL BLER may converge, however, when the DL TX power of the UE reaches the maximum value, the DL BLER does not converge. If UL interferences exist, the same problem would insist. Thus, in actual analysis, this method can be used to distinguish whether interferences exist. For methods to investigate the interferences, see the UMTS Interference Investigation Guidebook.

3.1.4 Call Failure Caused by Two Cells Using the Same PSC 3.1.4.1

Scenario One

Figure 1

Scenario one that may cause the same PSC problem

Cell A and Cell B (source cell) are configured as neighbor cell for each other, however, the geographical distance between Cell A and Cell B is huge. Cell A and Cell C has the same PSC, and Cell C and Cell B (source cell) is very close, however, Cell C and Cell B are not configured as neighbor cells for each other. Under this situation, the UE detects signals from Cell C and sends Event 1A request to be soft handed over to Cell C. The PSC in the Event 1A request is 123. After receiving the Event 1A request, the RNC checks from the neighbor cell list of Cell B (source cell) for cells with PSC of 123, then it finds Cell A. Then the RNC tries to build the radio link on Cell A. The RNC instructs the UE to add Cell A to its active set. Then, the update of the active set times out for the cell measured by the UE is different from the


7

Internal▲ cell where the radio link is built. 3.1.4.2

Scenario Two

Figure 2

Scenario two that may cause the same PSC problem

In this scenario, the UE has established the radio link with two cells, Cell B and Cell C. Cell A is the neighbor cell of Cell B, and Cell D is the neighbor cell of Cell C, and these two cells have the same PSC. When the UE is in soft handover state, the RNC would combine the neighbor cell lists of Cell B and Cell C, then the same PSC problem would happen. 3.1.4.3

Scenario Three

Figure 3

Scenario three that may cause the same PSC problem


8

Internal▲ Cell B and Cell D are not configured as neighbor cell for each other, however, these two cells are both included in the active set owing to the third-party handover among Cell B, Cell C, and Cell D. Cell A is the neighbor cell of Cell B, and Cell E is the neighbor cell of Cell D, and these two cells have the same PSC. The RNC would combine the neighbor cells of Cell B, Cell C, and Cell D in the active set, then the same PSC problem may occur.

3.1.5 Call Drops Caused by Engineering Causes 1. Reversely-connected antenna You can check whether the diversity is reversely connected by the PSC distribution figure of the drive test data. For the connection of the diversity, the PMS can be used to measure the cell performance. The antenna would only generate power when UEs try to access the network, and the measured value of the power equals to the demodulation power. You can check the ratio of two antennas, if the power of one antenna is lower than the other one in a long period of time, then the diversity must be reversely connected.

Figure 4

Measurement of antenna power on PMS


9

Internal▲ The balance level checking of two antennas in whole network can be implemented by OMCB measurement. However, you need to manually process the acquired data. 2. An excessive VSWR You can check the VSWR of the current site at the RNC SDR. If the VSWR is large than or equals to 1.4, then it must be adjusted. 3. Multi-band antenna problem In the network of some cities, multi-band antennas exist. The operator usually refuses to adjust the parameters of the multi-band antenna for fearing of affecting the subscribers of the existing 2G network. Then pilot pollution or overshooting may occur. To solve this problem, you should try to persuade the operator to change the antenna, so that 2G and 3G networks can have separate antennas. If these antennas cannot be changed, then the specific environment must be carefully studied before taking any actions. You can optimize the neighbor cells to avoid call drops. 4. Leakage of signals from indoor distribution system In most cities, call drops caused by signal leakage from indoor distribution system exist. You should persuade the operator to reconstruct the indoor distribution system. Or, the indoor distribution system can be merged to the whole network, which can be done by optimizing of the coverage of the ambient outdoor cells and addition of neighbor cells. 5. Call drop caused by unsteady transmission As the bottom level of transmission medium, E1 would report the loss of E1 electrical signals and reception failures at the remote end. Meanwhile, several E1s would be bound together as a group, and then E1 would report the fault of IMA group in nonoperating mode. The following table lists several E1 faults that must be handled and the related handling suggestions. Table 2 Common E1 faults and handling suggestions Fault Lost of E1 electrical signals

Causes

Solutions

The RX end detects no line circuit pulse or

1. Check whether the SA board is

cannot detect logic 1 within continuous

secure, and whether the E1 adapter

periods, then the LOS alarm is reported. This

is slack.


10

Internal▲ alarm is generally caused by the RX fault of the E1/T1 or broken lines, then the E1/T1

2. Check whether the pins of the adapter are damaged.

cannot detect the signals from the remote end.

3.

Check

whether

the

joint

connector of the E1 cable is damaged, and whether the joint connector is securely connected with the E1 cable. 4. Check whether the cabling of the

E1

cable

satisfies

the

engineering specification, whether the E1 cable bears any external force. 5. Use the E1 self-loop cable to recycle the line, if the alarm is cleared, then check the E1 cable at the peer end. Remote reception

It indicates the E1/T1 remote alarm. This

1. The TX line is faulty or

failure of the E1

alarm indicates the abnormal receptions at the

broken. Check whether the TX line

remote end. The remote end inserts the RAI

is correctly connected. For details,

indicator bit to the signals and then sends it to

see the Handling suggestions for the

the local end, and the local end reports the

LOS Alarm.

alarm after detecting the alarm. The remote reception error is reported.

2. Check whether the frame structures of the E1 frame at the local end and remote match. The E1 frame at both ends must both work at dual-frame or multi-frame mode. 3. Check for error codes at the TX line.

E1 frame out of sync

The first bit of slot 0 of both E1 and T1 carries the synchronous clock signals, which

1. Whether E1 and T1 work at the same state.

inform the RX end of the start of one frame. If

2. Check whether E1 frames are

the RX ends of the E1 and T1 are out of sync,

of

then data frames would be lost and the LOF

frame/multi-frame).

alarm is reported.

the

same

modes

(dual-

3. Check whether the impedance modes of E1/T1 matches. 4. Check for interferences from digital devices around E1/T1. 5. Check whether the clock signals are normal.

SSCOP link error

This alarm is caused by that the SSCOP signaling link is unsuccessfully established or

See the handling suggestions for E1 faults.


11

Internal▲ the SSCOP signaling from the remote end is not received within a certain period. Then the SSCOP link would be broken off, and this alarm is reported. IMA group in

After the IMA group is successfully

non-operating

configured, if IMA remains in non-operating

mode

mode for over 1s, then this alarm is reported.

See the handling suggestions for E1 faults.

Currently, some sites are configured with IP transmission. Therefore, the alarm of "Lack Ethernet electrical signals" also should be handled on site.

3.1.6 Call Drops Caused by 2G/3G Interoperability 1.

Optimization of 2G neighbor cells configured for 3G cells

If the 2G cells are congested, or interfered, then the success rate of 3G -> 2G handovers is low. During the neighbor cell optimization, this kind of neighbor cells must be removed from the list. 2.

Parameters must be refined based on different scenarios.

To improve 3G->2G handover success rate, the parameters must be detailed planned based on different scenarios. 3.

Compatibility of UEs

The 2G->3G handovers of some cells are slow. This is because some smuggled 3G handsets have some difficulties in supporting the 2G network. 4.

2G/3G data synchronization

To support 2G/3G handovers, the 2G/3G cells must be configured as the neighbor cells for each other firstly. If the cell information is updated timely, then the handover would fail and cell reselection cannot be performed. Therefore, the data of 2G/3G network should be synchronized timely.

3.1.7 Call Drops Caused by the System If the alarm is not caused by the causes listed in the above section, then it may be caused by the system. You need to check the alarm information of the equipment and system logs to further analyze reasons that cause call drops. For example, an abnormal NodeB would lead to the synchronization failures, which would lead to frequent This document contains proprietary information of ZTE Corporation and is not to be disclosed or used except in accordance with applicable agreements.

12

Internal▲ removal and addition of radio links, and then call drops may happen; call drops caused by poor DL signals may be because of abnormal RF module, and call drops caused by that the UE fails to report the measurement report Event 1A. It should be noted that in many foreign countries, the TX environment is bad and unstable. Therefore, influences of call drops caused by TX problem are huge.

3.2 Analyzing Call Drops by DT The following figure describes the flow chart for using DT and CQT to test call drops.

Figure 5

1.

Flow chart to test call drops by DT

Call drop data


13

Internal▲ The call drop data refers to the CNT test data and RNC signaling tracing data. 2.

Call drop spots

Use CNA to analyze the call drops to acquire the location where call drops happen. Then acquire the following data: pilot data collected before and after call drops, active set and monitoring set information collected by the cell phone, and signaling flow. 3.

Stability of the primary serving cell

The stability of the primary serving cell refers to its changes. If the primary serving cell is stable, then analyze RSCP and Ec/Io. If the primary serving cell changes frequently, then the handover parameters should be changed to avoid the ping-ping effect. 4.

RSCP and Ec/Io of the primary serving cell

Check the RSCP and Ec/Io of the optimal cell, and then 

When the RSCP is bad, the coverage is poor.



When the RSCP is normal, while the Ec/Io is bad, pilot pollution or DL

interference exists. 

When RSCP and Ec/Io are both normal, if cells in the active set of the

UE are not the optimal cells (which can be checked through playback of data), then the call drops may be caused by missed neighbor cell or untimely handovers; if cells in the active set of the UE are the optimal cells, then call drops may be caused by UL interferences or abnormal call drops. 5.

Reproducing of problems with DT

Since you cannot collect all necessary information by one DT, then multiple DTs shall be performed to collect sufficient data. In addition, multiple DTs can also help to ascertain whether the call drop is random or always happens at the same spot. Generally, if call drops always happen at the same spot, this problem must be solved, or if call drops happen randomly, multiple DTs must be performed to find inner reasons.

3.3 Analyzing Call Drops by Traffic Statistics When analyzing the traffic statistics, check the call drops index on the RNC firstly to learn the operating status of the whole network. Meanwhile, a cell-by-cell analysis can be performed to acquire the detailed call drop indexes of each cell. During the analysis, This document contains proprietary information of ZTE Corporation and is not to be disclosed or used except in accordance with applicable agreements.

14

Internal▲ the traffic statistic analyzing tool can be used to analyze the call drop situations of different services and the possible causes. Acquire data about cells with abnormal KPIs through the traffic statistics. If KPIs of these cells used to be normal, then the abnormal KPIs may be brought by software version, hardware, transmission, antenna, or data, then you can check these aspects based on the alarms. If no obvious abnormal cells exist, the statistics can be classified based on the carrier in each sector, then cells with poor KPIs can be screened out. Further analyze the traffic statistics of these cells, such as analyzing more related KPIs, such as analyzing data at a shorter interval, or analyzing KPIs that are more likely to cause call drops, such as handover. Meanwhile, you can analyze the reasons for call drops based on system logs. During the analysis, you should consider the effect of other KPIs when focusing on a certain KPI. It should be specified that the result of traffic statistics is meaningful only when the traffic volume reaches a certain amount. For example, a 50% of call drop rate does not mean that the network is bad. This value is meaningful only when the calling number, succeed calling number, call drop times all make statistical significances.

3.3.1 Procedure of KPI Analysis The commonly used KPI analysis method is the TOP cell method, which means the top cells will be screened out according to certain index, then these top cells are optimized and then the top cells are selected again. After several repetitions, the related KPI can be speedily converged. At the initial stage of network construction, there are few subscribers in the network. At this stage, the KPIs of many cells might be unstable, such as call drop rate. You can collect the data in seven days or longer periods, then select the top cell and then perform the optimization. For example, optimization of call drop rate of CS services. When selecting top cells, you can select the cell with call drop numbers exceeding the specified threshold, and then arrange the priority based on the call drop rate. The procedures of top cell selection are the same as the procedures of handling input information from other team of engineers (complains or single site acceptance), and are shown in the following figure.


15

Internal▲

Figure 6

Flow chart for top cell selection

3.3.2 Basic Methods to Analyze KPIs 3.3.2.1

Speedily Collecting the Field Data To locate the problem, you have collect data from many different spots between the UE and the pdn server. While, speedy and accurate collection of the field data is essential to locate and solve the problem and to improve the KPIs. Data collection can be divided into multiple layers. 1.

Collecting UE log, RNC signaling, KPI data, alarms, abnormal probes, and packet

captured at the Iub interface 2. NodeB and RNC debug log Some common skills are required to collect data of the first layer, and the network optimization & maintenance personnel can easily master these skills. At present, most This document contains proprietary information of ZTE Corporation and is not to be disclosed or used except in accordance with applicable agreements.

16

Internal▲ field questions can be located through the data analysis at this layer. Collection of the debug log of the second layer should be performed or remotely supported by the relative R&D engineers. Data at this layer can help to solve some deep layer problems. The following chapter focuses on the data collection tool and method for the first layer data, and only gives a brief introduction to that of the second layer. 3.3.2.2

Health Check of Sites For sites where alarms are reported, you should first perform the health check for the site, which mainly covers the following aspects: 

Alarms



Frequently added or removed common transport channels



UL & DL power



Radio link restore



Balance level between two antennas



Statistics of service failures

The RL restore rate is shown by the NodeB cell measurement recorded by PMC as shown in the following figure, and is accumulated since the establishment of cells. If the RL restore rate of a cell is lower than 80%, the cell is treated as abnormal, and the possible causes are as follows: 

UL interferences



Insufficient cell radius or overshooting



Reuse of the same PSC



Abnormal UL RF channel

For these possible causes, you may check them combining other measurement results and data analysis.


17

Internal▲

Figure 7

PMS cell performance measurement figure

3.3.3 KPI Analysis Tools 3.3.3.1

Signal Trace This tool traces signaling of RNC, you can trace the signaling at the Iu, Iur, Iub, and Uu interfaces, TNL signaling, and RNL signaling through this tool. The most commonly used method to check the KPIs is to trace the RNL signaling. This tool is very useful, and can trace the signaling on the basis of cell (trace signaling of multiple UEs) and IMSI (trace signaling of one UE). It should be emphasized that signaling tracing by cells can only trace the UE that initiates the call from this cell. The UE can be traced as long as it remains in the same RNC, even if it is handed over to other cells. However, if a UE initiates the call from other cells and then is handed over this call, and its call drop happens in this cell, it cannot be traced. Therefore, when you trace the signaling of a cell with high call-drop rate, the signaling of cells in close handover relation with this cell should also be traced, then the result would be more comprehensive. The RNC R&D engineers also develop a RNC signaling tracing tool, WinSigAn, which can mark the call drop spots more clearly.


18

Internal▲ 3.3.3.2

RNC Association Log This tool helps to record the context of the abnormal system flow, and then the context would be counted and analyzed to locate the network problem. It is usually used when the system is abnormal and no RNC signaling is traced. It can help to locate the problem by the time when the system exception happens. The exception can be queried on the basis of IMSI and CELL ID.

3.3.3.3

NodeB LMT Besides all functions of OMCB, NodeB Local Maintenance Terminal (LMT) can also provide detailed cell and UE information. The NodeB LMT consists of EOMS, EFMS, DMS, and PMS.

3.3.3.4

NodeB Exception Probe In the field of the WCDMA commercial network, this tool can effectively help to monitor the operating status of the NodeB. Different modules of the NodeB would record the information when exceptions happen, thus facilitating the location of problems. However, specialized knowledge is required. You have to understand the functions and interfaces of different boards. If the field engineers cannot analyze the report, they can simply send these data to the R&D engineers. The exception probe reported by different NodeBs can be saved on different OMCB servers based on the RNC they belong to. Then, this tool would download the file from different OMCB FTP, and then analyze them.

3.3.3.5

CTS CTS is the tool for the CN, and it can be used to perform deep signaling by IMSI. Unlike SignalTrace, which is applicable to the signaling tracing within one RNC, CTS can perform the signaling tracing across the RNC border, Therefore, it is applicable to the signaling tracing of VIPs. CTS can trace the interactive signaling among different NEs within the CN, and can trace the signaling at the Iu and Uu interfaces, and this is called deep tracing. The working principles of CTS is as follows: First establish an IMSI task on CTS server, and then sent this IMSI task to the CN, which is further sent to different modules through the arranged interfaces, then each module collects the signaling related to IMSI, and then the collected signaling is transmitted back to the CTS server through


19

Internal▲ the CN. The above interfaces are all private interfaces, thus this tool only work on ZTE CN and RNC. 3.3.3.6

UE Log DT is an important means to analyze KPIs. Many problems, signaling tracing at the network side and tracing of problems which are hard to be located, can be finally located after combining the UE logs. The commonly used DT software is QXDM/APEX(QCAT), CNT/CNA, and TEMS.

3.4 Radio Parameters Involved During Optimization 3.4.1 Radio Parameters Related with CS Call Drops Time To Trigger Time To Trigger is the interval between the moment that the events (1A, 1B, 1C, and 1D) are monitored and the moment that the events are reported. The setting of TTT would influence timely handover. The adjustment of handover parameters should first ensure that this cell is overlapped by other cells, then you can adjust the related radio parameters to ensure that the time that the UE passes the handover area is longer than the handover delay of the whole system, thus ensuring the continuity of the services. The other is to ensure that the handover area ascertained by the signals and radio parameters cannot be too large to avoid the increase of handover overhead and reduction of resource utilization ratio. For areas where the signals may change greatly, the trigger time of Event 1A must be reduced, and that of Event 1B must be increased. Meanwhile, the CIO of the corresponding neighbor cells should be adjusted so that Event 1A can happen earlier and Event 1B would happen later, thus ensuring successful handovers. For highways, the cells are sparsely distributed. If the vehicles drive too quickly and cannot access the new cell in time, call drops would happen. The optimization is the same as that for the optimization for street corners in dense urban, which is to make cells with good signals join the active set speedily to ensure continuity of services. For the adjustment of the related parameters, a whole new set of parameters must be assigned to the target cell.


20

Internal▲ 3.4.1.1

Cell Individual Offset The sum of the value of Cell Individual Offset (CIO) and the actually measured value is used in the evaluation of the events of the UE. The UE would use the original measurement value of this cell plus the CIO as the measurement result for the intrafrequency handover judgment. CIO can help to ascertain the cell edge. The larger this parameter is set, the easier the soft handover will be, and more UEs will be in soft handover state. However, more resources are consumed. This smaller is parameter is set, the more difficult the soft handover is. CIO is valid only for the neighbor cell. For Event 1A, the CIO can be set in the neighbor cell; for Event 1B, the CIO can be set in the cell to be removed. The formula is as follows: Formula of Event 1A triggering:

 NA  10 ⋅ LogM New + CIONew ≥ W ⋅10 ⋅ Log  ∑ M i  + (1 − W ) ⋅10 ⋅ LogM Best − ( R1a − H 1a / 2),  i =1  MNew is the measurement of the to-be-evaluated cells that has entered the report range. Mi is the mean measurement result of cells (exclude the best cell) in an active set. NA is the current cell number (exclude the best cell) in the active set. MBest is the measurement result of the optimal cell in the active set. W is the weight proportion of the best cell to the rest cells in the active set. R1a is the reporting range of Event 1A. H1a is the reporting hysteresis of Event 1A. Formula of Event 1B triggering:

10 ⋅ LogM Old + CIOOld

 NA  ≤ W ⋅10 ⋅ Log  ∑ Mi  + (1 − W ) ⋅10 ⋅ LogMBest − ( R1b + H1b / 2),  i =1 

Mnew is the measurement of the to-be-evaluated cells that has entered the report range. Mi is the mean measurement result of cells (exclude the best cell) in an active set. NA is the current cell number (exclude the best cell) in the active set. MBest is the measurement result of the optimal cell in the active set. W is the weight proportion of the best cell to the rest cells in the active set.


21

Internal▲ R1bis the reporting range of Event 1B. H1b is the reporting hysteresis of Event 1B. 3.4.1.2

Start/Stop Threshold for Compressed Mode Compressed mode is frequently used during inter-frequency and inter-RAT handovers. The compressed mode is started before the handover, and the system can use the time slot brought by compressed mode to perform the signal quality test on the interfrequency or inter-RAT neighbor cells. In the current system, the compressed mode is started through Event 2D, and stopped through Event 2F. The measurement value of RSCP or Ec/Io can be selected. Currently, the default value is RSCP. Generally, the quality and other related information of the target cell (inter-frequency or inter-RAT) must be acquired for the compressed mode. Meanwhile, the moving of the UE would lead to the deteriorate of the quality of the cell, therefore, the start threshold of the compressed mode should satisfy the condition that the UE can finish the measurement of the target cell and report for handover before call drops happens. For the stop threshold, it should help to avoid the frequent start or stop of compressed mode. In dense urban, the continuous coverage of the 3G should be ensured, thus avoiding unnecessary inter-RAT handovers and increase of system load. For edges of the 3G network and highways, the UEs should be handed over to the 2G network before the heavy fading. Under this condition, the trigger threshold of Event 2D should be raised so that the UE can initiate the compressed mode as early as possible.

3.4.1.3

Maximum DL TX power of the Radio Link If large amounts of call drops happen due to coverage causes, then the maximum DL TX power of the services can be increased appropriately. However, this is at the risk that the UEs at cell edges may consume too much power, and then affect the other UEs, and reduce the DL capacity of the system. For cells with a great deal of access failures caused by excessive load, this parameter can be set to a small value.

3.4.1.4

Inter-Frequency/Inter-RAT Handover Threshold The UE can be handed over to the inter-RAT/frequency neighbor cells when the measured value of the signals from these cells is higher than the threshold. This parameter can be set combining the start threshold of the compressed mode. If this


22

Internal▲ parameter is configured with a little value, then the handover can be triggered early. If this parameter is configured with a large value, then the handover will be prolonged.

3.4.2 Timer and Counter Related with Call Drop The following table lists the timer and counter related to the UE. Table 3 Timer and counter related to the UE Name

Description

Value Range

Default Value

T312

T312 in connected mode, and indicates the

Connected

time that UE waits from the synchronization

(1..15)s

1s

1

indicator from L1 when it starts to establish the DPCCH. N312

T312 in connected mode, and indicates the

(1, 2, 4, 10, 20, 50,

Connected

number of synchronization indicator that the

100, 200, 400, 600,

UE received from L1 before the DPCCH is

800, 1000)

established. T313

Indicates the waiting time of the UE in

(0..15)s

3s

Indicates the number of maximum number of

(1, 2, 4, 10, 20, 50,

20

out of sync indicators that the UE receives

100, 200)

CELL_DCH state after the DPCCH channel is established. N313

from L1. T314

Start: When the criteria for radio link failure

(0, 2, 4, 6, 8, 12, 16,

are fulfilled. The timer is started if radio

20)s

4s

bearer(s) that are associated with T314 exist or if only RRC connection exists only to the CS domain. T315

Start: When the criteria for radio link failure

(0,10, 30, 60, 180,

are fulfilled. The timer is started if radio

600, 1200, 1800)s

30s

bearer(s) that are associated with T314 exist or if only RRC connection exists only to the CS domain. N315

Indicates

the

synchronization T309

maximum indicators

number that

the

of

(1, 2, 4, 10, 20, 50,

UE

100, 200, 400, 600,

received from L1 after T313 is activated.

800, 1000)

Indicates the waiting time of the UE after

(1..8)s

1

3s

sends the inter-RAT handover requests.


23

UMTS CS Call Drop Analysis Guide zte

Recommend Documents