Vous êtes sur la page 1sur 35

No.

8 March 2007 Issue


29
Maintenance Experience
www.zte.com.cn

GSM Network Products

Typical Cases
Experience Collection
Preface
In this issue of ZTE's "Maintenance Experience", we continue to pass on various
field reports and resolutions gathered by ZTE Engineers and Technicians from
Maintenance Experience
around the world.
Editorial Committee
The content presented in this issue is as below:
Twelve GSM-MSS-related cases
Director: Zhou Susu
Two special documents
Three experience exchange documents Deputy Director: Chen Jianzhou

Have you examined your service polices and procedures lately? Are you Editor - in - Chief: Yang Cheng
confident that your people are using all the tools at their disposal? Are
Editors:
they trained to analyze each issue in a logical manner that provides for less
J i a n g G u o b i n g , Wa n g Ya p i n g , B a
downtime and maximum customer service? A close look at the cases reveals
Z e x u e , Z h a n g S h o u k u i , Wu F e n g ,
how to isolate suspected faulty or mis-configured equipment, and how to
solve a problem step by step, etc. As success in commissioning and service Yuan Yufeng, Tang Hongxuan, Chen
is usually a mix of both discovery and analysis, consider using this type of Huachun, Ding Guixiang, Gu Yu, Tian
approach as an example of successful troubleshooting investigations. Jinhua, Zhu Wensheng, Ling Changwen,
Zhang Zhongdong, Shu Huiyu, Tang
While corporate leaders maintain and grow plans for expansion, ZTE Wen, Liu Xianmin, Wang Zhaozheng,
employees in all regions carr y out with individual ef for ts towards Chen Taiming, Zhang Mingjing, Wang
internationalization of the company. Momentum continues to build, in all
H a i d o n g , Ch e n L e, L e i Ku n , Wa n g
levels, from office interns to veteran engineers, who work together to bring
Tiancheng, Zheng Hongliang, Wang Tao
global focus into their daily work.
Technical Senior Editors:
If you would like to subscribe to this magazine (electronic version) or review
Zhang Niantao,Pan Xiangqin
additional articles and relevant technical materials concerning ZTE products,
please visit the technical support website of ZTE Corporation Executive Editor:

http://support.zte.com.cn Li Fenglian

If you have any ideas and suggestions or want to offer your contributions, you
MaintenanceExperience
can contact us at any time via the following email: doc@zte.com.cn.
Newsroom
Thank you for making ZTE a part of your telecom experience!

Address: ZTE Plaza, Keji Road South,


Hi - Tech Industrial Park,
Maintenance Experience Editorial Committee Nanshan District, Shenzhen,
P. R. China
ZTE Corporation
Postal code: 518057
April 2007 Contact: Song Chunping
Email: doc@zte. com. cn
Tel: 86 - 755 - 26770600 26771195
Fax: 86 - 755 - 26772236
Contents
Fault Instance 2
The SDT Trunk Circuit Is Blocked 2
Exceptional Unblocking of the Trunk Circuit Caused Declining of the Call Completion Rate 3
The Subscriber Couldn't Send the Short Message 5
The HLR180 Database Recovery Alarm 6
Backing up the Data from the HLRDB to the 129 Failed 7
The Inter-Module Communication between the CFBI and the FBI Was Exceptional 9
Error in the Backup Subscriber Database on the HLR 10
The MSC Signaling Link Was Intermittently Disconnected and the Call Completion Ratio Declined 11
It Was Difficult for the Subscriber Roaming to a Foreign Country to Log in the Network 13
The HP Minicomputer Failed to Create a Volume Group 14
No Tone in Both Directions 15
Hardware Fault Instance 17

Special Document 23
The Principle of the ECC Board and Its Use in the MSS System 23
The PRI Interface Configuration Description 27

Experience Exchange 29
Precautions for Shuffle 29
Select Statement of the Database 30
The Problems Occurs When Debugging the Monitoring Service 31
April 2007 Issue 29

Fault Instance

The SDT Trunk Circuit Is Blocked


Wang Ke, ZTE Corporation

moment after analyzing the on-site alarm.


Phenomenon
Finally, The ZTE engineer found the fault
It was found that other PCMs except the reason: the intermittent disconnection of
first 4 PCMs having signaling among the 63 the signaling link caused inter-office ISUP
PCMs from the MSC2 to the GMSC2 were signaling loss, and the ISUP blocked the
blocked through observing the MSC2 after related circuit due to signaling exception.
the version upgrading on the site. One fiber This is a protection behavior, not a version
trunk circuit was used between MSC2 and problem.
the GMSC2.

Solution
Analysis
The problem was solved after plugging and
The ZTE engineer used UNBLOCK in unplugging the SDT board in the MSC2, and
the MSC2 dynamic management to unblock activating the circuits.
the circuits after finding the circuits were
blocked, but the operation failed. Then the
ZTE engineer successfully unblocked the
circuits in the GMSC2 dynamic management,
but the circuits were blocked again after a
while.

T he fault may be caused by a version


problem. It was found that the intermittent
disconnection of signaling occurred at some

 Maintenance Experience
Exceptional Unblocking of the Trunk Circuit Caused
Declining of the Call Completion Rate
Chen Yu, ZTE Corporation

Phenomenon
The ZTE engineer checked the MSC performance statistics after finding that the call
completion rate and call answering rate dramatically declined recently, as shown in Figure 1.

Figure 1 MSC Performance Statistics 1

Analysis
1. The ZTE engineer opened the service observation and checked the call loss
situations. It was found that there are a large numbers of 2145=No digital trunk
available call losses, which meant that the trunk circuit had problems, as shown
in Figure 2.

Figure 2 Call Loss

GSM Network Products 


April 2007 Issue 29 Fault Instance

2. The ZTE engineer selected Advance Filter, as shown in Figure 3.

Figure 3 Advance Filter

The filtering option in the call loss observation indicated that the call losses were all
from the module 5, which meant that one trunk circuit had problems.

3. The ZTE engineer checked the performance statistics data of the module 5, as shown
in Figure 4, which displayed that there was heavy congestion in the inter-office trunk.

Figure 4 MSC Performance Statistics 2

Solution
After opening the dynamic observation and checking the inter-office circuit of the module 5,
the ZTE engineer found that one trunk of the Siemens MSC was blocked. The problem was
solved after the trunk was unblocked. The performance statistics data were normal when being
observed again.

This E1 was blocked maybe due to no operation, and that of the opposite end was not blocked,
which causes the failure of lots of inter-office traffic. During the daily maintenance, the
performance statistics data and alarm information should be observed carefully.
 Maintenance Experience
The Subscriber Couldn't Send the
Short Message
Chen Yu, ZTE Corporation

Phenomenon
The subscriber couldnt send the short message but could
receive the short message. The short message sending failure
was directly displayed on the screen of the mobile phone.

Analysis
This is a burst phenomenon, which indicates that the Figure 1 MS Flags
HLR configuration, the MSC configuration or the SCP
configuration may be modified. First, The ZTE engineer After searching the MSC operating lods, the ZTE engineer

performed the signaling tracing and then compared it with found that much data were modified last night. For example,

the correct signaling process. It was found that there is no the value of the local IMSI in the number roaming analysis

MSCMAP signaling when the fault occurred, which meant was changed from no roaming to roaming international.

that the fault was on the MSC. The ZTE engineer checked the Mobile Digit Analysis, as
shown in Figure 2.
Solution
The ZTE engineer performed the test on the site. He/she
deleted the normal number (the value of roamflag in the
MSFlags of the number is 0) from the VLR, and then
compared the MSFlags information after updating the
location. It was found that the value of roamf lag was
changed to 2, as shown in Figure 1. It was found that the
value of roamflag of all the problematic numbers was 2.

Figure 2 Number Roaming Analysis

GSM Network Products 


April 2007 Issue 29 Fault Instance

The subscriber re-updated the location after this parameter was modifyied. The value of
roamflag recovered to 0, and the subscriber could send the message normally. However,
there were still some subscribers in the VLR whose information was not modified on the site.
The information of all the subscribers can be forcibly updated. In addition, this fault can also
be solved by switching on or switching off the mobile phone.

The HLR180 Database Recovery Alarm


Chen Yunlong, ZTE Corporation

Phenomenon
The recovery alarm of the 180 database occurred on the HLR alarm management at 3 oclock
AM, and the alarm automatically disappeared at 5 oclock AM.

Analysis
The 180 database was successfully performed at 4 oclock AM. every day. After checking the
history records, the ZTE engineer found no failure record. After checking the configuration file
hccfg.ini on the server 129, the ZTE engineer found that the set time was within 3:00~5:00, so
the time for backing up the 180 database was in the set time range of the configuration file.

After checking the time of the database system and that of the server 129, the ZTE engineer
found there was difference between them.

Solution
The alarm disappeared after the time of the database system and that of the server 129 are
synchronized.

 Maintenance Experience
Backing up the Data from the HLRDB
to the 129 Failed
Chen Yu, ZTE Corporation

Phenomenon
The 129 was the backup database of the subscriber database in one HLR office. One day
when the system performed automatic backup, backing up the data from the node 173 to
the 129 failed. After checking the database, the ZTE engineer found that the msdb database
file of the backup database of the node 173 on the 129 was hung up. The ZTE engineer
rebuilt the HLRDB to recover the data of the previous day.

This problem occurred again one week later, and it was solved by performing the above
operations.

Analysis
Rebuilding the HLRDB can solve this problem for a while, but this problem occurred every
week, so the system must have hidden troubles. The backup has specific regularity, so the
error log of the backup should be analyzed, as shown in Figure 1.

Figure 1 Error Log


GSM Network Products 
April 2007 Issue 29 Fault Instance

Check Show step details at the top right


Solution
corner of the log dialog box, and check the
detailed reasons of the backup failure when 1. The ZTE engineer manually deleted

the problem occurs. As shown in Figure 7, the file with incorrect format.

the corresponding data format in the backup


2. T h e Z T E e n g i n e e r m a n u a l l y
direction of the node 173 on the 129 is
performed remote backup after one
incorrect, so it is recommended to compare
oclock pm (the automatic remote
the backed-up database file and that to be
backup was performed at 1:30 p.m.)
backed up when performing remote backups.
on the day of every week when the

Through the analysis, it was concluded that fault occurred, and the operation

the format of the backed-up database is was successful. Then waiting for the

damaged due to abnormal backup. Rebuilding automatic backup of the system, the

the HLRDB after the phenomenon occurred ZTE engineer found that the backup

every week imported the data of the incorrect was nor mally perfor med. After

format to the backup HLRDB on the 129 (the several weeks of observation, the

data can be imported though the format is problem didnt occur again.

damaged). Because the damaged file was not


updated by the new backup successfully, the
automatic backup task was influenced.

After checking in the backup directory, the


ZTE engineer found that the size of the file
with incorrect format is less than those of
other files, and the date is the day when the
problems occurred for the first time.

 Maintenance Experience
The Inter-Module Communication between
the CFBI and the FBI Was Exceptional
Huang Wuxiang, ZTE Corporation

Phenomenon
During the MSC commissioning in Bengal, after checking the hardware statuses of one
central module and four MPM modules of the foreground, the ZTE engineer found that the
statuses of the FBI and the CFBI having inter-module communication were abnormal. The
fault indicators on the FBI board of the peripheral module were all on, and the ch1\ch2\ch3\
ch4 indicators on the CFBI board were all on.

Analysis
The ZTE engineer tried to interchange upper and lower connection lines of the fiber of the
FBI. The statuses of the first three pairs of FBIs became normal, and the ch1\ch2\ch3\ch4
indicators on the CFBI were all off, but the ch1\ch2\ch3\ch4 indicators on the last pair of
CFBI boards were still on.

In general, a pair of CFBI boards correspond to 1~4 pair(s) of FBI boards of one MPM
module. When the quantity of the corresponding FBI boards is less than four pairs, there
are free optical interfaces on the CFBI board. The free optical interfaces should be short-
circuited by the jumper within the board. If the free optical interfaces are not short-
circuited, it may cause the level-2 alarm of the CFBI board.

This problem can be detected through the fault management system, and this phenomenon
will affect the switching of the FBI board.

Solution
After the last X25 jumper was screened, the ch1\ch2\ch3\ch indicators on the CFBI were all
off, and the inter-module communication was recovered to be normal.

GSM Network Products 


April 2007 Issue 29 Fault Instance

Error in the Backup Subscriber


Database on the HLR
Liu Yecheng, ZTE Corporation

two recovery tasks, and eliminated


Phenomenon
the possibility of the set error of the
The level 1 alarm appeared in the current recovery task and the problem of the
a l a r m of t he H L R , d i s pl ay i ng H L R SQL program.
conf igurat ion ser v ice l ink backup
database error, and the backup database 4. After checking the master database

node nu mber wa s 180. W hen open i ng of the node 170, the ZTE engineer

the SQL2000 Enterprise Manager of the found that its remote backup task

server 129, the ZTE engineer found that the also failed, but its local backup task

status of the database180 was suspect; was normal. After comparing the

when checking the backup recovery task latest file locally backed up on the

at the same time, the ZTE engineer found database 170 with that remotely

that the recovery task in the morning was backed up, the ZTE engineer found

unsuccessful. that the sizes of these two files were


not consistent. The reason might

Analysis be that the backed-up database file


is not complete, which causes the
1. After the SQL was restarted, the
exception in the database180 after
database180 was invisible in the
implementing the recovery task.
SQL Enterprise Manger.

2. The ZTE engineer deleted the file Solution


of the database180 and rebuilt this The ZTE engineer used the latest file locally
database, but the recovery task was backed up on the database 170 to cover the
still unsuccessful. However, the latest file remotely backed up on the database
recovery task of the backup database 170, and performed the recovery task. The
183 was normal. recovery task was successfully performed
and the alarm disappeared.
3. The ZTE engineer compared these
10 Maintenance Experience
The MSC Signaling Link Was
Intermittently Disconnected and the
Call Completion Ratio Declined
Tian Jinhua, ZTE Corporation

Phenomenon
In the MSC system of one office, some signaling links to the Alcatel MSC office were
intermittently disconnected at about seven oclock p.m., which caused that the call completion
ratio declined.

Analysis
There was lots of alarm information when the fault occurred, and the description of some
information is shown in Table 1.

Table 1 Alarm Information Description

Occur Time Recovery Time Additional Info

SLID=52, Office=3, SLC=1, Module=4,


9/7/2006 19:01:55 9/7/2006 19:02:54
Slot =22, Reason Code =10

SLID=52, Office=3, SLC=1, Module=4,


9/7/2006 19:04:23 9/7/2006 19:04:28
Slot =22, Reason Code =18

It was found that through the alarm information:

1. Office=3 meant the office direction of the ZET MSC to the Alcatel MSC.

2. Reason Code=10 meant that the bit error ratio of the physical link was high, because
the clocks of two ends of the signaling are not synchronized, which caused too many bit
errors and signaling blocking; Reason Code =18 meant that the local sending buffer area
was congested, which was caused by heavy load of the link. GSM Network Products 11
April 2007 Issue 29 Fault Instance

The clock problem caused the blocking of some links of this office direction, which
accordingly increased the load on other links to this office. As a result, more signaling links
were congested, and then intermittently disconnected.

See Table 2 for statistics on some signaling to the Alcatel MSC during the failure period.

Table 2 Signaling Statistics

Time Category description Traffic Received

09-07-2006 19:05 71-AlcatelMSC2 0.98


09-07-2006 19:05 72-AlcatelMSC2 0.9
09-07-2006 19:05 73-AlcatelMSC2 0.96
09-07-2006 19:05 74-AlcatelMSC2 1

The signaling of the ZET MSC to the HLR and the SCP is transmitted by the Alcatel MSC,
so some subscribers is difficult to make calls and the call completion ratio declines when the
signaling of this office direction is congested.

The intermittent disconnection of the signaling and the decline of the call completion ratio
were all caused by the problem in the clock source. According to the feedbacks from the site,
the operator didnt provide the clock reference input when commissioning the office, and all
the modules adopted free oscillation mode. Later the engineer of the operator imported the
EK8 clock of the Alcatel MSC to the module 8 and the module 10, while the other modules still
adopted free oscillation mode. Therefore, the clocks in the switch are not synchronized, and the
communication fault occurred in the CFBI.

Solution
The ZTE engineer adjusted the clock source of the system and imported the clock all from the
module 2 to solve the clock problem. In addition, it is advised to enable the 2M high-speed
signaling between the ZET MSC and the Alcatel MSC, or enable the double-signaling-point
function for them, to solve the problem that there is too heavy load on each signaling link.

12 Maintenance Experience
It Was Difficult for the Subscriber Roaming to a
Foreign Country to Log in the Network
Han Liang, ZTE Corporation

Phenomenon
When a subscriber roamed to a foreign country, he/she found that the network search
was not performed again after being rejected to log in the network during the automatic
network search, and he manually logged in the network successfully after 3~4 times of
trying.

Analysis
1. According to the messages traced by the MSC, the MSC received the reply UDTS
from the signaling transfer point three times after sending the location updating
UDT four times. The failure reason is that the address could not be translated. The
UDT location updating UDT was successfully sent to the HLR for only one time,
which matched with the traced signaling of the HLR.

2. The above phenomenon also occurred when another card in the same IMSI
number section was used, while it was normal when the card is in other
IMSI number sections was used. The reasons might be that there were many
international routes and the routing was very complicated, and the translation data
of one STP were not complete.

Solution
When tracing signaling in the HLR and the MSC at the same time, the ZTE engineer found
that the problem was caused by that the data of IMSI number section analyzed by the HSTP
was not complete. And the network was logged in successfully the first time after the data
were modified.

GSM Network Products 13


April 2007 Issue 29 Fault Instance

The HP Minicomputer Failed to Create


a Volume Group
Liu Qixing, ZTE Corporation

Phenomenon Analysis
W hen debugging the HLR DB HP The ZTE engineer judged that the C4T0D0
minicomputer, it is necessar y to create equipment had problems, but the No. of the
t wo volu me g rou p s: H LR DBVG1 a nd equipment corresponded to the physical disk.
HLRDBVG2. The ZTE engineer directly After observing this command carefully, the
copied the vgcreate -s 16 /dev/hlrdbvg1 /dev/ ZTE engineer found that there was a blank
dsk/ c4t0d0 command to run when creating before c4t0d0.
the volume group. Running command failed,
and the prompt information is as follows: Solution
vgcreate: Physical volume "/dev/dsk/" is not a The command run successfully after the

block special file. blank was deleted.

Usage: vgcreate
# vgcreate -s 16 /dev/hlrdbvg1 /dev/dsk/
[-f]
c4t0d0
[-A Autobackup]
[-x Extensibility] Increased the number of physical extents per

[-e MaxPhysicalExtents] physical volume to 8447.

[-l MaxLogicalVolumes]
Volume group "/dev/hlrdbvg1" has been
[-p MaxPhysicalVolumes]
successfully created.
[-s PhysicalExtentSize]
[-g PhysicalVolumeGroupName] Volu me G roup conf ig u ration for /dev/
hlrdbvg1 has been saved in /etc/lvmconf/hlrd
VolumeGroupName PhysicalVolumePath...
bvg1.conf
"PhysicalVolumePath": must be a block
special file.

14 Maintenance Experience
No Tone in Both Directions
Huang Wuxiang, ZTE Corporation

Phenomenon
After performing dialing tests for all the trunks between the MSC and the BSC during the
commissioning, the ZTE engineer found that there was no tone in both directions of some
trunk circuits.

Analysis
After performing induction and searching, the ZTE engineer found that the following three
groups of data had problems, and the reason for no tone phenomenon of each group of data
was also different.
Module 3, Unit 24, Subunit 3 and Subunit 4;
Module 4 and all subunits of Unit 12 and Unit 13;
Module 5 and all subunits of Unit 12.

Solution
The followings are the separate solutions for the above three groups of data:
For Module 3, Unit 24, Subunit 3 and Subunit 4.
After checking the related configuration, PCM number, CIC number of both the MSC
side and the BSC side, the ZTE engineer found that they were all normal. The trunk in the
MSC side was in the TrunkIdle status and the trunk in the BSC side was in the normal
status.

After checking the trunk connection of the two subunits, the ZTE engineer found
that the two trunks in the BSC side were connected in inverse. The problem was
solved after the two trunks were reconnected.

GSM Network Products 15


April 2007 Issue 29 Fault Instance

For Module 4 and all subunits of After checking the related configuration,
Unit 12 and Unit 13. PCM numbers, CIC numbers of
The trunk in the MSC side was in both the MSC side and the BSC
the TrunkIdle status, and the trunk side, the ZTE engineer found all
in the BSC side was in the normal these were normal. The trunk in
status. The trunk connection was the MSC side is in the TrunkIdle
also normal. When checking the status and the trunk in the BSC side
related configurations of both the is in the normal status. The trunk
MSC side and the BSC side, the connection was also normal.
ZTE engineer found that the PCN The problem still occurred after the
numbers of the four subunits of Unit DTI board and the TIC board were
12 and Unit 13 in the BSC side were replaced respectively at both the
configured in inverse. The PCM MSC side and the BSC side.
numbers of Unit 12 should be 16, 17, After checking the corresponding
18 and 19; and the PCM numbers of EDRT board of the TIC board at the
Unit 13 should be 20, 21, 22 and 23. BSC side, the ZTE engineer found
The problem was solved after the that one chip circuit of this board
PCM numbers at the BSC side were had gone moldy. The problem was
reconfigured. solved after the EDRT board was
replaced.
For Module 5 and all subunits of
Unit 12.

16 Maintenance Experience
Hardware Fault Instance
Sun Guohua, ZTE Corporation

the DTI board working normally, the board could


The the HW Cables of
work normally, and the fault was eliminated.
Both Ends Are Plugged
in Reverse Summary
The orders of the HW cable of the two ends
Phenomenon
are different, but it can not be distinguished
When the HW cables were plugged according
from the appearance. In general, there is a
to the default physical connection of the
label at the lateral surface of the each end
background, and data were configured the after
of the HW cable, which indicates that this
the cable was powered on, it was found that
end should be connected to the DT side or
some new DT boards could not run normally.
to the DSNI side. If there is no label or the
label is indistinct, it is necessary to perform
Troubleshooting
tests through exchanging the two ends when
After the jumpers and the HW cables of the
connecting the HW cable.
new backboards are checked, and compared
with the corresponding jumpers and the
HW cables of the backboards of the normal
module, it was found that the jumpers and the The Communication
HW cables were all normal. The DSNI-SP Traffic between the PIII
board and the DTI boards worked normally
MP and the Background
after being inserted to the module working
normally, so the possibility that these two
Was not Steady
boards have faults was excluded. Phenomenon
The problem also occurred after the network board The communication traff ic between the
was switched over. However, using the cable of foreground and the background of the iHLR

GSM Network Products 17


April 2007 Issue 29 Fault Instance

in Kiamusze, Heilongjiang was not steady, the network port traffic test tool.
maybe because the MP network port was not
The ZTE engineer used the Ethernet tester
steady. Therefore, data synchronization was
to test them (reserve the FV1-FV4 protection
performed at the background. Sometimes,
LEDs on the MP). When the PIII MP was
when the synchronization operation reached
connected to the switch, the tester displayed
to 54%, 81% or 19%, it could not continue
that there was slight packet loss phenomenon
any more until the background displayed
during the test.
communication overtime messages.

After connecting the PIII MP to the Ethernet


Af ter perfor ming the data packet
tester directly, the ZTE engineer found that
transmission test for the PIII and the PII MP
the connection was broken during the test,
with the network port traffic test tool, the
and there are a lot of error packets through
ZTE engineer found that all the traffic tests
the statistics information from the Ethernet
of PII MP were steady. After the PIII MP
tester.
was connected with the laptop HP NC6000
through the crossover network cable, they Then the ZTE engineer connected the PIII
could be successfully pinged each other, MP to the common computer directly, and
and the communication between them was found that the PING packets are broken
normal through being tested by the traffic test during the test, and that there are about ten
tool. Therefore, the ZTE engineer thought broken packets each time. The test result was
that there were problems in the cooperation almost the same after the FV1-FV4 protection
between the PII MP and the SWITCH. LEDs were welded away.

T he PI I I a nd t he PI I M P u sed t he The ZTE engineer performed the reliability


intel82559ER chip before, and the two faulty test of the pulse train for the PIII MP network
PIII MPs used the intel82551ER chip. At the port. Through the test result of the pulse
same time, the connection status indicator on train, it was found that the immunity of the
the SWITCH port was off sometimes, and the PIII MP network port is not good.
SWITCH indicator was on sometimes after
In addition, the ZTE engineer also tested the
the PIII MP was reset. Therefore, the ZTE
eye pattern of the 82551ER. The eye pattern
engineer thought that there were problems in
of the 82551ER was almost same with that of
the cooperation.
the 82559ER. The quality of the eye pattern

Troubleshooting is very poor when the FV1-FV4 were welded,


and it is improved a little when the FV1-FV4
The ZTE engineer sent the on-site PIII MP
were removed. In general, all the eye pattern
and the SWITCH to the equipment room for
waveforms were not satisfied, which was
being test. The connections were all normal,
related to the PCB wiring of the PIII MP board.
and the traffic is steady after being tested by
18 Maintenance Experience
Summary clock slightly dithers, it will reset the MPPP board actively,
but the MPPP board needs to periodically communicate with
Process the PIII MP network port according to the on-site
the network board, which causes the power failure alarm on
situation.
the double network boards. The current program version has
improved this loss detection mechanism, and the board is not
easy to be reset and restarted when the static interferes with

The Network Board Is Powered the clock.

down
The reasons for the power failure of the network board are as
follows: The Signaling is All Intermittently
The MPPP board is powered down and cannot be Disconnected
connected with the network board;
The attribute impedance of the PCB board of the network
Phenomenon
layer where the network board resides is not steady; The STB board adopts the HDLC link, which is implemented
There is a fault or instability when the clock board by the communication port of the MPC860 on the board.
of the network layer outputs the clock; When being interfered with, the clock at the communication
The versions of the network board PP programs port of the MPC860 is also interfered with, which causes the
under the active MP hard disk and the standby MP blocking of the communication port, and further breaks the
hard disk are not consistent. HDLC link on the STB board.

In general, if the attribute impedance of the PCB board of In addition, it is found that there is some exception on the
the network layer where the network board resides is not STB board when the HDLC of the MPC860 sends the status
steady, the DT, IWF and ASIG and other resource boards flag bit, which hinders the sending of the status flag bit,
will work unstably, such as occasional power down. and further causes the intermittent disconnection of all the
signaling links.
If there is a fault or instability when the clock board of the
network layer outputs the clock, there will be clock-related Troubleshooting
alarm or notification messages.
It is advised to replace the new program of the STB board.
If The versions of the network board PP programs under
The processing mechanism is as follows:
the active MP hard disk and the standby MP hard disk are
not consistent, the power down alarm will occur on the two Periodically detect the communication port of the MPC860
network boards in turn only when the MP is switched over. when the signaling interruption is not affected. Once the
exception is found at the communication port, initialize
The MPPP board is easy to be reset and restarted when
the communication port immediately, and recover the
there is static interference, which further causes faults in the
communication processing module in the MPC860 as soon as
double network boards. The reason is that the MPPP board
possible to avoid the intermittent disconnection of the signaling.
strictly detects the clock sent from the network board. If the

GSM Network Products 19


April 2007 Issue 29 Fault Instance

when internal-module test on the DTI board


The Port of the Board Is
is performed. The main reason is that the
Abnormal cable of the module is configured to the full

Phenomenon capacity (which is convenient for the capacity


expansion in the further), and there are eight
When the man-machine command switching
cables inserted in the DSNI-SP slot of the
for the SYCK board, the FBI board and other
backboard that were connected to the trunk.
boards is performed at the background, it
In fact, the board of the trunk layer is not
is found that the switching of the master
configured to the full capacity, which means
board in some slot is successful, but the
that one end of some cables is inserted in
switching of the boards in the other slots
the DSNI-SP interface, but the other end is
is unsuccessful, and the system prompts
suspended.
failure in the standby ports. However, the
switching is successful by directly pressing The cable actually becomes a small antenna,
the board switching button at the foreground and the imported interference jumps into
of the rack, which indicates that the board the power of the DSNI-SP board and the
is fine, and that there are exceptions in the ground. This causes the signal imperfection
background server or the client end. of t he H W cable, a nd t he i nc re a se of
communication error codes in the module,
Troubleshooting which further causes the power failure of the
Restart or reinstall some software modules of floor.
the background server or the client end.
Troubleshooting
The suspended cable should be unplugged if
there is no load at the output end of the cable.
The DTI Board Is Reset To guarantee that the cable is not inserted by
when the DSNI Board mistake during the capacity expansion, label
in the Adjacent Slot Is the cable.

Plugged or Unplugged
Phenomenon
When the DDSN board is used separately
Signaling Byte Error in
(which means that the active and the standby
the GSM-HLR in Libya
DDSNs are not used together), the DTI
Phenomenon
board is reset when the DSNI board in the
The subscriber registration information in
adjacent slot is plugged or unplugged. It
is found that there are many error codes

20 Maintenance Experience
the CS was not consistent with that in the HLR. The tracing this slot, and the test was unsuccessful. However, when the
information of the MTP3 layer in the HLR was correct, but normal signaling board was put back to the former normal
the tracing information of the MTP3 layer in the CS was slot, the test was also unsuccessful. Therefore, the ZTE
not correct, and the subscribers having problems were in engineer thought that the backboard of the former slot had
the HLR1. It was judged that there were problems in the damaged this signaling board, and this signaling board
signaling sent by the HLR1. further damaged the former normal slot.

Troubleshooting Summary
After exchanging the 2M signaling boards of the two If it was judged that the signaling had problems, maybe it
slots, and switching the active MP to the left slot, the ZTE is caused by the MP level DSNI board, the MPMP or the
engineer found that the service was abnormal after the test. signaling board.

After switching back the active MP of the right slot, the ZTE The MPs of the left and right slots respectively correspond
engineer found the problem still occurred, which indicated to a set of data bus connected to the interfaces of the MPMP
that the signaling board was normal. and the signaling board, so the signaling board also use
different bus interfaces after the MP is switched. Therefore,
After switching back the active MP to the left slot, the ZTE
it cannot be confirmed that the MP is abnormal.
engineer found that sending and receiving are on the two
MPMPs respectively when using the Moniproj tool. After
unplugging the COMM1 (the left MPMP) that is in the
data sending status, the ZTE engineer found that several
problems occurred, such as location update failure and
the telecommunication service being not inserted when
performing the test,

There was no sending/receiving error, each error item


of the DPRAM was 0, and there was no alarm about bus
errors through checking from the Moniproj. During the
whole switching operation, the communication between
the modules was normal, so the main processor to the main
processor (MPMP) board is normal.

After modifying the configuration data and changeed the


slot of the signaling board, the ZTE engineer found that
the test service was normal after switching the MP, which
indicated that the signaling board was normal.

In order to verify whether the former slot had problems,


the ZTE engineer switched the normal signaling board to

GSM Network Products 21


April 2007 Issue 29 Fault Instance

Check whether there are error data among indicators are on at the same time, but the
the communication between modules by the tones are different.
Moniproj tool, such as data loss, frame error
The similar problem also occurs on the FBI
and the data error in the physical layer.
board. If the active and the standby indicators
Then use exclusive methods or substitution on the active and standby FBI boards are
methods to check whether the board has on at the same time, the the communication
problems. between modules will be interrupted.

Before or after performing the substitute test Troubleshooting


for the board and the slot, it is necessary to
Logically update the EPLD board, and it
observe the pins on the board carefully, and
is advised to replace the backboard of the
check whether they are twisted or burned, to
network layer.
avoid the inter-influence between the faulty
board and the slot.

The Active and Standby


Status Indicators of the
DSNI-SP Board Are on
At the Same Time
Phenomenon
The active and the standby indicators of one
pair of DSNI-SP boards are on at the same
time, and only the tones are different, which
causes the power failure of all the boards in
the trunk layer.

The reason is that the active and standby


signal of the active and the standby boards
are not logically interlocked, and the attribute
impendence of the PCB on the network layer
is not controlled well, which causes that the
active and the standby signal of the fixed
level oscillates. The duty ratios are different,
so it seems that the active and the standby

22 Maintenance Experience
Special Document

The Principle of the ECC Board and Its Use in


the MSS System
Sun Guohua, ZTE Corporation

Summary
The SDT board adopted by the current MSS system doesn't have echo cancellation
function, but some regions have requirements for the echo cancellation of the SDT board.
Develop the SDT board with echo cancellation function to meet the requirement or develop
the ECC board to implement the echo cancellation function indirectly. Developing the ECC
board can meets the future requirements after comparing the correcting time, performance/
price ratio, system implementation, current resource and the production stopping possibility
of the DTEC board.

At present, the resource-type echo cancellation board ECC is used as a kind of common
resource, similar to the ASIG. When the call connection requires echo cancellation, apply
for the echo cancellation resource. It needs to perform wire wrapping processing on the
ECC board when performing call connection. The trunk resources on the SDT board
are not possible to be used at the same time, so, there is no need to configure the echo
cancellation circuit and the SDT trunk circuit in the form of 1:1 when configuring the
resource-type echo cancellation circuit. In this way, not only the echo cancellation circuits
can be shared, but also the SDT board can substitute the DTEC board

The Principle of the ECC0209 Board


The hardware principle diagram of the echo cancellation board is shown in Figure 1.

GSM Network Products 23


April 2007 Issue 29 Special Document

Figure 1 Hardware Principe Of Echo Suppressor

The 8MHW signal from the T network is sent to the switching chip MT8986, is tapped off
to 2MHW by the MT8986, and then is sent to the echo cancellation chip. After the echo is
eliminated, the 2MHW is multiplied to the 8MHW through the MT8986, and then is sent back
to the T network.

The CPU reads the information in the module, performs proper processing on it, packs it as
HDLC packets and sends the packets to the corresponding HDLC channel. Finally the packets
are sent to the T network in the difference mode after timeslot combination with the 8MHW
data.

This board also has two DSP chips. The 8MHW from the T network can be sent to the input
end of the MT8986 directly, and it also can be sent to the input end of the MT8986 after going
through the mutual conversion between the A law and U law of the signal through DSP through
the jumper selection and adjusting the volume. In the same way, the output of the MT8986 can
be sent to the T network directly, and it also can be sent the T network after being processed by
the DSP chip.

The Jumper of the ECC0209 Board


X12, X13
When short circuiting the 3-2 pins, the 8MHW from the T network is sent to the input

24 Maintenance Experience
end of the MT8986 directly; when short circuiting HL4
the 2-1 pins, the 8MHW from the T network is
T he i nd icator i nd icates t hat t he DSP2 r u ns
sent to the input end of the MT8986 after being
normally.
processed by the DSP chip.

X14, X15 ECC0209 Parameter Index


When short circuiting the 1-2 pins, the 8MHW from
Each board uses four MT9300 echo cancellation chips,
the MT8986 is sent to the T network after being
namely four subunits, which totally have 120 (30*4=120)
processed by the DSP chip; when short circuiting
echo cancellation circuits, and can perform echo cancellation
the 2-3 pins, the 8MHW from the MT8986 is sent to
for 120 calls. Each board has two external 8M HWs.
the T network directly.

X20
The Position of the ECC0209
Select different jumpers when welding different Board in the Peripheral Module
FLASH (D2) chips on the hardware. When short
The DT layer (backboard BDT) has 16 (8 pairs of)
circuiting 2-1 pins, the FLASH chip on the board
slots. Each pair of slots can be inserted with one
is produced by INTEL; when short circuiting the
ECC board, and it is advised to insert the board at
2-3 pins, the FLASH chip is produced by the SST
the left slot. If the left slot of this pair is inserted
Company.
with one ECC board, the right slot must be idle.

X8
The MDT layer (backboard BMDT) has 16 (8 pairs
When short circuiting the 2-3 pins, use the TRU050
of) slots. Each slot can be inserted with the ECC
phase-locked loop, and D14 is welded on the board;
board. At the back of the BMDT, there are two (3
when short circuiting the 2-1 pins, use the phase-
8 HW cable) interfaces leading out the BD cables.
locked loop produced by ZTE, and D15 and G3 are
When the ECC board is inserted into the BMDT,
welded on the board.
the cables leaded out by the HW should be inserted
into the 25th ~32nd pins.
ECC0209 Indicators
HL1 The Echo Cancellation Principle
T he i ndicator i ndicates that the boa rd r u ns
The PSTN subscriber access adopts 2-wire loop. When
normally;
performing call services between the PSTN subscriber

HL2 and the mobile subscriber, echoes will be generated to the

The indicator indicates that the board is in the error mobile subscriber when converting the 2-wire to the 4-wire

status; (as shown in Figure 2).To effectively cancel the echo to the
mobile subscriber, it is necessary to add an echo canceller in
HL3
the mobile network.
T he i nd icator i nd icates t hat t he DSP1 r u ns
normally;

GSM Network Products 25


April 2007 Issue 29 Special Document

Figure 2 Echo Cancellation Principles

At present, the echo cancellation hardware is implemented by the DTEC board, which can
only cancel the echo for the tone in the receiving direction of the E1 interface. During the
call, the echo cancellation can be performed for the tone in the receiving direction of each
trunk interface circuit according to the requirement. Because the echo cancellation function
and the trunk circuit are in one-to-one correspondence, and not every incoming/outgoing call
needs echo cancellation, this echo cancellation method can not make good use of the echo
cancellation resource.

The resource-type echo canceller adopts the common echo cancellation resource to implement
the echo cancellation function. The basic idea of the resource-type echo cancellation board is
that the echo cancellation resource is common, and the connection of this call in the T network
needs one time of wire wrapping through the ECC board if the call needs echo cancellation.

According to the hardware design, there are two 8M HWs on the ECC board: S HW and R
HW, which are connected to the T network, The R HW connects the voice channel timeslot in
another direction, which acts as the echo cancellation processing reference timeslot; the S HW
connects the voice channel timeslot needing echo cancellation, which enters the ECC board
through the S HW in, and is connected out of the ECC board through the S HW out after echo
cancellation.

Figure 3 and Figure 4 describe the timeslot connection relations when the call in the trunk
terminating/originating office needs echo cancellation.

26 Maintenance Experience
Figure 3 Connection Of Echo Cancellation For Figure 4 Connection Of Echo Cancellation For The
The Trunk Incoming Call Trunk Oncoming Call

The voice channel direction of the echo cancellation on the ECC board is SHW inRHW out, and the reference voice
channel direction is RHW inSHW out.

The PRI Interface Configuration Description


Wang Bin, ZTE Corporation

Before configuring the data, the following data should be to the ISDN switch directly, and the ISDN switch is used

prepared and agreed on by both parties: for the subscriber side. In this case, the subunit should be

Whether the local office is subscriber side or configured as ISDN PRA (network side). If the switch is used

network side. In general, the network side is as the PBX and connected to other switches, the subunit

adopted when being connected to the gateway; should be configured as ISDN UPRA (subscriber side).

Whether the tr unk adopts E1 interface or T1


For the DT board of the T1, the subunit is configured as
interface;
ISDN T1 PRA (network side) or ISDN T1 UPRA (subscriber
The timeslot occupied by the D channel.
side) in the unit configuration. In general, the subunit is
configured as ISDN T1 PRA(network side), as shown in
Physical Configuration
Figure 1.
For the DT board of the E1, the subunit is configured as
ISDN PRA (network side) or ISDN UPRA (subscriber side)
in the unit configuration. In general, the switch is connected
GSM Network Products 27
April 2007 Issue 29 Special Document

number, so the alert box will pop up when


adding the trunk group, as shown in Figure 3.

Figure 3 Prompt Information

2. Configure the PRA trunk circuit


Figure 1 Subunit Cofiguration
Check PRA trunk group on the
Note: The old PP version of the DT1 doesn Pra trunk circuit assignment page,
t support the ISDN signaling. In this case, and click the modify button. Then
after the physical connection is normal, the select the corresponding circuits as
status indicator of the subunit flashes slowly. channel B or channel D For the E1
The dt304g0030408 can support the ISDN board, the No.7 circuit is allocated
signaling, and the status indicator of the as channel D, and the No.1~No.15
subunit should flash quickly. and the No.17~No.23 circuits are
allocated as channel B.
Truck Configuration 3. Con f ig u re t he outgoi ng rout e,

1. Add the t r u n k g roup, selecting outgoing route group and outgoing

PRA for Trunk Channel Type, and route chain, and the configuration

DSS1 for Incoming (Outgoing) method is same to that of configuring

Line Signaling Mark, as shown in the trunk route.

Figure 2.
4. Check the circuit status through
the dynamic management after the
trunk configuration is finished, as
shown in Figure 4.

Figure 2 Adding the Trunk Group

Note: Uncheck default origin digit on the


PRA Trunk page. The default origin digit Figure 4 Dynamic Management
checked by default, but there is no specific
28 Maintenance Experience
Experience Exchange

Precautions for Shuffle


Chen Yu, ZTE Corporation

The current network of an office has three BSCs, namely, BSC1, BSC2 and BSC3. There
are 125 sites and 373 cells in all. During the earlier stage of the site commissioning, the
same region is cross-distributed to different BSCs. After considering the network layout
and the influence on the network index in the future, it is planned to adjust the sites among
these three BSCs, which is called Shuffle.

When the sites are adjusted, the data of the interconnected MSC should be adjusted
according to the situation. When changing the data of the MSC side, do not omit the special
service number. If the HLR has area restriction, pay attention to the data change when
adjusting the corresponding cell.

1. The on-site special service number only relates to the LAC. The Telephone Code
of the corresponding special service number doesnt change when the cell is
adjusted from one BSC to another BSC. It neednt to be modified here, as shown
in Figure 1.

Figure 1 Special Service Number Management


GSM Network Products 29
April 2007 Issue 29 Experience Exchange

2. For Zone Code, there is an area restriction code defined in the HLR under the
area covered by each site. After the site is changed, the corresponding data should
be adjusted to the BSC where the site belongs. The corresponding data should be
modified here, as shown in Figure 2.

Figure 2 Zone Code Configuration

Based on the analysis above, cooperate with the BSC to change the sites. Only the CI data of
the cell needs to be adjusted to the corresponding LAC number at the MSC side. According to
the local situation, as shown in Figure 17, the corresponding CI items in GCI Configuration
and Zone Code Configuration need to be adjusted.

It is advised to add the corresponding data again after deleting the corresponding items of the
CI to be adjusted, instead of modifying them. The retransmitting table will take effect after
being adjusted.

Select Statement of the Database


Yang Yinping, ZTE Corporation

T h e d a t a b a s e m ay b e o p e r a t e d w h e n Select X from TABLE where CONDITION


implementing commissioning debugging or
Examples:
maintenance on the site. The query statement
is the most f requently used among the 1. Select * from a table
database statements. Therefore, knowing Select * f rom D_HIS_ DI A RY1
well the query statements is helpful to locate where nodeID=200
and process faults. (Select all the columns from the
30 Maintenance Experience
table D_HIS_DIARY1 where the nodeID is 200) 2. Association table query
Select imsi, isdn from baseinfo where secver=0x00 Select imsi from baseinfo a, camelinfo b where
(Select the IMSI and the MSISDN of the subscriber a.imsi=b.imsi and a.camelinfoind=0x00
from the table baseinfo where the authentication
3. Association table query
version is comp128-1)
Select imsi from camelinfo where imsi in (select
Select count(distinct(imsi)) from camelinfo
imsi from baseinfo where isdn>0x13828791004ffff
(Count the quantity of intelligent subscribers in the
and isdn<0x13828799999ffff)
database. The camelinfo table has two records for
each IMSI: OCSI and TCSI, so the statistics of the (Query the IMSI of the camel subscriber in one ISDN
same IMSI will not be repeated when using distinct nu mb e r se ct ion. Be cau se t he I MSI d ist r ibut ion is
as the key word) discontinuous, make use of other continuous parameters to
set this condition in order to carry out the query)

The Problems Occurs When Debugging the


Monitoring Service
Zhang Hongliang, ZTE Corporation

During the course of debugging the monitoring service, server. Because the process can run only after the watchdog
if there is a problem occurring in any link such as X25, program is started, it needs to add the RU_MN_AGENT.
MODEN parameter setting, S20 chip, monitoring trunk EXE and the WINNTTCP.EXE shortcut keys to the startup
board, E1 transmission, cable from the X25 to the MODEM0 options in order that the process can start automatically
and the telephone wire from the S20 to the MODEM, the when the computer is restarted.
monitoring service will be abnormal. Therefore, it should
be guaranteed that each link should be normal during the Fault Examples
course of debugging the monitoring service.
When the monitoring service is being debugged, it is found

In addition, if the client end is used as the SORM server, the that the 1st X25 link always can not be connected after the

monitoring center can not automatically establish connection link to the monitoring center is put through. This problem is

with the X25 after being restarted, and it needs to restart the solved after theS20 is restarted.

RU_MN_AGENT process at the client used as the SORM


GSM Network Products 31
April 2007 Issue 29 Experience Exchange

One day the power in the equipment room is Because there is no UPS in the cur rent
off, when the client end used as the SORM equipment room, the power is often off. In
server is being monitored. The monitoring order not to affect the monitoring service, the
service can not be connected normally after operator has moved the client end used as the
the client end is started, while the MODEMs SORM server. However, the server cabinet
of both ends have responded. However, the can not be connected to the monitoring center
MODEMs can not be started when the X25 is after it is powered on, and the MODEM of
restarted, and the system prompts MODEM the client end cannot communicate with that
NOT READY(NO CLOCK). The setting of the monitoring center.
of the opposite end is normal after being
The same problem also occurs after the
checked.
MODEM is reset. Maybe the telephone wire
Connect the MODEM to the port 1, firstly, from the S20 to the MODEM has problems,
use the AT&F to recover the setting to its but it is normal after being checked by the
factory setting, set each parameter according millimeter. However, this problem disappears
to the configuration document, and reconnect after the telephone wire from the S20 to the
the MODEM to the X25 card. Then the MODEM is replaced.
communication with the monitoring center
will be normal.

32 Maintenance Experience
No.8 March 2007 Issue
29
Maintenance Experience
www.zte.com.cn

GSM Network Products

Typical Cases
Experience Collection

Vous aimerez peut-être aussi