GSM Frequency Planning

GSM_P&O__01_200904 GSM Frequency Planning

Objective

Frequency Planning

Neighbor Cell Planning

BSIC Planning

Reference

GSM Cellular Network Design and Optimization

GSM Frequency Planning

GSM Frequency Hopping Principles

i

Contents

Chapter 1 Frequency Planning .................................................................................................................... 1

1.1 Cellular Structure Formation Rule ....................................................................................................... 1

1.2 Interference Model............................................................................................................................... 3

1.3 Frequency Multiplexing Technology and Interference Analysis ......................................................... 8

1.4 Packet Frequency Multiplexing Technology ....................................................................................... 8

1.4.1 4 x 3 Multiplexing Technology .................................................................................................. 8

1.4.2 3 x 3 Multiplexing Technology ................................................................................................ 14



1.4.5 MRP Multiple Frequency Multiplexing MRP ......................................................................... 18

1.4.6 Concentric Cell Technology .................................................................................................... 25

1.5 Cell Splitting ...................................................................................................................................... 30

1.6 Several Common Immunity Technology to Interference ................................................................... 31

1.6.1 Discontinuous Transmission (DTX) ........................................................................................ 32

1.6.2 Frequency Hopping (FH) ......................................................................................................... 32

1.6.3 Dynamic Power Control (DPC) ............................................................................................... 37

1.6.4 1 x 3 Multiplexing + Radio Frequency FH + DTX + DPC ..................................................... 37

1.7 Conclusion on Principles of Frequency Assignment in Engineering ................................................. 38

Chapter 2 Neighboring Cell Planning ....................................................................................................... 41

Chapter 3 BSIC Planning ........................................................................................................................... 49

1

Chapter 1 Frequency Planning

Key points

Frequency multiplexing and interference model analysis based on ideal cellular

structure

Several common immunity technologies to interference

1.1 Cellular Structure Formation Rule

In ideal situation, a base unit (base station area) of cellular structure is a

regular hexagon (handoff border). A certain number of regular hexagons

constitute a radio cluster. A full mobile network coverage is composed of two

adjacent radio clusters.

The radio cluster, a base unit of Frequency Multiplexing (FR), allocates all of

the available channels in a radio cluster to every base station area or sectoral

cell. Two same radio clusters are able to be adjacent to each other and ensure

mapping relationship between each base station areas or sectoral cells. The

channel group assigned to every base station area or sectoral cell is fixed.

Therefore mapping base station areas or sectoral cells in any adjacent radio

clusters are all co-frequency areas. This forms a comprehensive co-frequency

multiplexing pattern.

The radio cluster must meet the following conditions:

1) The radio clusters are able to be adjacent to each other.

2) The distance between any two co-frequency multiplexing area centers in

the adjacent radio clusters should be equal.

GSM_P&O_ _01_200904 GSM Frequency Planning

2

A

B

C

D

E

F

G

A

B

C

D

E

F

G R

60o

i

j

D

Figure 1.1-1 Constitution of a radio cluster

As shown in the above figure, i and j are two parameters. Given the two

parameters different values (cannot be 0 at one time), any area can be

reached from a certain area. Based on triangular relationship shown in the

above figure, the distance D between two co-frequency multiplexing areas

is:

22 jijiD

The number of base stations N included in the radio cluster based on the

above distribution is:

22 jijiN

Given the distance between the centers of two adjacent base station areas is

1, and semi diameter of base station area is R, then:

3/1R

Define RDq / as co-frequency multiplexing distance protection

coefficient or called as co-channel interference attenuation factor:

NR

Dq 3

(

1

-

1

) (

1

-

2

) (

1

-

3

) (

1

-

4

)

Error! Use the Home tab to apply 1 to the text that you want to appear here.

3

1.2 Interference Model

1. Co-frequency interference protection ratio B

Under the condition wherein wanted signals from the Tx end of a receiver

meet the defined quality, the parameter indicates the minimum ratio of

wanted RF signals to unwanted RF signals. Usually, the value of this

parameter is represented as dB.

2. Estimation on carrier-to-interference ratio in an N-multiplexing radio

cluster

A

B

C

D

E

F

G

A

B

C

D

E

F

G

A

B

C

D

E

F

G

A

B

C

D

E

F

G

A

B

C

D

E

F

G

A

B

C

D

E

F

G

A

B

C

D

E

F

G

A

B

C

D

E

F

G

A

B

C

D F

G

A

B

C

D

E

F

G

A

B

C

D

E

F

G

A

C

D

E

F

G

A

B

C

D

E

F

G

A

B

C

D

E

F

G

A

B

C G

A

B

E

F

G

A

B

C

D

E

F

G

A

D

E

F

A

B

C

D

E

Figure 1.2-1 Interfering resource

Regarding wave propagation characteristic, it could be described with the

preceding general model:

DiffkkkHdkHkdkkPL effeff 765loglog4log3log21


4

Since it is the ideal cellular system that has been observed and studied, both

emission power of each cell and antenna antenna apparent height are the

same, and there is no diffraction loss. Therefore the carrier-to-interference

ratio can be calculated as follow:

M

k

dkk

dkk

M

k

PL

PL

M

k

PL

PL

M

k

PLP

PLP

M

kk

kHeff

Heff

k

kkt

t

I

C

I

C

1

10/log)log42(

10/log)log42(

1

10/

10/

1

10/

10/

1

10/)(

10/)(

1

10

10

10

10

10

10

10

10

Indicate effHkkk log42'2

, d as cell semi diameter R, kd as

propagation distance D from each interfering resource to this cell.

As shown in figure 1-2, there are 6 the most intense interfering resources

around each cell, and 6 (or 12) the secondary most intense interfering

resources.

10/'210/'2

10'2

12

1

10/2log'26

1

10/log'2

10/log'2

)2(126

1010

10

kk

lk

k

Dk

k

Dk

Rk

DD

R

I

C

Indicate '2k/10 ( This is so-called propagation path loss slope

determined by the actual terrain environment.)

2

126)2(126

q

DD

R

I

C

Logarithm, it is:

)2

126log(10log'2)(

qkdB

I

C

(

1

-

6

)

(

1

-

5

)


5

No loss generality, indicate 40'2 k , 4 .

So

dB5.06log10)2

126log(10

4

.

We can see that contribution to interference made by the secondary most

intense interfering resource in the second circle is much less than that of the

most intense interfering resource in the first circle, which can be negligible.

Now we have established an interference model under ideal cellular

environment. We will use this model to study its interference when various

common multiplexing methods are introduced later.

3. Co-frequency interference possibility )/( BICP

Actually, because of non-ideal site location and rise-and-fall characteristic topography,

when mobile station is on the move, received signals are influenced by Rayleigh fast

fading and Gauss slow fading. No matter it is signal or interference, before it reaches

mobile station, its field strength instantaneous value and median value are all random

variables. Even though mobile station stays still, as a result of various existed

interference including movement of surrounding moving objects, its field strength

instantaneous value and median value are still random variables.

We can see that the value of receiver Rx end IC / is not static but a random variable.

Only if BIC / , there is no interference. Co-frequency interference appears with

certain possibility.

According to CCIR740-2 report, in 1979 France comes up with the idea that

when multipath fading complies with Rayleigh distribution and shadow

fading complies with Gauss distribution, co-frequency interference

possibility is :

du

uBICP

uBIC 10/)2(

2

101

}exp{1)/(

In the formula, u is integration variable, is standard deviation of

signal and interference, IC

.


6

BICZ p

100

10-1

10-2

10-3

10-4

20 40 60

dB

= 12

= 0 = 6 = 8

Figure 1.2-2 Co-frequency interference possibility

Co-frequency interference possibility in the typical circumstance is shown

as above.

Without losing its generality, indicate = 6, interference possibility

)/( BICP =0.1,

dBZ p 12 concluded from the chart, GSM

network requires co-frequency interference protection ratio B be less than 9

dB, generally B = 12dB in engineering. Therefore, in ideal interference

model carrier-to-interference ratio must be more than: 9(12) + 12 = 21 dB

(24dB).

William C.Y. Lee believes that indicating 6 dB margin is enough, so it is

concluded that in ideal interference model carrier-to-interference ratio must

be more than 9(12)+6=15dB (18dB).

4. Near End- Far End interference


7

A

B

C

D

d2

d1

2

1 d2

d1

Figure 1.2-3 Near end- Far end interference

1 Cell 1

2 Cell 2

According to interference model, indicate mobile station B relative to

mobile station A

dBd

dkdB

I

C9log'2)(

2

1

, then

69.11

2 d

d

. If

frequency used by mobile station B is adjacent to that of mobile station A,

when

69.11

2 d

d

, adjacent frequency interference protection ratio does not

match the condition, which causes call drop. The same circumstance also

appears in adjacent cells.

Lets check another extreme circumstance: given that Tx power of two

antennas in a cell is 34 dBm, the level received on spot D is -85 dBm, base

Comment [m1]:


8

station sensitivity is -110 dBm If uplink and downlink power are balanced,

then transmission power of mobile station D is -110+(34-(-85))=9dBm.

Now, when the very near mobile station C powers on, if it is working with

the maximum transmission power 30 dBm (1 W), given that the path loss

when the signal reaches cell 2 is same to that reaching mobile D, then

interference signal received by cell 2BTS is : 30-(34-(-85)= -89 > -110 + 9.

Therefore, call drop occurs.

1.3 Frequency Multiplexing Technology and Interference Analysis

Frequency multiplexing is a kind of technology commonly used in a GSM

network. It applies the same frequency to cover different areas. In addition,

it keeps certain distance between these areas using the same frequency, and

the distance is called co-frequency multiplexing distance.

If directional antenna is used, it is recommended to adopt 4 x 3 multiplexing

method. In certain areas with heavier traffic, other multiplexing methods can

be adopted according to machine capability, such as 3 x 3 and 2 x 6. No

matter which method it is adopted, its basic principle is that it should meet

the requirements of interference protection ratio after considering different

propagation conditions, different multiplexing methods, and multiple

interference factors. They are shown as follows:

Co-frequency protection ratio C/I 9 dB

Adjacent frequency interference protection ratio C/I -9 dB

400 kHz adjacent frequency interference protection ratio C/I -41 dB

1.4 Packet Frequency Multiplexing Technology

1.4.1 4 x 3 Multiplexing Technology

There are a variety of frequency multiplexing structures used by GSM, such

as 4 x 3, 3 x 3, and 2 x 6. Usually, all multiplexing methods are to classify


9

limited frequencies into a certain number of groups. In sequence these

groups form a cluster of frequency assigned to adjacent cell (shown in the

following figure). According to advices of GSM system criteria, 4 x 3 is

commonly used in various GSM systems. 4 x 3 multiplexing method is to

divide frequencies into 12 groups assigned to 4 stations in turn. That means

3 frequency groups can be used in each station. As result of long

multiplexing distance in this frequency multiplexing method, it can reliably

meet the specifications of co-frequency protection ratio and adjacent

frequency interference protection ratio required by GSM system. Therefore,

it makes GSM network operate in fine quality and good security.

A3

D2B1

D1

D3

C1B3

C2

B2

C3

A1

A2

A3

D2B1

D1

D3

C1B3

C2

B2

C3

A1

A2

A3

B1

B3B2

A1

A2

A3

B1

A1

A2A3

D2B1

D1

D3

A1

A2

A1

A3

D2B1

D1

D3

C1B3

C2

B2

C3

A1

A2

Figure 1.4-1 43 multiplexing

Indicating the value of cellular hexagon side length as 1, from the above

figure and the preceding interference models, it can be concluded as:


10

dBdBI

C18

)2.7(28

2log10)(

52.352.3

52.3

Subtracting the margin of 6dB suggested by William C.Y. Lee, the value is

exactly 12 dB.

Discussion on 4 x 3 frequency packet and multiplexing model applied in

engineering:

As the name implies, 4 x 3 multiplexing divides usable frequencies into 4 x

3 = 12 groups, and respectively marks them as A1, B1, C1, D1, A2, B2, C2,

D2, A3, B3, C3, and D3. Take the following table as an example:

A1 B1 C1 D1 A2 B2 C2 D2 A3 B3 C3 D3

1 2 3 4 5 6 7 8 9 10 11 12

13 14 15 16 17 18 19 20 21 22 23 24

25 26 27 28 29 30 31 32 33 34 35 36

Indicate A1, A2, and A3 as a large group, and assign it to 3 sectors in a base

station. Indicate B1, B2, B3, C1, C2, C3, D1, D2, and D3 as a large group,

and assign it to 3 sectors in an adjacent base station. Obviously, there are 6

frequency multiplexing methods as follows.


11

A1

A2

A3

D1

D2

D3

B1

B2

B3

C1

C2

C3

A1

A2

A3 B1

B2

B3

C1

C2

C3

A1

A2

A3

A1

A2

A3

1

A1

A2

A3

C1

C2

C3

B1

B2

B3

D1

D2

D3

A1

A2

A3 B1

B2

B3

D1

D2

D3

A1

A2

A3

A1

A2

A3

2

A1

A2

A3

D1

D2

D3

C1

C2

C3

B1

B2

B3

A1

A2

A3 C1

C2

C3

B1

B2

B3

A1

A2

A3

A1

A2

A3

3

A1

A2

A3

B1

B2

B3

C1

C2

C3

D1

D2

D3

A1

A2

A3 C1

C2

C3

D1

D2

D3

A1

A2

A3

A1

A2

A3

4

A1

A2

A3

C1

C2

C3

D1

D2

D3

B1

B2

B3

A1

A2

A3 D1

D2

D3

B1

B2

B3

A1

A2

A3

A1

A2

A3

5

A1

A2

A3

B1

B2

B3

D1

D2

D3

C1

C2

C3

A1

A2

A3 D1

D2

D3

C1

C2

C3

A1

A2

A3

A1

A2

A3

6

(1~6) Method (1-6)

If following the above frequency sequence packet method, there will be no

problems about co-frequency occurred in adjacent base stations. However,

adjacent frequency in end-on cells still exists: (see the positions indicated by

red arrowheads in the above figure)

Method 1: D1---A2; Method 2: D2---A3; Method 3: D1---A2;

Method 4: D2---A3; Method 5: D3---A1; Method 6: D3---A1.

Therefore, lets switch to another frequency packet method. See it in the

following table:

Comment [m2]:


12

A1 B1 C1 D1 A2 B2 C2 D2 A3 B3 C3 D3

1 2 4 3 5 8 7 6 9 11 10 12

13 14 16 15 17 20 19 18 21 23 22 24

25 26 28 27 29 32 31 30 33 35 34 36

The same 6 multiplexing methods:

No end-on adjacent frequency in Methods 1 and 4; Method 2: C1---A2;

Method 3: B2---A3;

Method 5: C1---A2, B2---A3, D3---A1; Method 6: D3---A1.

(1~6) Method (1-6)

A1

A2

A3

D1

D2

D3

B1

B2

B3

C1

C2

C3

A1

A2

A3

B2

B3

C1

C2

C3

A1

A2

A3

A1

A2

A3

1

A1

A2

A3

C1

C2

C3

B1

B2

B3

D1

D2

D3

A1

A2

A3 B1

B2

B3

D1

D2

D3

A1

A2

A3

A1

A2

A3

2

A1

A2

A3

D1

D2

D3

C1

C2

C3

B1

B2

B3

A1

A2

A3 C1

C2

C3

B1

B2

B3

A1

A2

A3

A1

A2

A3

3

A1

A2

A3

B1

B2

B3

C1

C2

C3

D1

D2

D3

A1

A2

A3 C1

C2

C3

D1

D2

D3

A1

A2

A3

A1

A2

A3

4

A1

A2

A3

C1

C2

C3

D1

D2

D3

B1

B2

B3

A1

A2

A3 D1

D2

D3

B1

B2

B3

A1

A2

A3

A1

A2

A3

5

A1

A2

A3

B1

B2

B3

D1

D2

D3

C1

C2

C3

A1

A2

A3 D1

D2

D3

C1

C2

C3

A1

A2

A3

A1

A2

A3

6

Comment [m3]:


13

Therefore, we recommend the above frequency packet multiplexing

methods 1 and 4. Base station of each system may not be right on the grid,

hence it can be all right if we adopt the preceding packet method classified

by frequency sequence. However, adjacent frequency problem occurred in

adjacent cells should be avoided.

We can see from the above example table that the largest station type is of

7.2M bandwidth. It can be concluded that this multiplexing method cannot

satisfy the requirement of network capacity expansion in the areas with

heavy traffic, as a result of its low frequency utilization rate. In some large

and medium cities with high population density, after many times

expansion, station distance is less than 1 km, coverage semi diameter is no

more than several hundred meters and some sites even cover 300 m.

Therefore, it is not realistic to increase network capacity by adopting

large-scaled cell splitting technology. There are two methods to solve the

problem of ever-increasing network capacity demand. One is to develop

GSM900/1800 two-frequency network, and the other is to adopt the closer

frequency multiplexing technology.


14


A3

C2B1

C1

C3

B3B2

A1

A2

A3

C2B1

C1

C3

B3B2

A1

A2A3

C2B1

C1

C3

B3B2

A1

A2

A3 C1

A1

A2

A3

C2B1

C1

C3

B3B2

A1

A2

A3 C1

A1

A2

A3

B1

B3B2

A1

A2

Figure 1.4-2 3 3 multiplexing

Indicating the value of cellular hexagon side length as 1, from the above

figure and the preceding interference models, it can be concluded as:

dBdBI

C3.13

)57.5(2)7(2

2log10)(

44

4


engineering:

3 x 3 multiplexing generally adopts baseband frequency-hopping, or it

adopts without frequency-hopping. However, it does not perform well. 3 x 3

multiplexing divides usable frequencies into 9 groups, and respectively

marks them as A1, B1, C1, A2, B2, C2, A3, B3, and C3, as follows:


15

A1 B1 C1 A2 B2 C2 A3 B3 C3

1 2 3 4 5 6 7 8 9

10 11 12 13 14 15 16 17 18

19 20 21 22 23 24 25 26 27

28 29 30 31 32 33 34 35 36

There are the following two multiplexing methods:

Method 1: Adjacent frequency in no-end-on cell; Method 2: C1---A2,

C2---A3, C3---A1.

Obviously, multiplexing method 1 is better.

A1

A2

A3

B1

B2

B3

C1

C2

C3

C1

C2

C3

B1

B2

B3 A1

A2

A3

A1

A2

A3

C1

C2

C3

B1

B2

B3

A1

A2

A3

C1

C2

C3

B1

B2

B3

B1

B2

B3

C1

C2

C3 A1

A2

A3

A1

A2

A3

B1

B2

B3

C1

C2

C3


16


A3

A1

A2

A3

A1

A2

A3

A1

A2

A3

A1

A2A3

A1

A2

A3

A1

A2

A3

A1

A2

Figure 1.4-3 1 x 3 multiplexing

Indicating the value of cellular hexagon side length as 1, from Figure 1-7

and the preceding interference models, it can be concluded as:

dBdBI

C43.9

)36.4(25

2log10)(

44

4


engineering:

1 x 3 is one of the closest methods in frequency multiplexing. It is generally

adopted in synthesizer hopping system. Meanwhile DTX, power control,

antenna diversity and other immunity technologies to interference are used

to make up for interference degradation caused by shortened multiplexing

distance. All non_bcch frequencies are divided into three groups: A1, A2,

and A3. Each of them is MA of three sectors in each base station, as shown

below:


17

A1 1 4 7 10 13 16 19 22 25 28 31 34

A2 2 5 8 11 14 17 20 23 26 29 32 35

A3 3 6 9 12 15 18 21 24 27 30 33 36

When frequency-hopping load (number of cell frequency/MA length) is less

than 50%, MAIO of 3 cells in the same base station should not be adjacent

frequency. In addition, MAIO of cells in the same direction in each station

and HSN of 3 cells in the same base station should be the same, and HSN in

adjacent base stations should be different. Base station distance with the

same HSN should be as far as possible, all of which should be guaranteed.


A1

A2

A3

A4

A5

A6

B1

B2

B3

B4

B5

B6

A1

A2

A3

A4

A5

A6

B1

B2

B3

B4

B5

B6

A1

A2

A3

A4

A5

A6

B1

B2

B3

B4

B5

B6

A1

A2

A3

A4

A5

A6

B1

B2

B3

B4

B5

B6

A1

A2

A3

A4

A1

A2

A3

A4

A1

A2

A3

A4

A5

A6

B1

B2

B3

B4

B5

B6

A1

A2

A1

A2A6

A1

A2

A3

A4

A5

A6

B1

B2

B3

B4

B5

B6

Figure 1.4-4 2 x 6 multiplexing


18

Obviously, 2 x 6 multiplexing model is not symmetrical model. Therefore,

the multiplexing distance of cells A1 and A4 are different from that of other

cells.

Indicating the value of cellular hexagon side length as 1, from Figure 1-8

and the preceding interference models, carrier-to-interference ratio of cells

A1 and A4 can be concluded as:

dBdBI

C86.16

)64.2(

1log10)(

4

4

Carrier-to-interference ratio of other cells can be concluded as:

dBdBI

C04.12

)2(

1log10)(

4

4

1.4.5 MRP Multiple Frequency Multiplexing MRP

The Multiple Multiplexing Pattern (MRP) technology divides the full band

of frequency into BCCH frequency band and a certain number of TCH

frequency bands, and these frequency bands are mutually orthogonal. In

addition, each band of load frequency is an independent layer. Frequencies

in different layers adopt different multiplexing method and frequency

multiplexing is closer and closer by layers.

This method divides the full band of frequency into two mutually

orthogonal bands, that is, BCCH frequency band and TCH frequency band,

planning with different multiplexing methods respectively. One of methods

to improve system capacity is to use closer multiplexing method. BCCH

channel plays a decisive role in the process of mobile station access and

switching. Therefore, in order to ensure the quality of BCCH channel, the

following benefits can be enjoyed, if using the frequency orthogonal to TCH

frequency band:

BCCH can use 4 x 3 or higher multiplexing coefficient to ensure the

quality of BCCH channel, while TCH uses relatively close

multiplexing method.


19

Separation among each layer of BCCH and TCH frequency band

reduces planning workload. Therefore, planning by layer is available.

In addition to that, a section of frequency may be kept for micro cell.

BSIC decoding has nothing to do with voice channel load. BCCH

frequency band and TCH frequency band are mutually orthogonal.

Therefore, the increase in TCH channel load has little influence on

BCCH channel. In addition, it does not have an impact on BSIC

decoding, and thereby improving switching performance.

Simplify the configuration of adjacent cell list. Some documents

indicate that long adjacent cell list will reduce switching performance,

while this method can simplify adjacent cell list, and thereby

improving switching performance.

BCCH independently uses a segment of frequency (12 frequency

points in 4 x 3 method), and thus length of adjacent cell list

(composed of BCCH frequency points) can be greatly reduced.

Give full play to immunity technologies to interference, such as

power control and DTX. BCCH cannot use dynamic power control

and DTX and it has been transmitting signal in the highest

transmission power. Therefore, BCCH and TCH will influence the

effect of these anti-interference technologies by using the same

frequency band.

Each layer in BCCH and TCH is comparatively independent, which

helps maintenance and expansion by layer. Increasing or deleting sites or

TRX in cells will not have an impact on existed BCCH frequency planning

and thus facilitating network maintenance.

MRP segmenting with 6 MHz frequency band

1 2 3 4 5 6 7 8 9 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

1 2 3 4 5 6 7 8 9 11 12

13 14 15 16 17 18 19 20

21 22 23 24 25 26

27 28 29 30TCH3(4)

BCCH(12)

TCH1(8)

TCH2(6)


20

Carrier No.

MRP is one of the hotspots in frequency planning technology development

in recent years. Some documents indicate that by using MRP simultaneously

integrated with frequency-hopping, DTX, power control and other immunity

technologies to interference can reduce average frequency multiplexing

coefficient to around 7.5, without influencing network quality.

Example:

TRX 2 3 4

20% 30% 50%

MRP 12/8 12/8/6 12/8/6/4

12+8/2=10 (12+8+6/3=8.7 (12+8+6+4)/4=7.5

TRX TRX quantity of the cell

Proportion of the cell

MRP MRP segments

Average frequency multiplexing coefficient

Frequency-hopping diversity gain

Large, medium, small

Comment [m4]:

Comment [m5]:


21

In the above table, the number of cells with 2TRX is 20%, that with 3TRX

is 30%, and that with 4TRX is 50%. Given that these cells are distributed

equally, thus average frequency multiplexing coefficient must be less than

actual multiplexing coefficient. Take the cells with 3TRX as an example.

Since the number of cells with 3TRX or above is actually 80%, and they are

distributed equally, thus the actual multiplexing coefficient on the third layer

is 6/0.8=7.5.

Extended MRP is the development of MRP concept. After being segmented,

each layer can include frequencies of each layer thereafter: Layer TCH0

includes frequency points in each layer from TCH1 to TCHn, layer TCH1

includes frequency points in each layer from TCH2 to TCHn, and so forth.

First, assign frequency points in Layer TCHn, then frequency points in

Layer TCHn-1, and so forth. However, this will have an impact on the

structure of MRP planning.

Extended MRP segmenting with 6 MHz frequency band

1 2 3 4 5 6 7 8 9 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

1 2 3 4 5 6 7 8 9 11 12

13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

21 22 23 24 25 26 27 28 29 30

27 28 29 30TCH3(4)

BCCH(12)

TCH1(8)

TCH2(6)

Carrier No.

Example:

Take frequency bandwidth of 7.2 MHz as an example. Classify 36 pairs of

carrier frequencies into four groups according to 12/9/8/7 by using MRP, as

follows:

Comment [m6]:


22

TCH1 TCH2 TCH3

60 61 62 63 64 65

66 67 68 69 70 71

72 73 74 75 76 77

78 79 80

81 82 83 84 85

86 87 88

89 90 91 92

93 94 95

Channel Type

Channel No.

Logic Channel

TCH1 TCH1 Traffic Channel



Channel BCCH adopts 4 x 3 multiplexing (Figure 1.4-5A), traffic channel

TCH1 adopts 4 x 3 multiplexing (Figure 1.4-5B), traffic channels of TCH2

and TCH3 adopt 2 x 3 multiplexing (Figure 1.4-5A and Figure 1.4-5B),

classify them into four groups.

Comment [m7]:


23

60

64

68

62

66

7063

67

7161

65

69

72

75

78

73

76

7972

75

787477

80

12-carrier frequencies of BCCH adopt 4 x

3 multiplexing method

(A)

9-carrier frequencies of TCH1 adopt 3 x 3

multiplexing method

(B)

Figure 1.4-5

89

91

93

9092

94 9092

9489

91

93

8183

85

8284

8682

84

8183

85

86


multiplexing method

(A)


multiplexing method

(B)

Figure 1.4-6


24

60728189

64 75 8391

85 93 6878

62738290

66 76 8492

70808594

63 7282 90

67 75 92 84

71 8678 94

65 77 8391

61748189

85936980

Figure 1.4-7 Configuration diagram of MRP carrier frequency with a frequency bandwidth of 7.2

MHz

Comparison about system capacities between packet multiplexing and MRP

technology

According to the preceding various analysis and introduction on

multiplexing technologies, now lets make a comparison on capacity

increasing among these four multiplexing methods (4 x 3 multiplexing

method).


25

43 3/2/2 3/3/2 1440 1

33 3/3/3 1788 124

13 4/4/4 2640 183

MRP1296

**

3/3/3 1788 124

6MHZ

26 2/2/2/2/2/2 2160 15

43 4/4/4 2628 1

33 5/5/5 3384 129

13 6/6/6 4272 163

MRP1296

**

6/6/6 4272 163

9.6MHZ

26 3/3/3/4/4/4 4416 168

Note: GOS = 0.02, 0.025 Erl/User

** () herein indicates multiplexing method of each carrier frequency

Multiplexing Method

Based Station Configuration

Average Capacity per Site

Capacity Ratio

1.4.6 Concentric Cell Technology

(1) Basic principle

Comment [m8]:


26

So-called concentric cell is to divide a common cell into two areas: external

layer and internal layer, or called as overlay and underlay. The scale of

overlay covers traditional cellular, while that of underlay covers mainly

around base stations. The differences between overlay and underlay are not

only on coverage scale, but also on frequency multiplexing coefficient.

Overlay generally adopts traditional 4 x 3 multiplexing method, while

underlay adopts closer multiplexing method, such as 3 x 3, 2 x 3 or 1 x 3.

Therefore, all carrier channels are classified into two groups. One is for

overlay and the other is for underlay. The reason why this structure is called

concentric cell is that overlay and underlay share co-location, a set of

antenna system and the same BCCH channel. However, public control

channel must belong to external channel group, which means call

establishment must operate on external channel. Diagram of concentric cell

structure is shown as follows:

f5

f2

f3

f6

f9

f12

f10

f11

f1

f7

f4

f8

/

/

Figure 1.4-8 Diagram of concentric cell structure

/ Overlay

Comment [m9]:


27

/ Underlay

According to different methods to realize concentric cell, they are common

concentric cell and intelligent underlay overlay (IUO). The difference

between these two concentric cells is mainly about underlay transmission

power and handoff algorithm between underlay and overlay.

Generally overlay transmission power of common concentric cell is lower

than overlay power, and thereby reducing coverage scale, increasing

distance ratio and satisfying co-frequency interference requirement. Handoff

between underlay and overlay in common concentric cell is based on power

and distance.

Transmission power in underlay of IUO (frequency adopts closer

multiplexing method, therefore, this layer is usually called as super layer) is

the same as that in overlay (usually called as regular layer), as a result of

handoff algorithm. Handoff algorithm of IUO is switched based on C/I. The

simple description of its realization process is as follows:

A call is established at the regular layer. Then, the BSC continuously

monitors the downlink super group channel C/I ratio of the call. When a

certain super channel C/I ratio reaches the available threshold (good C/I

threshold defined in IUO), the channel for the call is switched to the super

channel. At the same time, the BSC continues to monitor the channel C/I

ratio. When the C/I ratio reaches a bad threshold (bad C/I threshold defined

in IUO), the channel is switched back to the regular channel. Therefore, to

use IUO, the system must have the following functions:

A. Estimation on downlink co-frequency C/I ratio

B. Handoff algorithm concerning IUO

Handoff in cell from regular layer to super layer (measured C/I greater

than good C/I threshold)

Handoff in cell from super layer to regular layer (measured C/I less than

bad C/I threshold)

(2) Capacity


28

Since underlay adopts closer multiplexing method, each cell can be assigned

to more TRX, and thereby improving frequency utilization rate and

increasing network capacity. However, it should pay attention to the fact that

coverage semi diameter of underlay in concentric cell is less than common

cell and its traffic absorption is confined by call operation distribution and

coverage scale. The following table shows distribution on different call

operations. Under different coverage scales, make a comparison on capacity

between concentric cell and traditional 4 x 3 method. Indicate Si as underlay

coverage, Sout as outlay coverage area, the measure of capacity as Erlang:

Si / Sout

3TRX 2TRXout+2TRXin 4TRX 3TRXout+2TRXin

0.3 14.04 10.57 21.04 20.05

0.7 14.04 20.55 21.04 28.25

0.9 14.04 21.04 21.04 28.25

0.3 14.04 15.09 21.04 21.92

0.7 14.04 21.04 21.04 28.25

Coverage Ratio

Uniform distribution of traffic

Linear distribution of traffic

What needs to explain is that coverage ratio is concerned with frequency

multiplexing type. The closer frequency multiplexing type is, the more

intense co-frequency interference is and the less underlay coverage ratio is.

In addition to that it is concerned with configuration of handoff parameter

and surrounding environment. Therefore, coverage semi diameter is not

configured at will. It needs giving comprehensive consideration upon the

quality of network, which is hardly more than 50%.

Comment [m10]:


29

From the preceding analysis, the concentric cell technology improves the

capacity a little or even reduces the capacity when traffic is well distributed.

The denser the traffic around a cell, the more obvious the effect is. Above all,

capability increasing is limited. For a common concentric cell, the Tx power

of its internal layer is low, which is hard to absorb the traffic indoor.

Therefore, the frequency efficiency is low and the actual capacity is

increased by about 10% to 30%. For IUO, the Tx power of its internal layer

remains unchanged, which can absorb the traffic indoor and the handoff

based quality for capacity absorbing is flexible. Therefore, the actual

capacity is increased by 20% to 40% (relatively high).

(3) Features and application

A. Common concentric cell

The features of common concentric cell are as follows:

No need to change network structure.

Need to increase some special handoff algorithm, but generally the

realization is simple.

No specific requirements on mobile phone.

A limitation on capacity increase, generally it is 10-30%. It is concerned

with call operation distribution. Because of small power, underlay is hard to

absorb indoor traffic.

It is applicable in the situation that call traffic is highly concentrated around

base station and distributed outdoor.

Notice in application

Make a good network planning. On one hand, it should be applied in areas

of high call operation concentration, on the other hand, making a good

planning about coverage area of underlay. The area cannot impact quality

because of interference caused by close multiplexing, and it should absorb

enough call operation. If it is a bad planning, it will not only hardly increase

capacity, but also reduce network quality.

It would be better to integrated with technologies about reducing

interference, such as power control and DTX.


30

B. IUO

IUO mainly has the following features:

As a kind of concentric cell, IUO can utilize existing site, have little

modification on network and no specific requirements on mobile phone.

System function needs to increase measure and estimation on C/I and

special handoff algorithm.

Capacity has an augment of 20% - 40%, and it has nothing to do with call

operation distribution and call traffic absorbed by super layer. In addition to

that it can ensure quality on the basis o f increasing capacity.

Super layer can adopt closer multiplexing method. When the frequency is

enough wide, it can keep a segment of frequency for micro cellular.

It is applicable for the areas where high density of call operation is and

concentrated around base stations.

Notice in IUO application:

Make a good planning. Cell should be planned based on call operation

distribution and notice to reduce interference.

When cell channel is being assigned, reasonable configuration between

super layer frequency and regular layer frequency should be noticed.

In order to reduce interference, power control and DTX technology should

be integrated in application.

It would be better to adopt handoff based on C/I in regular layer.

1.5 Cell Splitting

In the beginning period of GSM network establishment, since there are not

too many users, channels are surplus. With ever-increasing users, blockage

occurs in the channels that were assigned to each base station. At this time

new channels can be added and assigned to the original base stations. If

users are constantly increasing, while the usable channels are exhausted,

then only cell splitting, increase in base stations and co-channel


31

multiplexing can meet demand of users. Usually, the semi diameter of

newly-splitted cell is only half of that of original cell.

Semi diameter of new cell=semi diameter of old cell/2 (1-5-1)

based on formula (1-4-1), the following formula works:

Coverage area of new cell=coverage area of old cell/4 (1-5-2)

Given that the highest traffic load of each new cell and old cell are the same,

then principally speaking, it is concluded as:

New traffic/unit area= 4 old traffic/unit area (1-5-3)

Therefore, the relation between cell splitting and increase on user capacity

can be represented as follows:

Tn = 4n T (1-5-4)

In the formula: Tn- network capacity after n times cell splitting

T0- network capacity prior to cell splitting

Formula (1-5-4) is applicable for cellular grid that is splitted to 4 less cells

at a ratio of 1:4. Simply speaking, after one time splitting, the number of

users can be increased to 4 times of the original one and the actual capacity

is less than that of its four times.

1.6 Several Common Immunity Technology to Interference

GSM system itself has many immunity technologies to interference, such as

frequency hopping, power control, discontinuous transmission based on

voice activity detection and so on. If effectively applicable, it will improve

C/I, thereby it can form a closer frequency multiplexing method, in addition

to increase on frequency multiplexing coefficient and frequency utilization

rate. Herein, we will introduce these technologies one by one and analyze

their gain through absolute mathematics model and artificial model.


32

1.6.1 Discontinuous Transmission (DTX)

TRAU BTS

BTS MS

480 ms

Comfort noise frame

Voice frame

During the period of voice activation, discontinuous transmission encodes

voice at 13 kbit/s. During silence period, it encodes comfort noise at 500

bit/s.

During silence period, discontinuous transmission has little contribution to

interference. It can be regarded that its power is zero (none activation). If

DTX activity factor is p

, then its gain is

pI

C

pI

CdBIC log10log10log10)(/

1.6.2 Frequency Hopping (FH)

Frequency hopping is one kind of spread spectrum communication. It is

applied in cell mobile communication system to improve system

anti-multipath fading capability. In addition to that it can curb co-frequency

interference on communication quality. Therefore, it is highly applicable.

Especially, nowadays when spectrum resource is more and more insufficient,

Comment [m11]:


33

frequency hopping becomes one of the most effective methods to improve

spectrum utilization rate.

In GSM, data of each logic frame is sent in a way of decentralized and

interleaved in 8 TDMA frames, while all of these data has been encoded in

convolution. If a part of these 8 blocks of burst is interfered or damaged, it

can recover the data that has been sent through convolution encoder.

However, if too many blocks are damaged, it is hard to recover the original

data. By frequency hopping, it is unlikely to make burst in one channel in

heavy fading area too long (it easily occurs in a still or moving-at-low-speed

mobile station that works on a fixed carrier), or to be interfered by a certain

strong co-frequency signal. Thus it is possible to get good transmission

effect by using channel coding and encoding. It is the simple principle to

improve communication quality by adopting frequency hopping technology.

Frequency hopping sequence used by GSM system is a kind of Poisson false

random variable sequence. It can provide 64 frequency hopping sequence at

most. Length is same to hyperframe (lasting 3 hours 28 min 53 sec 760ms)

to ensure, as soon as possible, that each sequence is orthogonal to each other,

so that ensure the effect of frequency hopping. Frequency hopping sequence

in GSM is mainly described by two parameters: HSN (Hopping-frequency

Sequence Number) and MAIO (Mobile Assignment Index Offset). Usually,

different HSN is assigned to different cell and different value of MAIO is

assigned to different channel of cells.

It is noticed that every channel in a same cell adopts a same HSN and only

value of MAIO offset is different, which ensures that every channel in a

same cell will not occupy same frequency points at a same time. In different

cells, as a result of different HSN, it adopts different categories of frequency

hopping sequence. Then it makes frequency hopping sequence in every cell

is not relevant as far as possible, so that interfering resource signals are

assigned to many channels to ensure coding and encoding effect. When

HSN=0, MAI is dup loop from low to high, it is called as Cyclic Hopping.

Since frequency hopping gain in this method is very low, usually it is not

adopted in GSM


34

GSM supports baseband frequency hopping and RF frequency hopping (or

called as Synthesized Frequency Hopping). Baseband frequency hopping

means that many transmitters work on their own fixed frequency points,

while on baseband signals from different channels switched to different

transmitters are sent according to frequency sequence. Baseband frequency

hopping can be easily realized, however, frequency hopping points are few

as a result of the limited TRX number. Frequency hopping artificial system

established by ZTE is mainly to support RF frequency hopping. Baseband

frequency hopping is only regarded as an exception of RF frequency

hopping (that is, the number of frequency points equals to the number of

TRXs). The advantage brought by frequency hopping is mainly about the

effect of Frequency Diversity and Interference Diversity. Frequency

diversity actually improves network coverage scale, and Interference

diversity increases network capacity

The number of available frequency hopping in baseband frequency hopping

is equal to the number of TRX. Therefore, it can only bring frequency

diversity gain, not interference diversity gain. However, now GSM

operators are more concerned about capacity problem. Coverage is not a

problem in most of cities. RF frequency hopping is a very effective method

to solve capacity problem.

RF frequency hopping is a trend of application in network planning.

Frequency diversity gain

Frequency diversity means its immunity ability to Rayleigh fading. Since

Rayleigh fading on different carriers is certain irrelevance (the more

frequency differential is, the less irrelevance is), then burst distributed on

different carriers will not influenced by the same Rayleigh fading. It means

a lot to the still and moving-at-low-speed mobile station. It is said that it can

provide a gain value of 6.5dB. However, MS moving at high speed and two

successive burst in a same channel are different on timing position, which

means they are irrelevant to Rayleigh change. They are seldom influenced


35

by the fading at a time when frequency diversity provided by low-speed

frequency hopping is very little.

Under the condition that MS moves at high speed, frequency points

assigned in cells have little impact on frequency hopping performance.

While under the condition of no frequency hopping, there are about

frequency diversity gain of 1 dB to 2dB. When MS moves at low speed

(TU3), because of frequency diversity effect, the number of assignment

frequency points has significant influence on system performance.

Frequency points increased by a time will obtain about gain value of

0.2~1dB, load rate can be increased by 10% or so.

(1) Interference diversity gain

Interference diversity means that it curbs capability of interfering signals in

other co-frequency multiplexing cell, that is, to provide frequency hopping

and interfering differential on the transmission path in order to improve

interference under the harshest conditions. It makes all users evenly enjoy

good communication quality, which is very important for the mobile

communication system with lots of users, especially critical to increase

communication traffic through augmenting frequency multiplexing rate.

Usually interference diversity effect should be provided, and the number of

frequency hopping points should not be less than 3.

MA},...,,,{ 321 nffff ,

TRX m (mn)

Regarding the above figure, given that mobile station is in air with fk at time

t. At that time, the possibility of interfering cell fk being activated is


36

nmCCp mnmn //

11

m

n

I

C

pI


Gain

(3) Frequency hopping planning and capacity analysis

If co-frequency point is 10 MHz, frequency hopping planning and capacity

analysis without adopting frequency hopping are as follows:

The multiplexing method of BCCH is 4X3, and the multiplexing method of

traffic channel is 3X3. 10 MHz has 50 frequency points. It leaves 37

frequency points after losing 1 protection frequency point and 12 BCCH

frequency points. Thus, each cell can be assigned 4 traffic frequency points

((37-1)/9), and only one frequency point is left. Then its most assignment

should be 5+5+5. Each cell can provide 37 channels

(1BCCH+2SDCCH+37TCH).

When RF frequency hopping technology is adopted, traffic channel can

adopt 1X3 multiplexing. When load is 50%, each cell can provide 6 service

logic frequency points. The reason why it is called logic frequency point is

that they all adopt the same 12 frequency hopping collection ((37-1)/3).

Only HSN is different from MAIO, then one frequency point is left, and the

most assignment becomes 7+7+7. It can provide 53 service traffic channels

(1BCCH+2SDCCH+53TCH) with increase on capacity by 43%. At this

time more than 90% of areas can have C/I with a value of 9dB. When DTX

and ZTE distinctive fast power control algorithm are adopted at the same

time, system capacity can be improved much better. If intelligence traffic

control technology is adopted, GSM can acquire soft capacity, and gain

more system capacity by sacrificing certain voice quality in hot traffic areas.

Comment [m12]:


37

1.6.3 Dynamic Power Control (DPC)

A3

A1

A2

A3

A1

A2

A3

A1

A2

A3

A1

A2A3

A1

A2

A3

A1

A2

A3

A1

A2

Seen from the above figure, in dynamic power control interfering only when

mobile station is at the border of a cell, BTS can work with the most

transmission power.

Obviously, the position of interfering mobile station is a possibility. The

circumstance is even more obvious in frequency hopping.

Indicate DPC factor as p:

pI

C

pI


Gain

1.6.4 1 x 3 Multiplexing + Radio Frequency FH + DTX + DPC

Lets observe and study 13multiplexing interference in specific and

check the contribution made by immunity technologies to interference to

reduce interference and increase system capacity.

Differentials on interfering degradation between 1 x 3 and 4 x 3

multiplexing methods:

CIR 43- CIR 13 =18 - 9.43 8.57 dB

Comment [m13]:


38

Interference diversity gain made by 1 x 3 frequency hopping with 50% of

frequency carrier:

10log10(2/1) = 3dB

Indicate the length of frequency hopping is 12 frequency points, then

frequency diversity gain is about 2 dB.

Indicate DTX activation factor is 0.5, then its gain is:

-10log10(0.5) = 3dB

Indicate DPC factor is 0.9, then its gain is:

-10log10(0.9) =0.5dB

The total gain is:

3+2+3+0.5=8.5Db

From the above analysis we can see that utilizing immunity technologies to

interference basically can recover interference degradation made by

intensified multiplexing methods.

1.7 Conclusion on Principles of Frequency Assignment in Engineering

Adjacent frequencies cannot be identical in the same base station.

Directly adjacent base station should avoid co-frequency (even

though the direction of their antenna central lobes are different, side

lobe and back lobe can also bring much interference.):

End-on cells cannot be co-frequency and should avoid adjacent

frequency, especially for BCCH and SDCCH carrier frequency

(usually they are 1st and 2nd carrier frequencies of the cell). When

frequency hopping is adopted, the starting hopping points of adjacent

base stations can be the same, while the algorithm of frequency

hopping cannot be the same.

When in common frequency hopping (that is synthesizer frequency

hopping), frequency hopping algorithm (HSN) of each cell in the

same station are all identical. However, starting frequency hopping


39

points (MAIO) cannot be adjacent frequency. Note: whatever adopts

CCB combiner cannot support frequency hopping.

Design of BSIC should also be noticed. BSIC = 8 x NCC + BCC,

BCC is available from 0-7, thus the near co-frequency and adjacent

frequency cell should not be the same as far as possible.

Co-frequency (especially BCCH carrier frequency) and the same

BSIC in short distance should be avoided as far as possible.

There are high mountains between base stations, which is not

regarded as adjacent station. If there are large scale of water between

base stations, which should be regarded as adjacent station.

Prior to frequency hopping, and no limitation on the using scale of

BCCH carrier, BCCH can be staggered as far as possible. A certain

segment of frequency band should be saved for 4X3 multiplexing

when in frequency hopping. If frequency is enough, BCCH can adopt

5 x 3 or even 6 x 3 multiplexing models to reduce interference among

BCCH.

In large or medium scale cities, different close frequency

multiplexing methods are adopted according to different functions

supported by equipments, such as MRP, 1 x 3, and 1 x 1 frequency

hopping. Meanwhile saving part of frequency points in advance for

micro cellular to establish layered network. Its frequency

multiplexing coefficient is small.

In large or medium scale cities, different frequency multiplexing

methods are adopted according to different functions supported by

equipments. Whether it is needed to establish layered network or not,

it is based actual conditions. Its frequency multiplexing coefficient is

a little bit more than that of large or medium scale cities.

In towns and villages, frequency resource are abundant, then regular

4 x 3 frequency multiplexing method can be adopted.

For stations build on high mountains for the geographic reason,

independent frequency points can be assigned.


40

The above are about some principles in frequency planning, while doing

frequency planning needs another important principle, that is, it should be in

accordance with the local actual circumstance. Morphology and base station

of each system are different, so are transmission of radio signal, which

requires us to learn more about the local actual circumstance before we

make frequency planning. We should not be confined in the common

frequency multiplexing methods, for frequency assignment methods is in

accordance with the local actual circumstance. Frequency planning should

satisfy the local situation as far as possible. If possible, we may use some

specific planning tools integrated with e-map to do field-strength prediction.

First, observe that whether the coverage area of each cell is reasonable.

Then make coverage modification or frequency planning revise for those

areas which are not satisfied with the interference requirements (generally

we set the co-frequency interference as 12dB or so and leave 3Db margin

while in prediction). After base station is in operation, it is to judge whether

coverage frequency planning is proper or not by line test and some statistic

data. For the areas with heavy interference, it is to solve the problems by

modifying coverage, revising frequency planning and other methods.

We make frequency planning in a way of geographic slicing, while we

should keep a certain number of frequency points (frequency is enough) or

make frequency band division at slicing border. The choice of border should

avoid hot areas or networking complex areas as far as possible.

Usually, planning should start from the densest area in base station. For

example, first from urban bustling areas, then to suburban base station with

less carrier assignment (usually choose O1/ or S1/1/1 as border). Notice that

there are rivers or l4arger lakes in the urban areas. Interference made by

strong transmission of water surface should be avoided.

As a result of irregularity of actual base station distribution, it is hard to

ensure that the frequency in the same layer carrier can be planned fully

following 4 x 3 or 3 x 3 and other common models. It needs flexible

modification according to actual circumstance.

41

Chapter 2 Neighboring Cell Planning

Knowledge point

Know about the principles and methods in neighboring cell planning.

Analyze cases of improper neighboring cell planning.

Neighboring cell planning determines consecutive coverage of a GSM

network and network performance indexes. The principles of neighboring

cell planning are as follows:

(1) Frequencies of the primary cell and neighboring cells must be

different.

(2) The number of neighboring cells must be less than or equal to 32, and

the OMCR can be configured for a maximum of 32 neighboring cells.

To acquire a better neighboring cell planning, the QoS and load of the

system must be taken into consideration. Actually, more neighboring cells,

more system load will bring because of handover. However, moderate

number of neighboring cells reduces call drops because of handover.

When you plan neighboring cells, take the following aspects into

consideration.

If there are a large number of neighboring cells, handover occurs frequently

and thus leads to signaling overload.

If there are fewer neighboring cells, call drop may occur because of

handover failure, or QoS is affected severely.

Generally, in planning neighboring cells, it is determined that cells are

distributed based on cellular structure. Therefore, pay attention to the

following points:

When you configure neighboring cells for microcells in urban areas,

two-layer adjacency is required, as shown below:


42

A3

D2 B1

D1

D3

C1 B3

C2

B2

C3

A1

A2

A3

D2 B1

D1

D3

C1 B3

C2

B2

C3

A1

A2

A3

D2 B1

D1

D3

C1 B3

C2

B2

C3

A1

A2 A3

D2 B1

D1

D3

C1 B3

C2

B2

C3

A1

A2

Figure 2-1

When you configure neighboring cells in remote areas such as suburban

areas or the countries, one-layer adjacency is required. This is because in

these areas, network is sparsely distributed and each cell is with wide

coverage. In this case, a long distance is between the first layer and the

second layer. For details, see the following figure:

Error! Use the Home tab to apply 1 to the text that you want to appear here. Error! Use the Home tab to apply 1 to the text that you want to appear here.

43

A3

D2 B1

D1

D3

C1 B3

C2

B2

C3

A1

A2

A3

D2 B1

D1

D3

C1 B3

C2

B2

C3

A1

A2

A3

D2 B1

D1

D3

C1 B3

C2

B2

C3

A1

A2 A3

D2 B1

D1

D3

C1 B3

C2

B2

C3

A1

A2

Figure 2-2

In a dual frequency network, cooperation and setting principles between the

two-layer network should be considered in configuring neighboring cells.

Therefore, the network adjacency should be configured according to

different principles of network sharing.

Generally, it is considered that cells are arranged in order based on cellular

shape. But actually, cells are hardly arranged in order because site selection

is affected by various factors. In this case, configuration should be

performed based on data that is obtained by simulating networking planning.

In addition, if the transmit power of a BTS is very large, the covered edge

zone takes a great proportion. In this case, the adjacency cannot be obtained

based on the geographic position; instead, it must be obtained by on-site

measurement, or configure more adjacencies, as shown below:


44

A

B

C

Figure 2-3 Neighboring cell configuration

From the figure, we can see that a cell phone walks along the curve line in

the covered edge zone. Theoretically, the cell phone selects service area A

first, then service area B, and last service area C. But actually, signals of

BTS B cannot size the control position in the curve line because of certain

complex factors in the radio propagation environment. In this case, if cells

of BTS C are not configured as the neighboring cells of sector 1 of BTS A,

the cell phone is always in sector 1 of BTS A, until call drop occurs or the

cell is selected again. To solve this problem, configure sectors 1 and 2 of

BTS A and sector 1 of BTS C as neighboring cells (sense frequency points).

However, you cannot configure all cells as neighboring cells. If all cells are

configured as neighboring cells, unexpected problems may occur, such as

cell reselection and handover.

The following lists improper neighboring cell planning.

One-way neighboring cell

Many neighboring cells

Few neighboring cells


45

The following lists problems that may occur because of improper

neighboring cell planning.

Call drop

Handover failure

Frequent handover

Isolated cells

Abnormal inter-cell handover

Unbalanced traffic

Decreased handover precision

Cases

Case 1

1. Fault Symptom

A BTS in the urban area is configured as S333, and the single frequency

GSM900 network is used at local with 1*3 RF frequency hopping mode.

The cutin failure rate in a sector of this BTS is constantly high. That is, cutin

failure rate for this cell from source cell A is about 80%, but indexes such as

call drop rate and failure rate of the voice channel allocation are normal.

2. Fault Analysis

The fault is not caused by hardware fault and interference. This is because

though the cutin failure rate is high, TCH allocation does not fail, which

indicates that MS can occupy TCH channels allocated by BSC. In addition,

severe interference does not exist because no call drop occurs on MS and

voice communication is with good quality. After analysis, it can be

determined that source cell A is far away from this cell with high cutin

failure rate and thus handover requests should be fewer. Therefore, the fault

may be caused by island effect.

3. Fault Locating


46

The fault is caused by island effect after checking hardware, transmission

stability, and interference and no fault occurs. Check neighboring cells of

source cell A to check whether cells with the same frequency and same color

codes as cell C exist. It is found that such a cell exists. Locate the fault

further. It is found that a very big square is set up between cell B and cell A,

which makes radio propagation conditions between cell A and cell B better.

MS senses signals, and these signals are ones of cell B, but BSC determines

send Handover Command to cell C. At the same time, the level of cell C

may be very low, which makes handover failure. This is handover failure

caused by island effect.

4. Troubleshooting

Modify the frequency point of cell C and add isolation cell B into the

neighboring cell table of cell A. Then, the fault is rectified.

5. Conclusion

In troubleshooting network faults, pay attention to environment change. For

example, whether radio signal propagation is affected, or whether radio

signals can be propagated better. If these factors are changed, engineering

parameters (such as antenna height, downtilt angle, and directional angle)

and cell parameters need to be adjusted (for example, add, delete, or modify

neighboring cells or frequency). Frequency resources of a GSM network are

limited. Therefore, with expansion of the network scale, island effect is

more likely to be generated, especially on handover. In addition, if

co-channel interference is severe, handover success ratio is severely

affected.

Case 2

1. Fault Symptom

A user at border areas of a province complains that he/she cannot disengage

the roaming signals from another province once receiving these signals, but

the roaming problem does not exist at home. These two provinces are not

neighboring cells.

2. Fault Analysis


47

Perform the drive test at local. It is found the following network results:

Figure 2-4

The user usually is located in point P, which is in cell A. Cell A and cell B

are neighboring cells and are home networks for the user. Cell C and cell D

are neighboring cells and are networks for roaming area. In addition, cell A

is not the neighboring cell for cell C and cell D.

Cell D has the same BCCH frequency point as neighboring cell B that is

defined by cell A; therefore, the mobile phone at point P may re-select cell

D and then re-select cell C through cell D. The neighboring cell table of cell

C and cell D does not define the frequency point of cell A. Therefore, the

user resides in the networking of the roaming area. If the user powers off the

mobile phone and then powers it on in cell C, the mobile phone searches

cell C and frequency points of the neighboring cells that are defined by cell

C. This is because the mobile phone keeps the frequency point of the cell

where it is located when being powered off. This leads to the roaming

problem.

3. Troubleshooting

Comment [m14]:


48

Define neighboring cells between two provinces. If neighboring cells cannot

be configured, modify the frequency point of cell B to rectify the fault.

49

Chapter 3 BSIC Planning

Knowledge point

Know about the definition, value range, and planning principles of BSIC.

1. Definition

In a GSM system, each BTS is allocated with a local color code, which is

called base station identity code (BSIC). In a physical position, if the mobile

phone receives BCCH carrier frequencies of two cells concurrently, and

these two cells are with the same channel ID, the mobile phone

differentiates these two cells by BSIC. In network planning, to decrease

co-channel interference, BCCH carrier frequencies of neighboring cells are

assured with different frequencies. However, in the cellular

telecommunication system, it is possible that BCCH carrier frequencies are

multiplex. For these cells with the same BCCH carrier frequency, make sure

that they have different BSICs, as shown below:

Figure 3-1 BSIC selection

In the figure, BCCH carrier frequencies of cells A, B, C, D, E, and F are

with the same absolute channel ID, and other cells use different channel ID

as the BCCH carrier frequency. Generally, cells A, B, C, D, E, and F are

requested to have different BSICs. When the BSIC resources are insufficient,

consider whether neighboring cells of these cells adopt different BSICs.

Here, take cell E as an example, if BSIC IDs are insufficient, preferentially

consider that cells D and E, B and E, and F and E use different BSICs

respectively, but cells A and E, and C and E use the same BSIC respectively.


50

The BSIC consists of the network color code (NCC) and base transceiver

color code (BCC), as shown in figure 3-2. The BSIC is transmitted over the

synchronization channel (SCH) of each cell. It mainly provides the

following functions:

Figure 3-2 BSIC composition

When the mobile phone receives codes from SCH, it is determined

that the mobile phone is synchronized with the cell. To correctly

translate the information of the downstream common signaling

channel, the mobile phone needs to know training sequence code

(TSC) used by the common signaling channel. According to GSM

specifications, TSC has eight fixed formats, which are presented from

ID 0 to ID 7. The TSC SN used by the common signaling channel of

each cell is determined by BCC of the cell. Therefore, one of

functions of BSIC is to notify the mobile phone of the TSC that is

used by the common signaling channel of the cell.

BSIC takes part in random access channel (RACH) translation, so it

can be used to prevent RACH that is transmitted to neighboring cells

by the mobile phone from being incorrectly translated as the access

channel of the cell by the BTS.

When the mobile phone is in use, it must measure the levels of BCCH

carrier frequencies of neighboring cells and then report the results to

the BTS according to specifications of the relevant neighboring cell

table on BCCH. At same time, for each frequency point in the

measurement report, the mobile phone must point out BSIC of the

measured carrier frequency. In specified conditions, that is, two or

BCC

BSIC

NCC


51

more cells contained in neighboring cells of a cell adopt the same

BCCH carrier frequency; the BTS differentiates these cells based on

BSIC. This prevents incorrect handover or even handover failure.

When the mobile phone is in use, it must measure signals of

neighboring cells and report the results to the network. In each report,

contents of only six neighboring cells are included. Therefore, take

control for that the mobile phone reports situations of only cells

having handover relationships with the current cell. In BSIC, the

higher three bits (representing NCC), are used for control. In this case,

carrier frequencies can control the mobile phone to report situations

of only neighboring cells in the allowed range of NCC through the

broadcast parameter "allowed NCC".

2. Format

Format of BSIC: NCC-BCC.

Value range of NCC: 0-7.

Value range of BCC: 0-7.

3. Setting and Effect

In many cases, different GSMPLMNs adopt the same frequency, but their

network planning is independent. To ensure that neighboring BTSs that are

with the same frequency point have different BSICs, neighboring

GSMPLMNs adopt different NCCs.

However, situations in China are special. Actually, the GSM network of

China Telecom is an integer and independence GSM network. Though

China Telecom subordinates a majority of local mobile offices, these offices

belong to the same carrier frequency, China Telecom. However, China is a

vast land and it is hard to manage the GSM network in a united manner.

Therefore, the entire GSM network is divided and managed by local mobile

offices (or their equal agencies) of cities and provinces. These local mobile

offices plan their networks independently. To ensure that BTSs with the

same BCCH frequency that are used by border areas of cities and provinces

have different BSICs, NCCs of cities and provinces are coordinated by

China Telecom in a united manner.


52

BCC is a component part of BSIC, which is used to identify different BTSs

that are with the same BCCH carrier frequency in the same GSMPLMN.

And the BCC value should meet the requirements described above as much

as possible. In addition, according to GSM specifications, the TSC of the

BCCH carrier frequency in a cell need to be the same as the BCC of the cell.

Usually, vendors assure this consistency.

4. Precautions

Make sure that neighboring cells that are with the same BCCH carrier

frequency have different BSICs. Especially, when two or more cells

contained in neighboring cells of a cell adopt the same BCCH carrier

frequency, these two cells must have different BSICs. Pay attention to

configurations of cells in the border areas of cities and provinces. Otherwise,

inter-cell handover failure may occur.

GSM Frequency Planning

Documents

frequency hopping fh

neighboring cell planning

concentric cell technology

common immunity technology

interference analysis

ideal cellular structure

interference model analysis

contents chapter