gsm planning

Upload: jamal-ahmad

Post on 07-Apr-2018

221 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/6/2019 GSM Planning

    1/10

    Frequency Planning

    Frequency Planning

    Abstract

    This is a technical document detailing a typical approach to Frequency Planning Process.

    Page 1 of 10

  • 8/6/2019 GSM Planning

    2/10

    Frequency Planning

    CONTENTS

    Frequency Planning

    (1.0) Introduction Page 3

    (2.0) Frequency Re-use Page 4

    (3.0) Co-channel Interference and System Capacity Page 5

    (4.0) Design Criterion Page 6

    (4.1) Example Page 7

    (5.0) Frequency Channel Allocation Page 7

    (5.1) Example Page 7

    (6.0) BSIC Planning Page 8

    (6.1) Example Page 8

    (7.0) Automatic Frequency Planning Page 9

    (8.0) Frequency Hopping Page 9

    (8.1) Frequency Hopping Techniques Page 10

    Page 2 of 10

  • 8/6/2019 GSM Planning

    3/10

    Frequency Planning

    Frequency Planning

    (1.0) Introduction:

    The Cellular concept is a system with many low power transmitters, each providing coverageto only a small portion of the service area. Each base station is allocated a portion of the totalnumber of channels available to the entire system, and nearby base station are assigneddifferent group of channels so that the interference between base stations is minimised. Thechannels assignment in case of GSM900, E-GSM900 and DCS1800 (or GSM1800) is asshown in Figure-(1.1) below,

    Fig.- (1.1) Channels Assignment

    As shown the Uplink and Downlink band are separated by 20 MHz of guard band in case ofGSM and DCS and 10 MHz in case of E-GSM. The channel separation between Uplink andDownlink is 45 MHz in case of GSM and E-GSM and is 95MHz in case of DCS network. Eachchannel(carrier) in GSM system is of 200 KHz bandwidth, which are designated by AbsoluteRadio Frequency Channel Number (ARFCN). If we call Fl(n) the frequency value of the carrier

    ARFCN n in the lower band(Uplink), and Fu(n) the corresponding frequency value in theupper band (Downlink), we have:

    GSM 900 Fl(n) = 890 + 0.2*n 1 n 124

    Fu(n)

    Fu(n) = Fl(n) + 45

    E-GSM 900 Fl(n) = 890 + 0.2*n

    Fl(n) = 890 + 0.2*(n-1024)0 n 124

    Fu(n)

    975 n 1023

    Fu(n) = Fl(n) + 45

    DCS 1800 Fl(n) = 1710.2 + 0.2*(n-512) 512 n 885 Fu(n) = Fl(n) + 95

    Table (1.1) ARFCN

    Page 3 of 10

    880 890 915 935 960925

    E-GSM900UPLINK

    GSM900UPLINK

    GSM900DOWNLINK

    E-GSM900DOWNLINK

    45 MHz

    45 MHz

    1710 1785

    DCS1800UPLINK

    1805 1880

    DCS1800DOWNLINK

    Guard Band

    95 MHz

  • 8/6/2019 GSM Planning

    4/10

    Frequency Planning

    Hence we have 124 channels in GSM900, 174 channels in E-GSM900 and 374 channels inDCS1800.

    (2.0) Frequency Re-use:

    One important characteristic of GSM networks is frequency planning wherein given the limitedfrequency spectrum available, the re-use of frequencies in different cells is to be plannedsuch that high capacity can be achieved keeping the interference under a specific level.

    A cell in a GSM system may be omni-directional or sectored represented by hexagons. InGSM system a tri-sectored cell is assumed and the frequency plan is made accordingly. Tounderstand the frequency re-use planning, consider a GSM system having S channels(ARFCNs) allocated, wherein each cell (sector) is allocated k channels, assuming that allthree sectors have same number of k channels. If the S channels are divided among N basestations each having three sectored cell, then the total number of available radio channelscan be expressed as,

    S = 3kN

    This explains N base stations each having three sectors and each sector having k channels.

    The N base stations, which collectively use the complete set of available frequencies, inwhich each frequency is used exactly once is called a Cluster. If the cluster is replicated Mtimes then the total number of channels, C, can be used as measure of capacity and is givenby,

    C = M3kN = MS

    The Cluster size N is typically equal to 3, 4, 7, or 12. Deciding a cluster size posses acompromise between capacity, spectrum allocated and interference. A cluster size of 7 or 12gives least interference frequency plan but as the cluster size is big enough hence re-use atfar away distance hence lesser capacity and would also require bigger frequency spectrum.Consider an example where k equals 1 that is one frequency per sector. With a cluster size of7 would require minimum spectrum of,

    S = 3 x 1 x 7 = 21 ARFCNor21 x 0.2 MHz = 4.2 MHz of spectrum that is about 16% of total available spectrum in

    GSM900.

    Adding one more frequency per sector would take the requirement to 42 ARFCN or 33% oftotal spectrum. On the other hand a cluster size of 3 would require (k = 1),

    S = 3 x 1 x 3 = 9 ARFCNor9 x 0.2 MHz = 1.8 MHZ which is about 7% of total spectrum available.

    Addition of one more frequency still results in about only 14% of spectrum required. But herea big compromise is made on interference, as the cells are quite closely located hence re-usewould pose a major problem. Studies have revealed that cluster size of 4 gives the bestbalance between capacity & interference, with k equal to 2 meaning two frequencies persector gives,

    S = 3 x 2 x 4 = 24

    Page 4 of 10

  • 8/6/2019 GSM Planning

    5/10

    Frequency Planning

    or24 x 0.2 MHz = 4.8 MHz that is about 19% of total spectrum available.

    Figure 1.2 illustrates the frequency reuse for cluster size of 4, where cells labelled with the

    same letter use the same group of channels.

    Fig.- (1.2) 4 x 3 Re-use pattern

    (3.0) Co-channel Interference and System capacity:

    Frequency re-use implies that in a given coverage area there are several cells that uses thesame set of frequencies. These cells are called co-channel cells and the interference betweensignals from these cells is called co-channel interference. Unlike thermal noise which, can beovercome by increasing the S/N ratio, co-channel interference cannot be combated by simpleincrease in carrier power. This is because an increase in carrier power increases theinterference to neighbouring co-channel cells. To reduce co-channel interference, co-channelcells must be physically separated by a minimum distance in order to provide sufficientisolation due to propagation.

    In a cellular system where the size of each cell is approximately the same, co-channelinterference is independent of the transmitted power and becomes the function of the radiusof the cell (R), and the distance to the centre of the nearest co-channel cell (D). Figure- (1.3)explains the relation between the cell radius R, cluster size N and the re-use distance D,

    Here,Outer Cell radius: RInner Cell radius: r = 0.5 x (3)1/2 x R.Re-use distance: D = R x (3 x (i2 + j2 + ij))1/2

    D/R = (3 x N)1/2

    Page 5 of 10

    B1

    B2

    B3

    C1

    C2

    C3

    B1

    B2

    B3

    D1

    D2

    D3

    B1

    B2

    B3

    B1

    B2

    B3

    B1

    B2

    B3

    B1

    B2

    B3

    C1

    C2

    C3

    A1

    A2

    A3

    A1

    A2

    A3

    A1

    A2

    A3

    A1

    A2

    A3

    A1

    A2

    A3

    C1

    C2

    C3

    C1

    C2

    C3

    C1

    C2

    C3

    C1

    C2

    C3

    D1

    D2

    D3

    D1

    D2

    D3

    D1

    D2

    D3

    D1

    D2

    D3

    D1

    D2

    D3

  • 8/6/2019 GSM Planning

    6/10

    Frequency Planning

    The Cluster size, N = (i2 + j2 + ij)

    Fig.- (1.3) Re-use distance calculation.

    Where i and j are non-negative numbers. To find the nearest co-channel neighbour of aparticular cell, one must do the following: (1) move i cells along any chain of hexagons andthen (2) turn 60 degrees counter-clockwise and move j cells. This is illustrated in the figureabove fori = 1 &j = 2 for a cluster size of 7.

    By increasing the ratio of D/R, the spatial separation between co-channel cells relative to thecoverage distance of a cell is increased. Thus interference is reduced due to improvedisolation from the co-channel cells. The relation between the re-use distance ratio D/R andthe co-channel interference ratio C/I is as below,

    (D/R)

    = 6 (C/I)

    (Note: C/I is in dB and should be converted to numeric values for calculation)

    Here, is the propagation index or attenuation constant with values ranging between 2 to 4.

    (4.0) Design Criterion:

    An optimal frequency plan requires minimal interference between co-channel and adjacentchannel cells, GSM Rec. 05.05 has defined the interference ratios for co-channel andadjacent channel cells. The actual interference ratio shall be less than a

    specified limit, called the reference interference ratio. The reference interference ratio shall befor base station and all types of MS,- for cochannel interference : C/Ic = 9 dB

    - for adjacent (200 kHz) interference : C/Ia1 = - 9 dB

    - for adjacent (400 kHz) interference : C/Ia2 = - 41 dB

    For the network planning purpose it is recommended that a value of C/Ic 9 dB and the first

    adjacent channel C/Ia -9 dB. This implies that the first adjacent channel should not be usedin the same sector cell or the same base station.

    Page 6 of 10

    r

    R

    D

    i

    j

    i

    jD

  • 8/6/2019 GSM Planning

    7/10

    Frequency Planning

    (4.1) Example:

    As an illustration let us consider that we require to design a system with C/I of 12 dB and we

    have from field drive test results the value of as 3.5, inserting these values in equation

    (D/R)

    = 6 (C/I) we have,

    (D/R)3.5 = 6 x 10.78 = 64.75

    3.5 Log(D/R) = Log(64.75) = 1.81

    (D/R) = Antilog(1.81/3.5)

    This gives (D/R) = 3.29.

    With this we can back calculate the required cluster size from equation D/R = (3 N)1/2 as,

    N = (3.29)2 / 3 = 3.61

    Hence a cluster size of 4 will satisfy our required C/I criteria rather if we back calculate forCluster of size 4 then we get C/I of 19dB.

    (5.0) Frequency Channel Allocation:

    In GSM systems we divide the total allocated spectrum into two sub-groups one for Controlinformation with traffic referred to as BCCH frequency and other only for traffic referred to asTCH (or non-BCCH) frequency. In case where the network has Microcells then the total bandallotted is divided for BCCH and TCH, wherein each band is further sub-divided forMacrocellular & Microcellular applications. Figure (1.3) explains the concept,

    Fig.- (1.3) Frequency band allocation.

    The re-use may differ for both the groups, as little or no compromise is made for BCCHfrequency interference whereas certain compromise could be made for TCH frequencyinterference. Typically a cluster size of 4 or 7 is considered for BCCH re-use whereas acluster size of 3 or 4 is used for TCH re-use. The number of channels in each group dependson the spectrum allocated and C/I criteria for re-use in each case.

    (5.1) Example:

    Page 7 of 10

    TCH TCHBCCH

    Micro CellMacro Cell

  • 8/6/2019 GSM Planning

    8/10

    Frequency Planning

    As an example consider C/I criteria of 12 dB for BCCH then the cluster size of 4 gives thebetter result whereas if the C/I criteria is 9 dB for TCH, gives the cluster size of 3. The figure(1.1) illustrates the case 4 x 3 re-use pattern for BCCH and the figure (1.4) below illustratesthe case of 3 x 3 re-use pattern for TCH,

    Fig.- (1.4) 3 x 3 re-use pattern.

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

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

    A1 B1 C1 A2 B2 C2 A3 B3 C3

    TCH 13 14 15 16 17 18 19 20 21

    For DCS1800 planning with cluster size of 7 the frequency grouping is as follows,In case of DCS1800 where a large band of spectrum is available the BCCH and TCH re-usecan be kept the same.

    Set A1 B1 C1 D1 E1 F1 G1 A2 B2 C2 D2 E2 F2 G2 A3 B3 C3 D3 E3 F3 G3

    BCCH 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

    TCH1 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42

    TCH2 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63

    (6.0) BSIC Planning:

    In addition to the assignment of frequency group to a cell, a Base Station Identity Code(BSIC) must be assigned in association with the frequency group. This will eliminate thepossibility of incorrect cell identification and will allow the evolution to future cell architecture.The BSIC is a two-digit code wherein the first digit is indicates NCC (Network Colour Code)and the second digit indicates BCC (Base Station Colour Code). The NCC and BCC havevalues ranging from 0 to 7, where the NCC is fixed for an operator, signifying at any givenpoint there can be maximum of 8 operators in an area. The BCC defines the cluster numberwhich means a group of 8 clusters carry unique identity which are re-used for another groupof 8 clusters and so on. The principal for allocation of the BSIC is the same as for the RFcarriers but at cluster level rather than cell level. The concept can be understood in thefollowing example,

    (6.1) Example:

    Assume a network with 100 base stations each having three sectors. The BCCH and TCHshare the same re-use plan 4 x 3. Which means we have cluster of 4 base stations, and in all

    we have 100/4 = 25 clusters. Assume NCC code allocated is 6, which gives us clustersstarting from number 61 to 67. Hence seven clusters form a group and hence we have 25/7that is 3 groups of 7 clusters plus additional 4 clusters which form part of the 4 th group. The

    Page 8 of 10

    A1

    A2

    A3C1

    C2

    C3B1

    B2

    B3A1

    A2

    A3

    A1

    A2

    A3

    A1

    A2

    A3

    A1

    A2

    A3C1

    C2

    C3

    A1

    A2

    A3

    B1

    B2

    B3

    B1

    B2

    B3

    C1

    C2

    C3

    C1

    C2

    C3

    C1

    C2

    C3

    C1

    C2

    C3

  • 8/6/2019 GSM Planning

    9/10

    Frequency Planning

    reuse of these 7 clusters group for BSIC numbered from 61 to 67 is shown in the figure (1.5)below,

    Fig.-(1.5) BSIC 7 re-use cluster plan.

    It should be noted that since BSIC are defined at cell (sector) level, hence there are everypossible chances that the three sectors within the same site can have different BSIC. Thereason being as BSIC is used for cell identification hence cells with same BCCH frequencybut different BSIC can be easily discriminated by the MS.

    (7.0) Automatic Frequency Planning:

    Automatic frequency planning is an feature offered by the planning tools to speed up the workof channel assignment and presents more reliable frequency assignment to sites. AFP(Automatic frequency planning) works on complex algorithm whose calculations are based onthe interference table data, field strength grids and an optional demand density grid (or trafficdistribution table). It allows human interaction at certain points such as assigning penalties todifferent clutter types or allowing interference results to be neglected especially in coverage

    boundaries of the network. AFP is of immense help and provides guidelines in the caseswhere frequency assignment is required for big complex network. Basic frequency planningtool is a standard feature of all available planning tools, however the advanced AFP toolbased on complex algorithm is provided as an optional feature.

    (8.0) Frequency Hopping

    The principle of Frequency Hopping used within GSM is that successive TDMA bursts of aconnection are transmitted via different frequencies-the frequencies belonging to therespective cell according to network planning. This method is called Slow Frequency Hopping(SFH) since the transmission frequency remains constant during one burst. In contrast to FastFrequency Hopping (FFH) where the transmission frequency changes within one burst.

    The effect of frequency hopping is that link quality may change from burst to burst, ie a burstof high BER may be followed by a burst of low BER, since

    Page 9 of 10

    61

    62

    63

    64

    67

    61

    62

    63

    64

    65

    66

    61

    62

    63

    64

    65

    66

    67

    61

    62

    63

    64

    65

    66

    67

    Represent a

    cluster of 4 siteseach having 3

    sectors

  • 8/6/2019 GSM Planning

    10/10

    Frequency Planning

    Short term fading is different on different frequencies,

    The interference level is different on different frequencies.

    The results of frequency hopping are improvement in the received quality in fading situationand interference averaging.

    (8.1) Frequency Hopping Techniques:

    The hopping techniques can be broadly classified into two main categories. They are,

    Base band Hopping

    Synthesised Hopping

    As Frequency Hopping is a subject in it self, a separate document will be writtenconcentrating on Frequency Hopping Techniques in near future.

    HSN

    The test set offers two basic forms of frequency hopping

    algorithms: cyclic and pseudo random. When set to the cyclic

    form, the test set and the mobile station are cycled through

    a fixed repeated pattern of frequencies. There are a total

    of 64 different frequency patterns that the test set can

    generate and use. The hopping sequence the mobile station

    uses depends on the Hopping Sequence Number (HSN) specified

    in the test set. An HSN of zero corresponds to the cyclic

    hopping sequence, and values 1 through 63 correspond to thepseudo random patterns. The ARFCNs used in the hopping

    sequence pattern are determined by the contents of the test

    set's Mobile Allocation (MA) Table. The entry of the MA

    Table at which the hopping sequence begins is called the

    Mobile Allocation Index Offset (MAIO). Note that an MAIO of

    zero corresponds to the first entry of the MA Table.

    The hopping sequence number identifies the frequency hopping pattern that a mobile radioshould use when communicating with the system.

    Page 10 of 10