1 GSM Principles and Call Flow

1.1 GSM Frequency Band Allocation

GSM cellular system can be divided into GSM900M and DCS1800M accordin to
!requency band" #it$ carrier !requency interval o! %00 &'( and u) and do#n
!requencies as !ollo#s*
+able 1,1 GSM !requency allocation
Frequency band-M'(." Band#idt$-M'(." Frequency number" Carrier !requency
number -)air.
GSM900 /) 8900911 Do#n 9210930" %1" 101%4" 1%4
DCS1800 /) 151001581 Do#n 180101880" 51" 11%0881" 254
6/)7 and 6do#n7 are classi!ied accordin to base station. Base station
transmittin , mobile station receivin is 6do#n78 mobile station transmittin ,
base station receivin is u).
9it$ t$e e:)andin services" GSM )rotocol adds ;GSM-e:)anded GSM !requency
band. and <GSM -e:)anded GSM !requency band includin rail#ay service. to
t$e oriinal GSM900 !requency band. +$e !requency band allocation is as
!ollo#s*
+able 1,% ;GSM=<GSM !requency allocation
Frequency band-M'(." Band#idt$ -M'(." Frequency number" Carrier !requency
number -)air. ;GSM /) 8800911 Do#n 9%10930" 21" 001%4 " 951010%2" 154
<GSM /) 8530911 Do#n 9%10930" 40" 001%4" 911010%2" 199
1.2 Multiple Access Technology and Logical Channel
1.%.1 GSM Multi)le Access +ec$noloy
>n cellular mobile communications system" since many mobiles stations
communicate #it$ ot$er mobiles stations t$rou$ one base station" it is
necessary to distinuis$ t$e sinals !rom di!!erent mobile stations and base
stations !or t$em to identi!y t$eir o#n sinals. +$e #ay to t$is )roblem is
called multi)le access tec$noloy. +$ere are no# !ive ?inds o! Multi)le access
tec$noloy" namely* Frequency Division Multi)le Access -FDMA." +ime Division
Multi)le Access -+DMA." Code Division Multi)le Access -CDMA." S)ace Division
Multi)le Access -SDMA." and )olar division multi)le access -@DMA..
GSM multi)le access tec$noloy !ocuses on +DMA" and ta?es FDMA as
com)lement. +$e !ollo#in only introduces FDMA and +DMA tec$noloies.
>. FDMA
FDMA divides t$e #$ole !requency band into many sinle radio c$annels
-transmittin and receivin carrier !requency )airs.. ;ac$ c$annel transmits
one )at$ o! s)eec$ or control in!ormation. Any subscriber $as access to one o!
t$ese c$annels under t$e control o! t$e system.
Analo cellular system is a ty)ical e:am)le o! FDMA a))lication. Diital cellular
system also uses FDMA" but not t$e )ure !requency allocation. For e:am)le"
GSM ta?es FDMA tec$noloy.
>>. +DMA
+DMA divides a broadband radio carrier into several time division c$annels
accordin to time -or timeslot.. ;ac$ subscriber ta?es one timeslot and sends
or receives sinals only in t$e s)eci!ied timeslot. +DMA is a))lied in diital
cellular system and GSM.
GSM ado)ts a tec$noloy combined #it$ FDMA and +DMA.
1.%.% +DMA Frame
+$e basic conce)tion o! GSM in terms o! radio )at$ is burst. Burst is a
transmission unit consists o! over one $undred o! modulation bits. >t $as a
duration limit and ta?es a limited radio !requency. +$ey are e:)orted in time
and !requency #indo# #$ic$ is called slot. +o be s)eci!ic" in system !requency
band" central !requency o! slot is set in every %00 &'( -in FDMA.. Slot occurs
)eriodically in eac$ 11=%3 ms" #$ic$ is about 0.155 ms -in +DMA..+$e interval
bet#een t#o slots is called timeslot. >ts duration is used as time unit" called
burst )eriod -B@..
+ime=!requency ma) illustrates t$e conce)t o! slot. ;ac$ slot is e:)ressed as
one little rectanle #it$ 11=%3ms lent$ and %00 &'( #idt$. See Fiure 1,1.
Similarly" t$e %00 &'( band#idt$ in GSM is called !requency slot" equal to radio
!requency c$annel in GSM )rotocol.
Burst re)resents di!!erent meanin in di!!erent situation. Sometimes it
concerns time 0 !requency 6rectanle7 unit" and sometimes not. Similarly"
timeslot sometimes concerns time value" and sometimes means usin one o!
every ei$t slots )eriodically.
/sin a iven c$annel means transmittin burst #it$ a )articular !requency at
)articular time" t$at is" a )articular slot. Generally" t$e slot o! a c$annel is not
continuous in time.
@$ysical c$annel combines !requency division multi)le access and time division
multi)le access toet$er. >t consists o! timeslot !lo# t$at connects base station
-BS. and mobile station -MS..+$e )osition o! t$ese timeslots in +DMA !rame is
!i:ed. Fiure 1,% s$o#s t$e com)lete structure o! +DMA !rame" includin
timeslot and burst. +DMA !rame is a re)etitive 6)$ysical7 !rame in radio lin?.
Ane +DMA !rame consists o! ei$t basic timeslots" about 30=12B4.311ms in
total. ;ac$ timeslot is a basic )$ysical c$annel #it$ 113.%1 elements" covin
11=%3B0.115ms.
+$ere are t#o ?inds o! multi!rames" consistin o! %3 and 11 continuous +DMA
!rames res)ectively. Multi!rames are a))lied #$en di!!erent loical c$annels
are multi)le used in one )$ysical c$annel.
+$e %3 multi!rame" #it$ a )eriod o! 1%0 ms" is used in tra!!ic c$annel and
associated control c$annel. Amon t$e %3 bursts" %4 are used in tra!!ic and %
are used in sinalin.
+$e 11 multi!rame" #it$ a )eriod o! 2030=12B%21.281 ms" is s)ecially used in
control c$annel.
Many multi!rames toet$er !orm a su)er !rame. Su)er !rame is a continuous
11C%3+DMA !rame" t$at is to say" a su)er !rame consists o! !i!ty,one %3 +DMA
multi!rames or t#enty,si: 11 +AMA multi!rames. +$e )eriod o! su)er !rame is
1"2%3 +DMA !rames" or 3.1% s.
Many su)er !rames toet$er !orm a $y)er !rame.
A $y)er !rame consists o! %"048 su)er !rames #it$ a )eriod o! 1%"122.5s" or 2
$ours and %8D 12DD 530DDD. >t is used in encry)ted voice and data. ;ac$ )eriod o!
$y)er !rame consists o! %"511"348 +DMA !rames numbered !rom 0 to %"511"348.
+$e !rame number is transmitted in sync c$annel.
1.%.2 Burst
Burst is t$e messae layout o! a timeslot in +DMA c$annel" #$ic$ means eac$
burst is sent to a timeslot o! +DMA !rame.
Di!!erent messae in t$e burst determines its layout.
+$ere are !ive ?inds o! bursts*
Eormal burst* used to carry messaes in +C'" FACC'" SACC'" SDCC'" BCC'"
@C' and AGC' c$annels
Access burst* used to carry messae in <AC' c$annel
Frequency correction burst* used to carry messae in FCC' c$annel
Sync$roni(ation burst* used to carry messae in SC' c$annel
Dummy burst* transmitted #$en no s)eci!ic messae transmission request !rom
system ->n cells" standard !requency sends messae continuously.
;ac$ ?ind o! burst includes t$e !ollo#in elements*
+ail bits* >ts value is al#ays 0 to $el) equali(er Fude start bit and sto) bit to
avoid lost sync$roni(ation.
>n!ormation bits* >t is used to describe tra!!ic and sinalin in!ormation" e:ce)t
idle burst and !requency correction burst.
+rainin sequence* >t is a ?no#n sequence" used !or equali(er to enerate
c$annel model -a #ay to eliminate dis)ersion.. +rainin sequence is ?no#n by
bot$ transmitter and receiver. >t can be used to identi!y t$e location o! ot$er
bits !rom t$e same burst and rou$ly estimate t$e inter!erence situation o!
transmission c$annel #$en t$e receiver ets t$is sequence. +rainin sequence
can be divided into ei$t cateories in normal burst. >t usually $as t$e same
BCC settin #it$ cells" but #$en accessed to burst and sync$roni(ation bust"
trainin sequence is !i:ed and does not c$ane #it$ cells. For e:am)le" in
access burst" trainin sequence is !i:ed -occu)yin 41 bits.. +$e 23,bit messae
diit o! t$e random access burst includes BS>C in!ormation o! t$e cell. BS>C
settins o! t$e same BCC' s$ould be di!!erent" in order to avoid mis,decodin
o! random access burst !rom nei$borin cells into local access.
Guard )eriod* >t is a blan? s)ace. Since eac$ carrier !requency can carry a
ma:imum o! ei$t subscribers" it is necessary to uarantee t$e non,overla))in
o! eac$ timeslot in transmission. Alt$ou$ timin advance tec$noloy
-introduced later. is used" bursts !rom di!!erent mobile stations still s$o# little
sli)s8 t$ere!ore" )rotection interval is ado)ted to allo# transmitter to !luctuate
in a )ro)er rane in GSM. An t$e ot$er $and" GSM requires )rotection bits to
?ee) constant transmission am)litude o! t$e e!!ective burst -e:ce)t )rotection
bits. and )ro)erly attenuate t$e transmission am)litude o! mobile station. +$e
am)litude attenuation o! t#o sequential bursts as #ell as )ro)er modulation
bit stream can reduce t$e inter!erence to ot$er <F c$annels.
+$e !ollo#in is a detailed introduction to t$e structure and content o! burst*
Access burst
>t is used !or random access -c$annel request !rom net#or? and s#itc$over
access..
>t is t$e !irst burst t$at t$e base station needs in u)lin? modulation.
Access burst includes a 41,bit trainin sequence" 23,in!ormation bit" and its
)rotection interval is 38.%1 bits. +$ere is only one ?ind o! trainin sequence in
access burst. Since t$e )ossibility o! inter!erence is rat$er little" it is
unnecessary to add e:tra ?inds o! trainin sequences. Bot$ trainin sequence
and )rotection interval are loner t$an normal bursts in order to o!!set t$e bu
o! timin advance inorance in t$e !irst access o! mobile station -or s#itc$ over
to anot$er B+S. and im)rove demodulation ability o! t$e system.
Frequency correction burst
>t is used !or !requency sync$roni(ation in mobile station" equal to an
unmodulated carrier. +$is sequence $as 14% constant bits !or !requency
sync$roni(ation. >ts structure is )retty sim)le #it$ all constant bits bein 0.
A!ter modulated" it becomes a )ure sine #ave. >t is used in FCC' c$annel !or
mobile station to !ind and modulate sync$roni(ation burst o! t$e same cell.
9$en mobile station ets t$e !requency t$rou$ t$is burst" it can read t$e
in!ormation o! !ollo#in bursts -suc$ as SC' and BCC'. in t$e same )$ysical
c$annel. @rotection interval and tail bit are t$e same #it$ t$at o! normal
burst.
Sync$roni(ation burst
9it$ a 34,bit trainin sequence and t#o 29,bit in!ormation !ields"
sync$roni(ation burst is used !or time sync$roni(ation o! mobile station in SC'
c$annel. >t belons to do#nlin?. Since it is t$e !irst burst required to be
modulated by mobile station" its trainin sequence is relatively lon and easy
to be detected.
Eormal burst
>t $as t#o 18,bit rou)s used in messae !ield. +o be more s)eci!ic" t#o 18,bit
rou)s are used to transmit subscriber data or voice toet$er #it$ t#o stealin
!las. Eormal burst is used to describe #$et$er t$e transmitted is tra!!ic
in!ormation or sinalin in!ormation. For e:am)le" to distinuis$ +C' and
FACC' -#$en +C' c$annel is used as FACC' c$annel to transmit sinalin" t$e
stealin !la o! t$e 8 $al! bursts s$ould be set to 1. >t $as no ot$er use in
c$annels e:ce)t in +C' c$annel" but can be rearded as t$e e:tension o!
trainin sequence and al#ays set to 1.Eormal burst also includes t#o 2,bit tails
and a )rotection interval o! 8.%1 bits. +$e only bu is t$at t$e receiver $as to
store t$e )recedin )art o! burst be!ore modulation. Eormal burst $as a total
o! %3 bits" 13 o! #$ic$ are in!ormation bits. >n order to et %3 bits" it co)ies
t$e !irst !ive bits to t$e end o! t$e trainin sequence and t$e last !ive bits to
t$e $ead o! t$e trainin sequence. +$ere are ei$t ?inds o! suc$ trainin
sequence -t$ese ei$t sequences $ave t$e least relevancy #it$ eac$ ot$er..
+$ey corres)ond to di!!erent base station color code -BCC" 2 bits. res)ectively
to distinuis$ t$e t#o cells usin t$e same !requency.
Dummy burst
+$is ?ind o! bust is sometimes sent by B+S #it$out carryin any in!ormation. >ts
!ormat is t$e same #it$ normal burst. +$e encry)ted bits are c$aned into
mi:ed bits #it$ certain bit model.
1.%.4 Goical C$annel
>n real net#or?in" eac$ cell $as several carrier !requencies and eac$
!requency $as ei$t timeslots" )rovin ei$t basic )$ysical c$annels. Goical
c$annel carries out time multi)le:in in one )$ysical c$annel. >t is classi!ied
accordin to t$e ty)e o! in!ormation in )$ysical c$annel. Di!!erent loical
c$annel transmits di!!erent ty)e o! in!ormation bet#een BS and MS" suc$ as
sinalin and data service. GSM de!ines di!!erent burst ty)e !or di!!erent
loical c$annel.
>n GSM" loical c$annel is divided into dedicated c$annel -DC'. and common
c$annel -CC'." or tra!!ic c$annel -+C'. and control c$annel -CC'. sometimes.
>. +C'
+C' carries coded voice or subscriber data. >t is divided into !ull rate +C'
-+C'=F. and $al! rate +C' -+C'='. #it$ %%.8 bit=s in!ormation and 11.4 &bit=s
in!ormation res)ectively. /sin $al! o! t$e timeslots in +C'=F can et +C'='. A
carrier !requency can )rovide ei$t ?inds o! +C'=F or si:teen ?inds o! +C'='.
Hoice c$annel ty)es are as !ollo#s*
;n$anced !ull rate s)eec$ +C' -+C'=;FS.
Full rate s)eec$ +C' -+C'=;FS.
Full rate 9.3 &bit=s +C' -+C'=F9.3.
Full rate 4.8 &bit=s +C' -+C'=F4.8.
Full rate I%.4 &bit=s +C' -+C'=F%.4.
>>. CC'
CC' is used to transmit sinalin or sync$ronous data. >t mainly consists o!
broadcast c$annel -BCC'." common control c$annel -CCC'." and dedicated
control c$annel -DCC'..
>>>. BCC'
Frequency Correction C$annel -FCC'.
>t carries t$e in!ormation !or !requency correction in mobile station. +$rou$
FCC'" mobile station can locate a cell and demodulate ot$er in!ormation in t$e
same cell" and reconi(e #$et$er t$is carrier !requency is BCC' or not.
Sync C$annel -SC'.
A!ter FCC' decodin" mobile station $as to decode SC' in!ormation. +$is
in!ormation contains mobile station !rame sync$roni(ation and base station
identi!ication. Base station identi!ication code -BS>C. occu)ies si: bits" t$ree o!
#$ic$ are @GME color codes ranin !rom (ero to seven" and t$e ot$er t$ree
are base station color codes -BCCs. ranin !rom (ero to seven.
<educed +DMA !rame -<FE. occu)ies %% bits.
BCC'
Generally" eac$ B+S $as a transceiver containin BCC' in order to broadcast
system in!ormation to mobile station. System in!ormation enables mobile
station to #or? e!!iciently in null state.
>H. CCC'
@ain C$annel -@C'.
@C' is a do#nlin? c$annel used to )ae mobile station. 9$en t$e net#or?
#ants to communicate #it$ a certain mobile station" it sends )ain
in!ormation mar?ed as +MS> or >MS> t$rou$ @C' to all t$e cells in GAC area
accordin to t$e current GAC reistered in mobile station.
Access Grant C$annel -AGC'.
AGC' is a do#nlin? c$annel used !or base station to res)ond t$e net#or?
access request o! mobile station" t$at is" to allocate a SDCC' or +C' directly.
AGC' and @C' s$are t$e same radio resource. &ee) a !i:ed number o! bloc?s
!or AGC' or Fust borro# @C' #$en AGC' requires #it$out ?ee)in s)ecial
AGC' bloc? -AGB..
<andom Access C$annel -<AC'.
<AC' is an u)lin? c$annel used !or mobile station to request SDCC' allocation
in random net#or? access a))lication. +$e request includes t$e reason to build
2,bit -call request" )ain res)onse" location u)date request and s$ort messae
request. and 1,bit re!erence random number !or mobile station to identi!y its
o#n access rant messae.
H. DCC'
Stand,alone Dedicated Control C$annel -SDCC'.
SDCC' is a bi,directional dedicated c$annel used to transmit in!ormation o!
sinalin" location u)date" s$ort messae" aut$entication" encry)ted
command" c$annel allocation" and com)lementary services. >t can be divided
into SD=8 and SD=4.
Slo# Associated Control C$annel -SACC'.
SACC' #or?s #it$ tra!!ic c$annel or SDCC' to transmit subscriber in!ormation
and some s)eci!ic in!ormation at t$e same time. /)lin? mainly transmits radio
measurement re)ort and t$e !irst layer $ead in!ormation8 do#nlin? mainly
transmits )art system in!ormation and t$e !irst layer $ead in!ormation. +$e
in!ormation includes quality o! communications" GA>" C;GG >D" BCC' sinal
strent$ in nei$borin cells" ECC limit" cell o)tions" +A" and )o#er control
level.
Fast Associated Control C$annel -FACC'.
FACC' #or?s #it$ +C' to )rovide sinalin in!ormation #it$ a rate and
timeliness muc$ $i$er t$an t$at )rovided by SACC'.
+$ere is anot$er control c$annel called cell broadcast c$annel -CBC'. besides
t$e t$ree control c$annels mentioned above. >t is used in do#nlin? and carries
s$ort messae service cell broadcast -SMSCB. in!ormation. CBC' uses a )$ysical
c$annel same as SDCC'.
H>. C$annel Combination
Goical c$annel is ma))ed to )$ysical c$annel accordin to certain rules. +$e
c$annel combinations s)eci!ied in GSM )rotocol are as !ollo#s*
+C'=F J FACC'=F J SACC'=+F
+C'='-0"1. J FACC'='-0"1. J SACC'=+'-0"1.
+C'='-0"0. J FACC'='-0"1. J SACC'=+'-0"1. J +C'='-1"1.
FCC' J SC' J BCC' J CCC' -main BCC'.
FCC' J SC' J BCC' J CCC' J SDCC'=4-0..2. J SACC'=C4-0..2.-BCC'
combination.
BCC' J CCC'-BCC' e:tension.
SDCC'=8-0. .5. J SACC'=C8-0. .5.
H>>. /ncombined BCC'=SDCC' and Combined BCC'=SDCC'
@ain in!ormation transmits in t$e timeslot 0 o! BCC'. +imeslot 0 $as t$e
!ollo#in sub c$annels*
Broadcast c$annel -BC'.* FCC'" SC'" BCC'
CCC'* @C'" AGC'
DCC' -combined BCC'=SDCC'.* SDCC'" SACC'" CBC' - i! usin cell broadcast.
@$ysical c$annel timeslot 0 is made o! multi!rames loically. ;ac$ multi!rame
is %21.4 ms in lent$. Multi!rame $as di!!erent c$annel con!iurations" suc$ as
combined BCC'=SDCC' and uncombined BCC'=SDCC'. Di!!erent con!iuration
$as di!!erent )ain ca)acity.
/ncombined BCC'=SDCC'
;ac$ !rame o! /ncombined BCC'=SDCC' can $ave nine )ain bloc?s. +$e
timeslot 0 o! BCC' carrier !requency does not $ave SDCC' c$annel or CBC'
c$annel.
Combined BCC'=SDCC'
;ac$ multi!rame o! combined BCC'=SDCC' can $ave t$ree )ain bloc?s. +$e
timeslot 0 o! BCC' carrier !requency contains !our SDCC' subc$annels -no
CBC'. or t$ree SDCC' and one CBC' subc$annel.
+$e con!iuration o! combined BCC'=SDCC' $as a reat in!luence on )ain
ca)acity. ;ac$ multi!rame $as only t$ree )ain bloc?s instead o! nine in
uncombined BCC'=SDCC'" #$ic$ means t$e )ain ca)acity o! cells #it$
combined BCC'=SDCC' is only one t$ird o! t$at o! cells #it$ uncombined
BCC'=SDCC'.
1.2 Data +ransmission
<adio c$annel $as totally di!!erent c$aracteristics !rom #ired c$annel. <adio
c$annel $as a stron time,varyin c$aracteristic. >t $as a $i$ error rate #$en
t$e sinal is in!luenced by inter!erences" multi)at$ !adin" or s$ado# !adin. >n
order to solve t$ese )roblems" it is necessary to )rotect t$e sinals t$rou$ a
series o! trans!ormation and inverse trans!ormation !rom oriinal subscriber
data or sinalin data to t$e in!ormation carried by radio #ave and t$en to
subscriber data or sinalin data. +$ese trans!ormations include c$annel codin
and decodin" interleavin and de,interleavin" burst !ormattin" encry)tion
and decry)tion" modulation and demodulation.
1.2.1 Hoice Codin
Modern diital communication system usually uses voice com)ression
tec$noloy. GSM ta?es tone and noise !rom $uman t$roat as #ell as t$e mout$
and tonue !ilter e!!ect o! acoustics as voice encoder to establis$ a model. +$e
model )arameters transmit t$rou$ +C' c$annel.
Hoice encoder is based on residual e:cited linear )rediction encoder -<;>@. and
its com)ression e!!ect is strent$ened t$rou$ lon term )redictor -G+@.. G+@
im)roves residual data encodin by removin t$e vo#el )art o! voice.
Hoice encoder divides voice into several %0 ms voice bloc?s and sam)les eac$
bloc? #it$ 8 ?'(" so eac$ bloc? $as 130 sam)les. ;ac$ sam)le is quanti!ied
t$rou$ !requency A 12 bits -!requency K 14 bits.. Since t$e com)ression rates
o! !requency A and !requency K are di!!erent" add t$ree and t#o 607 bits to
t$e quanti!ication values res)ectively" and t$en eac$ sam)le ets 13 bits
quanti!ication value. +$ere!ore" 1%8 &bit=s data !lo# is obtained a!ter
diiti(in but be!ore encodin. +$is data !lo# is too !ast to transmit in radio
)at$ and $as to be com)ressed in encoder. 9it$ !ull s)eed encoder" eac$ voice
bloc? is encoded into %30 bits to !orm a 12 &bit=s source codin rate. Ee:t is
c$annel codin. 9it$ %0 ms as a unit" %30 bits are out)ut a!ter com)ression
encodin" so t$e encodin rate is 12&bit =s.
Com)ared #it$ t$e direct codin transmission o! voice in traditional @CM
c$annel" t$e 12?b)s voice rate o! GSM is muc$ lo#er. More advance voice
encoder can reduce t$e rate to 3.1?b)s -$al! rate encodin..
1.2.% C$annel Codin
C$annel codin is used to im)rove transmission quality and remove t$e
in!luence o! inter!erential !actors on sinals at t$e )rice o! increasin bits and
in!ormation. +$e basic #ay o! codin is addin some redundant in!ormation to
t$e oriinal data. +$e added data is calculated on t$e basis o! oriinal data
#it$ certain rules. +$e decodin )rocess o! receivin end is Fudin and
correctin errors #it$ t$is redundant bit. >! t$e redundant bit o! received data
calculated #it$ t$e same #ay is di!!erent !rom t$e received redundant bit"
errors must $ave occurred in transmission. Di!!erent code is used in di!!erent
transmission mode. >n )ractice" several codin sc$emes are al#ays combined
toet$er. Common codin sc$emes include bloc? convolutional code" error
correctin cyclic code and )arity code.
>n GSM" eac$ loical c$annel $as its o#n codin and interleavin mode" but t$e
)rinci)le is tryin to !orm a uni!ied codin structure.
;ncode in!ormation bit into a uni!ied bloc? code consistin o! in!ormation bits
and )arity c$ec? bits.
;ncode bloc? code into convolutional code and !orm codin bits -usually 413
bits..
<eassemble and interleave codin bits and add a stealin !la to !orm
interleavin bits.
All t$ese o)erations are based on bloc?. +$e bloc? si(e de)ends on c$annel
ty)e. A!ter c$annel codin" all c$annels -e:ce)t <AC' and SC'. are made o!
434,bit bloc?" t$at is" 413 coded in!ormation bits )lus 8,bit $eader -$eader is
used to distinuis$ +C' and FACC'.. +$en t$ese bloc?s are reinterleaved
-concernin c$annel..
>n +C'=F voice service8 t$is bloc? carries one s)eec$ !rame o! in!ormation. >n
control c$annel" t$is bloc? usually carries one )iece o! in!ormation. >n +C'='
voice service" s)eec$ in!ormation is transmitted by a bloc? o! %%8 coded bits
bloc?.
For FACC'" eac$ bloc? o! 413 coded in!ormation bits is divided into ei$t sub
bloc?s. +$e !irst !our sub bloc?s are transmitted by even bits o! t$e !our
timeslots borro#ed !rom t$e continuous !rames o! +C'" and t$e rest !our sub
bloc?s borro#s odd bits o! t$e !our timeslots !rom t$e !our continuous !rames
delayed !or t#o or !our !rames a!ter t$e !irst !rame. ;ac$ 413 coded bit bloc?
$as a stealin !la -8 bits." indicatin #$et$er t$e bloc? belons to +C' or to
FACC'. >n t$e case o! SACC'" BCC' or CCC'" t$is stealin !la is dummy.
+$e sync$ronous in!ormation in Do#nlin? SC' and t$e random access
in!ormation in u)lin? use s$ort coded bit bloc?s transmitted in t$e same
timeslot.
>n +C'=F" a %0ms s)eec$ !rame is encoded into 413,bit code sequence. +$e %30
bits o! t$e 12 &bit=s %0ms s)eec$ !rame can be divided into t$ree cateories*
10 most im)ort bits" 12% im)ortant bits and 58 unim)ortant bits. Add 2 )arity
c$ec? bits to t$e 10 most im)ortant bits" and t$ese 12 bits toet$er #it$ 12%
im)ortant bits and 4 tail bits are convolutionally encoded - #it$ 1=%
convolutional codin rate . into 258 bits" )lus t$e 58 unim)ortant bits" and t$e
413 bits code sequence is obtained.
>n BCC'" @C'" AGC'" SDCC'" FACC' and SACC'" data is transmitted by Gin?
Access @rocedure on t$e Dm c$annel -GA@Dm.. ;ac$ GA@Dm !rame $as 184 bits"
toet$er #it$ 40 bits error correctin cyclic code and 4 tail bits" t$rou$ 1=%
convolutional codin rate" and t$e 413 bits code sequence is obtained.
;ac$ SC' contains %1,bit messae !ield. Amon t$em" 19 bits are !rame number
and 3 bits are BSC number. +$ese %1 bits )lus 10 )arity c$ec? bits and 4 tail
bits are 29 bits. +$rou$ 1=% rate convolutional codin" 58 bits are obtained"
#$ic$ occu)y an entire SC' burst. .
<AC' messae only $as 8 bits" includin 2,bit setu) cause messae and 1,bit
discrimination symbol. An t$e basis o! t$ese 8 bits" add 3 bits o! color code
-obtained t$rou$ t$e MAD % o! t$e 3,bit BS>C and 3,bit )arity c$ec? code."
)lus 4 tail bits to et 18 bits. +$rou$ 1=% rate convolutional codin" 23 bits are
obtained" #$ic$ occu)y an entire <AC' burst.
1.2.2 >nterleavin
>! s)eec$ sinal is modulated and transmitted directly a!ter c$annel codin"
due to )arametric variation o! mobile communication c$annel" t$e lon trou$
o! dee) !eedin #ill a!!ect t$e succeedin bits" leadin to error bit strins.
+$at is to say" a!ter codin" s)eec$ sinal turns into sequential !rames" #$ile in
transmission" error bits usually occur suddenly" #$ic$ #ill a!!ect t$e accuracy
o! continuous !rames. C$annel codin only #or?s !or detection and correction
o! sinal error or s$ort error strin. +$ere!ore" it is $o)ed to !ind a #ay to
se)arate t$e continuous bits in a messae" t$at is" to transmit t$e continuous
bits in a discontinuous mode so as to c$ane t$e error c$annel into discrete
c$annel. +$ere!ore" even i! an error occurs" it is only about a sinle or very
s$ort bit stream and #ill not interru)t t$e decodin o! t$e entire burst or even
t$e entire in!ormation bloc?. C$annel codin #ill correct t$e error bit under
suc$ circumstances. +$is met$od is called interleavin tec$noloy. >nterleavin
tec$noloy is t$e most e!!ective code rou)in met$od to se)arate error codes.
+$e essence o! interleavin is to dis)erse t$e b bits into n bursts in order to
c$ane t$e adFacent relations$i) bet#een bits. Greater n value leads to better
transmission )er!ormance but loner transmission delay. +$ere!ore" t$ese t#o
!actors must be considered in interleavin. >nterleavin is al#ays related to t$e
use o! c$annel. GSM ado)ts secondary interleavin met$od.
A!ter c$annel codin" +$e 413 bits are divided into ei$t rou)s8 eac$ rou)
contains 15 bits. +$is is t$e !irst interleavin" also called internal interleavin.
A!ter !irst interleavin" t$e continuity o! in!ormation in a rou) is bro?en. As
one burst contains t#o rou)s o! 15,bit voice in!ormation" i! t$e t#o,rou) 15
bits o! a %0 ms voice bloc? a!ter !irst interleavin are inserted to t$e same
burst" t$e loss o! t$is burst #ill lead to %1L loss o! bits !or t$is %0 ms voice
bloc?. C$annel codin cannot restore so muc$ loss. +$ere!ore" a secondary
interleavin" also called inter,bloc? interleavin" is required bet#een t#o voice
bloc?s.
A!ter internal interleavin" t$e 413 bits o! a voice bloc? B are divided into ei$t
rou)s. >nterleave t$e !irst !our rou)s o! voice bloc? B -B0" B1" B%" and B2.
#it$ t$e last !our rou)s o! voice bloc? A -A4" A1" A3" and A3." and t$en -BA"
A4." -B1" A1." -B%" A3." and -B2" A5. !orm !our bursts. >n order to brea? t$e
consistency o! bits" )ut bloc? A at even )osition and bloc? B at odd )osition o!
bursts" t$at is" to )ut B0 at odd )osition and A4 at even )osition. Similarly"
interleave t$e last !our rou)s o! bloc? B #it$ t$e !irst !our rou)s o! bloc? C.
+$ere!ore" a %0 ms s)eec$ !rame is inserted into ei$t normal bursts a!ter
secondary interleavin. +$eses ei$t bursts are transmitted one by one" so t$e
loss o! one burst only a!!ects 1%.1L voice bits. >n addition" as t$ese bursts $ave
no relations #it$ eac$ ot$er" t$ey can be corrected by c$annel codin.
+$e secondary interleavin o! control c$annel -SACC'" FACC'" SDCC'" BCC'"
@C'" or AGC'. is di!!erent !rom voice interleavin #$ic$ requires t$ree voice
bloc?s. +$e 413,bit voice bloc? is divided into ei$t rou)s a!ter internal
interleavin -t$e same as t$at o! voice bloc?." and t$en t$e !irst !our rou)s
are interleaved #it$ t$e last !our rou)s -t$e same interleavin met$od as t$at
o! voice bloc?. to et !our bursts.
>nterleavin is an e!!ective #ay to avoid inter!erence" but it $as a lon delay.
>n t$e transmission o! a %0 ms voice bloc?" t$e delay )eriod is -9M8.,5N31 bursts
-SACC' occu)yin one burst." #$ic$ is 25.1 ms. +$ere!ore" MS and trun? circuit
$ave ec$o cancellers added to remove t$e ec$o due to delay.
1.2.4 ;ncry)tion
Security is a very im)ortant !eature in diital transmission system. GSM
)rovides $i$ security t$rou$ transmission encry)tion. +$is ?ind o! encry)tion
can be used in voice" user data" and sinalin. >t is used !or normal burst only
and $as not$in to do #it$ data ty)e.
;ncry)tion is ac$ieved by OA< o)eration o! )oison random sequence -enerated
t$rou$ A1 alorit$m o! encry)tion ?ey &c and !rame number. and t$e 114
in!ormation bits o! normal burst.
+$e same )oison random sequence enerated at receivin end and t$e received
encry)tion sequence toet$er )roduce t$e required data a!ter OA< o)eration
1.2.1 Modulation and Demodulation
Modulation and demodulation is t$e last ste) o! sinal )rocessin. GSM
modulation ado)ts GMS& tec$noloy #it$ B+ bein 0.2 at t$e s)eed o! %50.822
&bit=s and Hiterbi alorit$m. +$e !unction o! modulation is to add a certain
!eature to electromanetic #ave accordin to t$e rules. +$is !eature is t$e
data to transmit. >n GSM" t$e )$ase o! electromanetic !ield bears t$e
in!ormation.
+$e !unction o! demodulation is to receive sinals and restore t$e data in a
modulated electromanetic #ave. A binary numeral $as to be c$aned into a
lo#,!requency modulated sinal !irst" and t$en into an electromanetic #ave.
Demodulation is t$e reverse )rocess o! modulation.
1.4 +imin advance
Sinal transmission $as a delay. >! t$e MS moves a#ay !rom B+S durin callin"
t$e sinal !rom B+S to MS #ill be delayed" so #ill t$e sinal !rom MS to B+S. >!
t$e delay is too lon" t$e sinal in one timeslot !rom MS cannot be correctly
decoded" and t$is timeslot may even overla) #it$ t$e timeslot o! t$e ne:t
sinal !rom ot$er MS" leadin to inter,timeslot inter!erence. +$ere!ore" t$e
re)ort $eader carries t$e delay value measured by MS. B+S monitors t$e arrive
time o! call and send command to MS #it$ t$e !requency o! 480 ms" )rom)tin
MS t$e timin advance -+A. value. +$e rane o! t$is value is 0032-00%22 us."
and t$e ma:imum coverae area is 21?m. +$e calculation is as !ollo#s*
1=%C2.5us=bitC32bitMcN21?m
2.5us=bit is t$e duration )er bit -113=155.8 32bit is t$e ma:imum bit !or time
coordination8 c is li$t velocity -transmission rate o! sinal.8 1=% is related to
t$e round,tri) o! sinal.
Accordin to t$e )recedin descri)tion" 1bit to 114 m" due to t$e in!luence o!
multi,)at$ transmission and t$e accuracy o! MS sync$roni(ation" +A error may
be about 2 bits -1.3?m..
Sometimes a reater coverae area is required" suc$ as in coastal areas.
+$ere!ore" t$e number o! c$annels t$at eac$ +<O contains must be reduced.
+$e met$od is to bind odd and even timeslots" so t$ere are only !our c$annels
-0=1" %=2" 4=1" and 3=5. !or eac$ +DMA !rame in e:tended cell. Allocate
c$annels 0" %" 4" and 3 to MS. 9it$in 21 &M around B+S" t$e +A value o! MS is in
t$e normal rane 0,328 !or t$e area beyond 21 &M" +A value stays at 32. +$is
tec$noloy is called e:tended cell tec$noloy. +$e ma:imum value o! +A in B+S
measurement re)ort is 32J113.%1N%19.%1 bit" so t$e ma:imum radius o!
coverae area is*
1=%C2.5usC -32J113.%1. C2C108m=sN1%0?m
>n real sc$eme" in order to im)rove t$e utili(ation o! +<O" bot$ common +<Os
and dual timeslot +<Os can be included. BCC' must be in dual timeslot +<O to
receive random access !rom any area. +$e calls #it$in 21 ?m are allocated to
common +<O8 t$e calls #it$in 21 ?m01%0 ?m and t$e s#itc$ed in calls are
allocated to dual timeslot +<O. >! t$e system detects t$e s#itc$ed in call is
#it$in 21?m" it #ill s#itc$ over t$is call to common +<O. >! t$e MS in
conversation oes beyond 21 ?m" an intra,cell s#itc$over #ill be carried out.
+$ere!ore" bot$ t$e ca)acity requirement !or remote areas and t$e coverae
requirement !or local areas can be satis!ied.
1.1 System >n!ormation
System in!ormation is sent to MS !rom net#or? in broadcast !orm. >t in!orms all
t$e MSs #it$in t$e coverae area o! location area" cell selection and re,
selection" nei$bor cell in!ormation" c$annel allocation and random access
control. By receivin system in!ormation" MS can quic?ly and accurately locate
net#or? resources and ma?e !ull use o! all ?inds o! services t$at net#or?
)rovides. +$ere are 13 ty)es o! system in!ormation* ty)e1" %" %bis" %ter" 2" 4"
1" 1bis" 1ter" 3" 5" 8" and 12.
System in!ormation is transmitted on BCC' or SACC'. MS receives system
in!ormation in di!!erent mode !rom di!!erent loic c$annel.
>n idle mode" system in!ormation 10 4" 5" and 8 are transmitted on BCC' 8
>n communication mode" system in!ormation 1 and 3 are transmitted on SACC'8
+$e content o! system in!ormation is as !ollo#s*
System in!ormation 1cell c$annel descri)tion J <AC' control )arameter"
transmitted on BCC'
System in!ormation % !requency descri)tion o! nei$bor cell J <AC' control
in!ormation J net#or? color code -ECC. )ermitted" transmitted on BCC'" used
!or cell re,selection
System in!ormation %bis ;:tended nei$bor cell BCC' !requency descri)tion J
<AC' control in!ormation" transmitted on BCC'" used !or cell re,selection.
System in!ormation %ter ;:tended nei$bor cell BCC' !requency descri)tion"
transmitted on BCC'" used !or cell re,selection.
System in!ormation 2 Cell identity J location area identity -GA>. J control
c$annel descri)tion J cell selection J cell selection )arameter J <AC' control
)arameter" transmitted on BCC'.
System in!ormation 4 GA> J cell selection )arameter J <AC' control
)arameter J CBC' c$annel descri)tion J CBC' mobile con!iuration"
transmitted on BCC'.
System in!ormation 1 Eei$bor cell BCC' !requency descri)tion" transmitted
on SACC' c$annel" used !or cell $andover.
System in!ormation 1bis ;:tended nei$bor cell BCC' !requency descri)tion"
transmitted on SACC' c$annel" used !or cell $andover.
System in!ormation 1ter ;:tended nei$bor cell BCC' !requency descri)tion"
transmitted on SACC' c$annel" used !or cell $andover.
System in!ormation 3 Cell Global >denti!ication -CG>. J cell o)tionECC
@ermitted" transmitted on SACC'.
System in!ormation 5 cell re,selection )arameter
System in!ormation 8 cell re,selection )arameter
BCC' is a lo#,ca)acity c$annel" every 11 multi!rames --%21 ms. $ave only !our
!rames -one in!ormation bloc?. to transmit a %2 byte GA@Dm messae.
;ac$ in!ormation unit contains*
Cell c$annel descri)tion contains all t$e !requencies used in t$is cell.
<AC' control in!ormation contains )arameters suc$ as Ma: <etrans"
+OPinteer" CBA" <;" ;C" and AC CE.
Eei$bor cell BCC' !requency descri)tion contains t$e BCC' !requency t$at
t$e nei$bor cell uses.
Allo#ed @GME is used to )rovide ECC @ermitted t$at MS monitors on BCC' +<O.
Control c$annel descri)tion contains )arameters suc$ as MS
A++AC'=D;A++AC' allo#ed >ndicator A++" BS,AG,BG&S,<;S" CCC',CAEF" BA,
@A,MF<MS" and +2%1%.
Cell selection contains )arameters suc$ as )o#er control -@9<C. indication"
discontinuous +ransmission -D+O. indication" and <AD>A,G>E&,+>M;A/+.
Cell selection )arameter contains )arameters suc$ as cell re,selection
$ysteresis" MS,+O@9<,MAO,CC'" and <OG;H,ACC;SS,M>E.
CBC' c$annel descri)tion contains c$annel ty)e and +DMA deviation -t$e
combination mode o! dedicated c$annel." timeslot number -+E." trainin
sequence code -+SC." $o))in !requency c$annel indication '" mobile
allocation inde: o!!set -MA>A." $o))in !requency sequence number -'SE. and
absolute radio !requency c$annel number - A<FCE..
CBC' mobile con!iuration contains t$e relations$i) bet#een $o))in c$annel
sequence and cell c$annel descri)tion.
Cell re,selection )arameter contains C;GG<;S;G>ED" cell bar quali!y -CBQ." cell
reselection o!!set -C<A." tem)orary o!!set -+A." and )enalty time -@+..
1.3 Cell Selection and <e,Selection
1.3.1 Cell Selection
9$en a MS is s#itc$ed on" it tries to contact GSM @GME t$at t$e S>M )ermits
and select a )ro)er cell to e:tract control c$annel )arameters and ot$er
system in!ormation. +$is )rocess is called cell selection.
+$e )riority levels o! cells include normal" lo#" and barred. Go# )riority level
cell is selected #$en t$ere is no )ro)er normal cell.
A )ro)er cell means*
+$e cell belons to t$e selected net#or?8
+$e cell is not barred8
+$e cell is not in t$e national )ro$ibited roamin location area8
+$e )at$ loss bet#een MS and B+S is under t$e limit set by net#or?.
+$e )riority level o! a cell is determined by C;GGPBA<PQ/AG>FR -CBQ. and
C;GGPBA<PACC;SS -CBA..
1.3.% Cell Selection @rocess
+o )er!orm cell selection and re,selection" MS requires all t$e !requencies
monitored to stay at t$e un#ei$ted averae value o! <elev <GAPC.
>. Cell Selection 9$en MS Storin Eo BCC' >n!ormation
MS searc$es all <F c$annels -at least 20 c$annels !or 900 M" 40 !or 1800 M" and
40 !or @SC1900. in t$e system to obtain t$e <elev o! eac$ <F c$annel" and
calculate t$e <GAPC based on at least !ive sam)les in t$ree to !ive seconds"
and t$en arrane t$ese levels in descendin order to select t$e )ro)er BCC'.
MS selects t$e cells #it$ normal )riority !irst. >! t$e )ro)er cells $ave lo#
)riority" MS #ill select t$e cell #it$ t$e $i$est <elev. MS $as already decoded
and identi!ied all t$ese !requencies by no#. >! t$ere is no )ro)er cell" MS #ill
?ee) on searc$in. >t ta?es a ma:imum o! 0.1 s to sync$roni(e a BCC' +<O and
1.9 s to read t$e sync$roni(ed BCC' +<O data" e:ce)t t$at it ta?es
nM1.9s-nS1.to obtain t$e system in!ormation.
>>. Cell Selection 9$en MS Storin BCC' >n!ormation
>! MS stores t$e BCC' !requency list o! t$e !ormer selected net#or?s" MS #ill
)er!orm measurement sam)lin )rocedure -only !or t$e stored BCC' +<O.
accordin to t$is list. >! t$e cell selection #it$in t$is list !ails" common cell
selection #ill be )er!ormed. >! all t$e cells $ave lo# )riority level" MS #ill
select t$e cell #it$ t$e $i$est <elev. MS $as already decoded and identi!ied
all t$ese !requencies by no#. 9$en a 900 M MS enters t$e 900=1800 net#or?"
MS #ill )robably c$oose 900 M net#or? and inore t$e )riority level" because
t$e MS stores all t$e 900 M !requency in!ormation in BCC' !requency list.
>>>. Cell Selection Criteria
@arameter C1 is t$e )at$ loss criteria !or cell selection" C1 o! t$e service cell
must e:ceed 0" t$e !ormula is as !ollo#s*
C1N <GAPC , <OG;HPACC;SSPM>E, MAO --MSP+O@9<PMAOPCC', @." 0. -%,1.
For DCS 1800 cells*
C1 N <GAPC , <OG;HPACC;SSPM>E, MAO --MSP+O@9<PMAOPCC' J @A9;<
AFFS;+, @." 0.
>n t$e !ormula*
<GAPC* Averae value o! <elev
<OG;HPACC;SSPM>E* Minimum <elev t$at MS allo#s
MSP+O@9<PMAOPCC'* Ma:imum transmit )o#er on control c$annel
@* Ma:imum transmit )o#er o! MS
@A9;< AFFS;+@o#er o!!set related to MSP+O@9<PMAOPCC' used by
DCS1800 cells.
1.3.2 Do#n Gin? Failure
Do#nlin? !ailure criteria are based on DSC. 9$en a mobile )$one stays in a
cell" DSC is initiali(ed to an inteer most close to 90=E - E is BSP@APMF<MS"
rane value* %09.. ;ac$ time #$en mobile )$one success!ully decodes a
messae on its )ain subc$annel" DSC increases by 1" but DSC cannot e:ceed
t$e initial value8 #$en decodin !ails" DSC decreases by 4. 9$en DSCTN0"
do#nlin? !ailure occurs. Do#n sinalin lin? !ailure #ill lead to cell re,
selection.
1.3.4 Cell <e,Selection @rocess
>n cell re,selection" mobile )$one #ill sync$roni(e and read t$e in!ormation
!rom si: BCC' +<Os -in BA list. #it$ stronest sinals outside t$e service area.
For multi,!requency mobile )$ones" t$e +<Os #it$ stronest sinals may be in
di!!erent !requency bands.
>n idle mode" mobile )$one monitors all t$e BCC' +<Os in BA list and averaes
eac$ <elev !rom BCC' +<O #it$in 1 s to Ma: U1" --1 M E J 3. D>H 5. M
BSP@APMF<MS = 4V s. E is t$e number o! BCC' +<Os outside service area in BA
list. ;ac$ <GAPC requires at least !ive level measurement sam)les and $as to be
u)dated !rom time to time. Service area sam)les t$e <elev at least once !or
eac$ )ain bloc? to mobile. <GAPC is calculated by averain t$e level
sam)les received !rom 1s to Ma: U1s" !ive consecutive )ain bloc?s o! t$at
MSV.
;ac$ <GAPC u)date is !ollo#ed by t$e u)date o! t$e si: BCC' +<Os outside t$e
service area in BA list. And t$e latter u)date may be even !aster.
Mobile )$one decodes all t$e BCC' data in a service cell every ot$er 20 s and
t$e BCC' data bloc?s related to cell re,selection )arameters o! t$e si: BCC'
+<Os #it$ stronest sinals every ot$er !ive minutes. 9$en t$e mobile )$one
detects t$at a ne# BCC' +<O becomes one o! t$e si: +<Os #it$ stronest
sinals" t$is BCC' +<O data s$ould be decoded #it$in 20 s. Mobile )$one
c$ec?s t$e BS>Cs o! t$e si: BCC' +<Os #it$ stronest sinals to ma?e sure t$ey
are in t$e same cell. >! t$e BS>C o! a +<O is c$aned" t$e MS #ill reard t$e +<O
as ne# +<O and reread t$e BCC' data.
MS #ill re,select a nei$bor cell as service cell under certain condition. +$is
condition includes several !actors" suc$ as <GAPC" cell restriction -decided by
cellPbar and cellPbarPquali!y." and access state o! t$e nei$bor cell.
Cell re,selection ado)ts C% alorit$m. +$e calculation !ormula is as !ollo#s*
9$en @;EAG+R +>M; is not 11111
C%NC1JC;GGP<;S;G;C+PAFFS;+0+;M@A<A<RPAFFS;+M' -@;EAG+RP+>M;0+.8
9$en @;EAG+RP+>M; is 11111
C%NC1,C;GGP<;S;G;C+PAFFS;+.
9$en OS0" !unction '-:. N08 #$en OIA" !unction '-:. N1.
+ is a timer8 its initial value is 0. 9$en a cell is included in t$e si: nei$bor
cells #it$ stronest sinals by MS" t$e timer + o! t$is cell beins to time8 #$en
a cell is e:cluded !rom t$e si: nei$bor cells #it$ stronest sinals by MS" +
#ill be reset.
C;GGP<;S;G;C+PAFFS;+ adFusts t$e value o! C%.
A!ter + starts" +;M@A<A<RPAFFS;+ #ill modi!y t$e C% alorit$m accordin to
t$e de!ined value be!ore t$e )enalty time in order to avoid a micro cell or a
cell #it$ small coverae area is selected by a !ast movin MS. >! t$e de!ined
)enalty time is out" t$e tem)orary o!!set #ill be inored. @enalty time can
avoid t$e !requent cell re,selection in t$ose coverae areas li?e e:)ress
$i$#ay.
+$ese )arameters in C% alorit$m #or?s only #$en
C;GGP<;S;G;C+>AEP>ED>CA+>AE is activated. At$er#ise" MS #ill inore t$e
settin o! C;GGP<;S;G;C+PAFFS;+" +;M@A<A<RPAFFS;+" and @;EAG+RP+>M;"
under suc$ circumstances" C%NC1.
Cell re,selection #ill be triered under t$e !ollo#in conditions*
+$e C% value o! a certain cell -belonin to t$e same location area #it$ t$e
current cell. e:ceeds t$at o! t$e current cell by 1 seconds successively8
+$e C% value o! a certain cell -belonin to di!!erent location area !rom t$e
current cell. e:ceeds t$e sum o! t$e C% value o! t$e current service cell and
cell selection $ysteresis value by 1 seconds successively8
+$e current service cell is barred8
MS detects do#nlin? !ailure8
+$e C1 value o! t$e service cell is less t$an 0 !or 1 seconds successively.
1.5 Frequency 'o))in
9it$ t$e ever ro#in tra!!ic volume and t$e limited !requency resource"
!requency reuse is more and more aressive. +$ere!ore" t$e )roblem o! $o#
to reduce !requency inter!erence becomes more and more remar?able. +$e
essence o! anti,inter!erence is to !ully utili(e t$e current s)ectrum" time
domain" and s)ace resources. +$e ?ey measures include !requency $o))in"
discontinuous transmission -D+O." and )o#er control. Frequency $o))in also
can e!!ectively reduce t$e in!luence o! !ast !adin.
1.5.1 +y)es o! Frequency 'o))in
GSM radio inter!ace uses slo# !requency $o))in -SF'. tec$noloy. +$e
di!!erence bet#een slo# !requency $o))in and !ast !requency $o))in is t$at
t$e !requency o! latter c$anes !aster t$an !requency modulation. >n GSM" t$e
!requency remains t$e same durin burst transmission. +$ere!ore" GSM
!requency $o))in belons to slo# !requency $o))in.
>n !requency $o))in" t$e carrier !requency is controlled by a sequence and
$o)s #it$ time. +$is sequence is !requency $o))in sequence. Frequency
$o))in sequence is a sequence o! !requencies decided by $o))in sequence
number -'SE." mobile allocation inde: o!!set -MA>A. and !rame number -FE.
t$rou$ a certain alorit$m in t$e mobile allocation containin E !requencies.
+$e E c$annels o! di!!erent timeslots can use t$e same $o))in sequence. +$e
di!!erent c$annels o! t$e same timeslot in t$e same cell ado)t di!!erent MA>A.
Frequency $o))in can be divided into !rame $o))in and timeslot $o))in
accordin to time domain and <F $o)in and baseband $o))in accordin to
im)lementation mode.
Frame $o))in* t$e $o))in !requency c$anes once in eac$ +DMA !rame
)eriod. ;ac$ +<O can be rearded as a c$annel. +$e +C' o! BCC' +<O cannot
Foin in t$e !requency $o))in in a cell. +$e $o))in +<O s$ould $ave a
di!!erent MA>A. Frame $o))in is an e:ce)tion o! timeslot $o))in.
+imeslot $o))in* t$e timeslot !requency o! eac$ +DMA !rame c$anes once.
+$e +C' o! BCC' +<O can Foin in t$e !requency $o))in" #$ic$ $a))ens in
baseband $o))in.
<F $o))in* bot$ transmission and rece)tion o! +<O Foin in t$e !requency
$o))in. +$e number $o))in !requencies can e:ceed t$e number o! +<Os in
t$e cell.
Baseband $o))in* eac$ transceiver #or?s at a !i:ed !requency. +O does not
Foin in !requency $o))in. Frequency $o))in is )er!ormed t$rou$ t$e
$andover o! banseband sinal. +$ere!ore" t$e number o! $o))in !requencies
cannot e:ceed t$e number o! +<Os in t$e cell.
+$e t#o !requency $o))in modes above are based on B+S. As !or MS" since
eac$ MS $as only one +<O unit" <F $o))in is t$e only mode.
>. Baseband 'o))in
+$e system $as multi)le baseband and +<O )rocessin unit. ;ac$ +<O
)rocessin unit $as a !i:ed #or?in !requency8 eac$ baseband )rocessin unit
)rocesses one line o! service in!ormation and sends t$e )rocessed in!ormation
to t$e +<O unit #it$ bus to)oloy in time sequence accordin to !requency
$o))in rule. +$is ?ind o! !requency $o))in is called 6baseband $o))in7.
>n baseband $o))in" eac$ transceiver #or?s #it$ a !i:ed !requency. +$e bursts
on t$e same s)eec$ )at$ are sent to eac$ transceiver. Baseband $o))in is
based on t$e $andover o! baseband sinals. Since t$e transceiver o! eac$ B+S
$as a !i:ed #or?in !requency" bot$ broadband combiner and cavity combiner
can be ado)ted. +$e number o! +<Os decides t$e ma:imum number o!
!requency $o))in. +$e )roblem !or baseband $o))in is t$at i! one +<O board
!ails" t$e corres)ondin code #ord #ill be lost" t$us a!!ectin all t$e calls
under $o))in mode in t$e cell.
>>. <F 'o))in
/nder t$is mode" eac$ line o! service in!ormation is )rocessed by !i:ed
baseband unit and !requency band unit. +$e #or?in !requency o! !requency
band unit is )rovided by !requency combiner. /nder t$e control o! control unit"
!requency can be c$aned accordin to certain rules. >n <F $o))in" t$e
!requencies used by a +<O to $andle all t$e bursts o! a call come !rom t$e
!requency c$ane o! combiner" instead o! t$e $andover o! baseband sinals.
+$e number o! +<Os is not limited by carrier !requency. As t$e #or?in
!requency o! +<O c$anes" #$ic$ means t$e !requency o! t$e in)ut )ort to
combiner c$anes" only broadband combiner can be ado)ted. +$is ?ind o!
broadband combiner leads to about 2dB insertion loss in t#o,in,one
combination and t$e loss is reater in t$e lin? insertion o! multi,combiner. GSM
)rotocol does not s)eci!y #$ic$ ?ind o! !requency $o))in is used in GSM B+S.
+$e mode o! !requency $o))in can be decided by o)erators accordin to t$e
equi)ments.
1.5.% Frequency 'o))in Alorit$m
+$e )arameters related to !requency $o))in alorit$m are as !ollo#s*
CA* cell allocation" t$e collection o! !requencies used by a cell
FE* +DMA !rame number" broadcasted on sync c$annel. FE -00%511345.
sync$roni(es B+S #it$ MS
MA* mobile allocation" t$e collection o! radio !requencies used !or MS
!requency $o))in. >t is a subset o! CA. MA contains E !requencies" 1IEI34.
MA>A* mobile allocation inde: o!!set" -00E,1.. Durin communication" t$e radio
!requency at air inter!ace is an element o! MA. Mobile allocation inde: -MA>" 00
E,1. is used to determine t$e element o! MA. +$at is to say" t$e actual
!requency used is decided by MA>. MA>A is t$e initial o!!set o! MA> and it is used
to avoid t$e contention o! !requency by several c$annels at t$e same time.
'SE* $o))in sequence number -0032.. >t determines t$at t$e $o))in
sequence #it$ concentrated !requencies is ado)ted in !requency $o))in.
9$en 'SEN0" t$e $o))in is cyclic $o))in8 #$en 'SEW0" t$e $o))in is
random $o))in.
+$e )ro)er settin o! )arameters is based on t$e understandin o! t$e use o!
eac$ )arameter in $o))in alorit$m and t$e $o))in t$eory. +$e )ro)er
settin ensures t$e $ealt$y #or?in state o! t$e system.
<emar?s* For t$e cyclic $o))in in discontinuous transmission -D+O." t$e
number o! $o))in !requencies s$ould avoid E mod 12 N 0" because under suc$
condition" t$e )robability o! transmission and measurement o! SACC' !rame at
t$e same !requency is rat$er $i$" and t$e $arms are obvious.
9$en 'SEN0" S equals t$e !rame number" in ot$er cases" S is only related to
!rame number and !requency $o))in number. 9$en 'SE is !i:ed and !rame
number is t$e same" S must be t$e same. +$ere!ore" as t$e +<Os o! eac$ sync
cell $ave t$e same !rame number" di!!erent $o))in rou)s in sync cells can
ado)t t$e same 'SE. A )ro)er con!iuration o! MA>A can avoid t$e inter,cell or
intra,cell !requency collision #it$in t$e same B+S. +$e aressive !requency
reuse ado)ts t$is t$eory.
1.5.2 Bene!its o! Frequency 'o))in
>n GSM" !requency $o))in $as t#o bene!its* !requency diversity and
inter!erence averain.
>. Frequency Diversity
Frequency $o))in can reduce t$e in!luence o! sinal strent$ c$ane due to
multi)at$ transmission. +$is e!!ect equals t$at o! !requency diversity. >n
mobile communications" <aylei$ !adin leads to t$e reat c$ane o! radio
sinal in a s$ort time. +$is ?ind o! c$ane is related to !requency* a more
inde)endent !adin accom)anies a reater !requency di!!erence. +$e %00 &'(
interval enerally ensures t$e inde)endence o! inter,!requency !adin" #$ile
t$e 1 M'( interval can !ully uarantee t$is ?ind o! inde)endence. +$rou$
!requency $o))in" all t$e bursts containin t$e code #ord o! t$e same s)eec$
!rame are )rotected !rom t$e damae o! <aylei$ !adin in t$e same #ay.
Statistics s$o#s t$at !requency $o))in ain is related to environmental
!actors" es)ecially to t$e movin s)eed o! MS. 9$en t$e MS moves at a $i$
s)eed" t$e location di!!erence bet#een t#o bursts on t$e same c$annel is also
a!!ected by ot$er ?inds o! !adin. +$e $i$er t$e s)eed is" t$e lo#er t$e ain
#ill be. Frequency diversity bene!its a lot to a lare number o! MSs movin at
lo# s)eed.
Frequency $o))in ain is also related to t$e number o! !requencies. 9$en t$e
number o! !requencies decreases" t$e $o))in ain !alls. +$e relations$i)
bet#een t$e number o! !requencies and $o))in ain can be e:)lained in t$is
#ay* !requency $o))in is )seudo s)ectrum s)read" and t$e $o))in ain is t$e
)rocessin ain a!ter transmission !requency band s)read. +$e basic #ay to
test !requency $o))in ain is to calculate t$e di!!erences bet#een di!!erent
C=> at di!!erent $o))in !requencies under t$e same F;<. +$ese C=> di!!erences
are t$e !requency $o))in ain.
>>. >nter!erence Averain
Frequency $o))in )rovides t$e diversity o! inter!erence on transmission
c$annel" so t$at all t$e bursts containin t$e code #ord o! t$e same s)eec$
!rame are )rotected !rom t$e damae o! inter!erence in t$e same #ay.
+$rou$ error correction codin and interleavin o! t$e system" t$e oriinal
data can be restored !rom t$e rest )art o! t$e received !lo#. +$e $o))in ain
is obtained only #$en t$e inter!erence is in narro#band distribution. >! t$e
inter!erence is in broadband distribution" all t$e bursts #ill be destroyed and
t$e oriinal data cannot be restored. +$ere!ore" no $o))in ain is obtained.
+$e common inter!erence a!ter !requency $o))in can be rearded in
narro#band distribution.
>n !requency $o))in" error rate tends to increase in t$e test" but #e !eel t$e
conversation quality im)roves. >t is because alt$ou$ t$e error rate increases"
t$e in!luence o! inter!erence is $omoeni(ed in !requency $o))in" t$e s)eec$
restorin ability im)roves because o! t$e interleavin and de,interleavin
be!ore. >n G@<S data services" !requency $o))in can be $arm!ul #$en t$e
data rate is rat$er $i$ -CS4..
1.9 @o#er Control
1.9.1 @o#er Control Avervie#
@o#er control is to c$ane t$e transmission )o#er o! MS or B+S -or bot$. in
radio mode #it$in certain area. @o#er control can reduce t$e system
inter!erence and im)rove t$e s)ectrum utili(ation and )rolon t$e service time
o! MS battery. 9$en t$e <elev and quality is ood" t$e transmission )o#er o!
t$e )eer end can be reduced to lo#er t$e inter!erence to ot$er calls.
>n GSM" )o#er control can be used in u)lin? and do#nlin? res)ectively. +$e
)o#er control rane !or u)lin? MS is %0 dB020dB. Based on t$e )o#er class o!
MS -most MSs belons to class 4" #$ic$ means t$e ma:imum transmission )o#er
is 22 dbm." eac$ ste) can c$ane % dB. +$e do#nlin? )o#er control rane is
decided by equi)ment manu!acturer. Alt$ou$ #$et$er to ado)t u)lin? or
do#nlin? )o#er control !unction is decided by net#or? o)erators" all MSs and
B+S equi)ments must su))ort t$is !unction. BSS manaes t$e )o#er control in
t$e t#o directions.
+o !acilitate BCC' !requency )ull,in and t$e measurement o! <elev -includin
t$e <elev o! nei$bor cell BCC' !requency." GSM )rotocol s)eci!ies t$at no
)o#er control is allo#ed !or t$e timeslots in t$e do#nlin? o! BCC' +<O.
1.9.% MS @o#er Control
+$e )o#er control o! MS includes t#o adFustment staes* stable adFustment
stae and initial adFustment stae. Stable adFustment is t$e common #ay to
im)lement )o#er control alorit$m. >nitial adFustment is used at t$e beinnin
o! call connection. 9$en a connection occurs" MS sends sinals #it$ nominal
)o#er -be!ore receivin )o#er adFustment commend" t$e nominal transmission
)o#er o! MS is t$e ma:imum transmission )o#er on BCC' o! t$e cell. >! MS
does not su))ort t$is )o#er level" it #ill ado)t ot$er )o#er level most close to
t$is level" suc$ as t$e ma:imum )o#er level su))orted by t$e classmar? o! MS
in indication messae establis$ment.. +$ere!ore" MS accesses to net#or?
t$rou$ <AC' #it$ t$e ma:imum )o#er broadcast on BCC'. 9$en MS )o#er is
lo#er t$an t$is value" it #ill transmit #it$ its ma:imum transmission )o#er.
+$e system s)eci!ies t$at t$e )o#er level o! t$e !irst messae t$at MS sends on
DC' is also t$is value. +$e system control beins a!ter MS receives t$e )o#er
control command in SACC' in!ormation bloc? !rom SDCC' or +C'.
Since B+S can su))ort multi,call at t$e same time" t$e <:lev s$ould be quic?ly
reduced in t$e ne# connection. At$er#ise" ot$er calls su))orted by t$is B+S
#ill deteriorate and t$e calls in ot$er cells #ill also be a!!ected. +$e )ur)ose
o! initial adFustment stae is to quic?ly reduce t$e transmission )o#er o! MS to
et t$e stable M<" so MS can be adFusted accordin to stable )o#er control
alorit$m.
+$e required )arameters in u)lin? )o#er control" t$e e:)ected u)lin? <:lev"
and t$e u)lin? received quality can be adFusted accordin to t$e situation o!
t$e cell. A!ter receivin a certain number o! u)lin? M<s" t$e system com)ares
t$e actual u)lin? <:lev and received quality obtained by inter)olation"
!ilterin" and ot$er met$ods #it$ t$e e:)ected values and calculate t$e )o#er
level t$at t$e MS s$ould be adFusted to t$rou$ )o#er control alorit$m. >! t$e
calculated )o#er level di!!ers !rom t$e out)ut )o#er level o! MS and meets
certain limit conditions -suc$ as ste) limit o! )o#er adFustment and rane limit
o! MS out)ut )o#er." t$e system #ill send )o#er adFustment command.
+$e command o! c$anin MS )o#er and t$e required time advance #ill be
sent to MS in t$e layer 1 $eader o! eac$ do#nlin? SACC' in!ormation bloc?. MS
#ill con!iure t$e )o#er level it uses no# in its u)lin? SACC' in!ormation bloc?
and send it to B+S in measurement re)ort. +$is level is t$e )o#er level o! t$e
last burst in t$e )revious SACC' measurement cycle. 9$en MS receives t$e
)o#er control in!ormation in SACC' in!ormation bloc? !rom DC'" it #ill
transmit #it$ t$is )o#er level. Ane )o#er control messae does not ma?e t$e
MS s#itc$ to t$e required level immediately. +$e ma:imum c$ane rate o! MS
)o#er is % dB !or every 30 ms. For 1% dB" be!ore MS receives t$e ne:t )o#er
control messae" it #ill not end as one SACC' measurement cycle ta?es 480
ms. >n addition" it ta?es t$ree measurement cycles to send )o#er control
messae and e:ecute t$e command. +$ere!ore" t$e )o#er control cycle s$ould
not be too s$ort in order to ensure its accuracy. See Fiure 1,10.
Fiure 1,1 ;:ecution o! )o#er control command
+$e )ur)ose o! u)lin? )o#er control adFustment is to minimi(e t$e di!!erence
bet#een t$e actual u)lin? <:lev and received quality and t$e e:)ected u)lin?
<:lev and received quality. +$e )ur)ose o! inter)olation and !ilterin is to
)rocess t$e lost measurement re)orts and remove tem)orary nature to ensure
t$e stability o! )o#er control alorit$m.
+$e di!!erence bet#een initial adFustment and stable adFustment is t$at t$e
e:)ected u)lin? <elev and received quality and t$e lent$ o! !ilter in initial
adFustment are di!!erent !rom t$at o! stable adFustment" and t$e initial
adFustment only $as do#nlin? adFustment.
1.9.2 B+S @o#er Control
B+S )o#er control is an o)tional !unction. >t is similar to MS )o#er control" but
it only uses stable )o#er control alorit$m. +$e required )arameters are <:lev
t$res$old -lo#er limit." and t$e ma:imum transmission level can be received
-u))er limit.. +$e <elev is divided into 34 levels ranin !rom 0 to 32. Gevel 0 is
t$e lo#est <:lev8 level 32 is t$e $i$est <:lev.
B+S )o#er control is divided into static )o#er control and dynamic )o#er
control. Dynamic )o#er control is t$e !ine tunin based on static )o#er
control. +$ere are si: ste)s -% dB=ste). o! static )o#er control accordin to
@rotocol 0101. >! t$e ma:imum out)ut )o#er is 43 dBm -409." t$e ste) 3 is 24
dBm.
Static )o#er control ste) is de!ined in t$e cell distributes list o! data
manaement system" #$ic$ s)eci!ies t$e ma:imum out)ut )o#er -su))ose t$is
value is @n. o! static )o#er control. For ste) 11 o! dynamic )o#er control" t$e
corres)ondin value rane is @n dB0@n,20dB. 9$en t$e ma:imum )o#er control
still cannot satis!y t$e requirement" adFust static )o#er control ste) to
im)rove t$e ma:imum out)ut )o#er o! dynamic )o#er control @n.
1.9.4 @o#er Control @rocessin
>. Measurement <e)ort >nter)olation
;ac$ measurement re)ort $as a sequence number. >! net#or? detects
incontinuous sequence numbers" it means some o! t$e measurement re)orts
are missin. +$e net#or? #ill com)lete t$e re)orts based on inter)olation
alorit$m.
+$e net#or? receives measurement re)orts n and nJ4. >t detects t$e sequence
numbers are not continuous" so it uses an alorit$m to add nJ1" nJ%" and nJ2
to com)lete t$e re)orts.
+$e )ur)ose o! measurement re)ort inter)olation is to avoid call loss #$en t$e
)o#er is too lo#.
>>. Measurement <e)ort Filterin
Eet#or? #ill not Fude t$e state o! MS based on only one measurement result"
because t$at is too incom)re$ensive" in addition" t$e MS may be !luctuatin.
+$ere!ore" !ilterin is required. Filterin combines several continuous
measurement results toet$er to determine t$e state o! MS durin t$is )eriod
o! time. +A $as !ilters !or <:lev and received quality o! u)lin? and do#nlin?
+$e )ur)ose o! measurement re)ort !ilterin is to remove tem)orary nature
and ensure t$e alorit$m stability.
>>>. @o#er Control AdFustment
Calculate t$e )o#er adFustment value based on t$e di!!erence bet#een t$e
<:lev and t$e e:)ected value.
@o#er control adFustment based on <:lev
@o#er control module com)ares t$e estimate value o! <:lev obtained t$rou$
)re,)rocessin o! measurement re)ort #it$ t$e e:)ected value" and calculates
t$e ste) lent$ o! adFustment. >n )o#er control alorit$m" variable ste) is
o!ten used !or quic? )o#er control.
@o#er control adFustment based on received quality
@o#er control module com)ares t$e estimate value o! received quality
obtained t$rou$ )re,)rocessin o! measurement re)ort #it$ t$e e:)ected
value" and calculates t$e ste) lent$ o! adFustment. 9$en t$e received quality
is bad" im)rove t$e transmit )o#er8 #$en t$e received quality is ood" reduce
t$e transmit )o#er. +$is ?ind o! )o#er control ado)ts !i:ed ste).
Com)re$ensive decision !or )o#er control
Consider bot$ <:lev and received quality and ado)t di!!erent )o#er control
strateies in di!!erent conditions to ?ee) t$e stability and e!!iciency o! )o#er
control alorit$m.
9$en t$e received quality requires t$e im)rovin o! transmit )o#er #$ile t$e
<:lev requires t$e reducin o! it" t$e system #ill ma?e a com)re$ensive
decision to )er!orm no )o#er control adFustment" because bad received quality
and ood <:lev re)resent stron net#or? inter!erence. /nder suc$
circumstances" im)rovin transmit )o#er #ill !urt$er increase t$e
inter!erence.
1.10 >mmediate Assinment @rocedure
+$e )ur)ose o! immediate assinment is to establis$ a radio connection -<<
connection. bet#een MS and system at /m inter!ace.
1.10.1 Eet#or? Access Gicense and <andom Access <equest
+$e request o! MS !or c$annel assinment is controlled by its o#n access level
and t$e access rant level broadcast in cell. ;ac$ MS $as one access level o!
t$e ten levels !rom 0 to 9. >n addition" it may also $ave one or several levels o!
t$e !ive s)ecial access levels !rom l1 to 11. Access level is stored in S>M card.
BCC' system in!ormation broadcasts access levels and s)ecial access levels
t$at t$e net#or? rants and t$e in!ormation t$at #$et$er all MSs allo#
emerency call or allo# s)ecial access levels only. >! t$e mobile oriinated call
is not emerency call" t$e MS can access to net#or? only #$en it belons to
t$e ranted access level or ranted s)ecial access level. >! t$e mobile
oriinated call is emerency call" t$e MS can access to net#or? only #$en all
t$e MSs in t$e cell allo# emerency call or it belons to t$e ranted s)ecial
access level.
9$en an MS #ants to establis$ connection #it$ t$e net#or?" it sends a c$annel
request to net#or? t$rou$ <AC' c$annel. C$annel request in!ormation
contains 8,bit use!ul sinalin in!ormation" amon #$ic$ 2 bits03 bits are used
as t$e minimal indicator o! access cause. +$e system )rocesses di!!erent
c$annel requests based on t$is rou$ indication. >t di!!erentiates t$e ranted
calls !rom t$e denied calls and assins )ro)er c$annels !or t$e ranted calls.
+$is ?ind o! )rocess is es)ecially use!ul #$en t$e net#or? is overload and t$e
!lo# control is required. Since t$e c$annel ca)acity is limited" t$is indicator
cannot trans!er all t$e in!ormation !rom MS" suc$ as t$e detailed cause o!
c$annel request" user identity and t$e !eatures o! mobile equi)ment. +$ese
?inds o! in!ormation are sent in t$e !ollo#in SABM messaes. +$e 8,bit
in!ormation also contains t$e random discriminator sent by t$e MS and t$e
immediate assinment command -it contains in!ormation about t$e assined
c$annel.. >mmediate assinment command carries t$e discriminator sent by t$e
)revious MS. MS com)ares t$is discriminator #it$ its o#n discriminator and
Fudes #$et$er it is t$e messae !or itsel! !rom net#or?. Since t$ere are at
most 1 bits in t$e 8 bits in!ormation carryin discriminator" only 2% MSs can be
di!!erentiated at t$e same time. Furt$er discrimination o! t$e MSs requires t$e
res)onse in!ormation at /m inter!ace. C$annel request in!ormation belons to
internal in!ormation o! BSS.
>n GSM" <AC' is a ?ind o! AGA'. >n order to reduce t$e collision on <AC' durin
MS access to net#or? and im)rove t$e e!!iciency o! <AC' c$annel and MS
access. GSM s)eci!ies t$e required access alorit$m !or MS. +$is ?ind o!
alorit$m de!ines t$ree )arameters* +:Pinterer +" t$e ma:imum
retransmission times <;+" and )arameter S related to + and c$annel
combination.
+ re)resents t$e number o! timeslots bet#een t#o transmissions #$en
continuous c$annel requests are sent. S is an intermediate variable de)ends on
+ and t$e con!iuration o! CCC'. See t$e descri)tion o! t$is )arameter in
C$a)ter 5. <;+ is t$e MS ma:imum retransmission times allo#ed in order to
avoid access collision. ;ac$ time a!ter MS sends access request" +21%0 is to
receive -or reFect. immediate assinment messae. MS #ill retransmit access
request !or t$e messaes t$at are not received or reFected #$en +21%0 times
out under t$e )remise t$at <;+ is not e:ceeded and restart t$e +21%0. 9$en
t$e retransmission times reac$es <;+ and +21%0 times out" +21%3 #ill be
started to receive -or reFect. immediate assinment messae. 9$en +21%3
times out" cell re,selection #ill be initiated.
1.10.% >nitial >mmediate Assinment
A!ter decodin t$e c$annel request in!ormation" B+S sends a c$annel required
messae to BSC. +$is messae contains im)ortant additional in!ormation and
t$e estimation o! +A by B+S. A!ter receivin t$is messae" BSC selects a )ro)er
c$annel !or t$is request and activates t$e land resources by sendin a c$annel
active messae to B+S. B+S returns a c$annel active ac?no#lede messae to
BSC. >! BSC receives t$is messae" B+S #ill send an immediate assinment
command or immediate assinment e:tended messae on CCC'. >n order to
im)rove c$annel e!!iciency" GSM introduces t$e messae layout o! immediate
assinment e:tended t$at contains t$e assinment in!ormation o! t#o MSs. +$e
immediate assinment messae contains t$e assinment in!ormation o! one MS.
Accordin to GSM s)eci!ications" MS must identity t$e immediate assinment
-e:tended. in!ormation !or t$e last t$ree c$annel requests.
>! t$ere is no c$annel to activate" BSC #ill send an immediate assinment
reFect or immediate assinment e:tended reFect messae to MS. A!ter
receivin t$e reFect messae" MS sto)s +21%0 based on one o! t$e last t$ree
c$annel requests and starts +21%%. Durin t$e s)eci!ied time o! +21%%" MS $as
no access to net#or? and turns into idle mode. Be!ore +21%% times out" MS
cannot initiate connection attem)t e:ce)t emerency call #it$in t$e same
cell.
A!ter receivin immediate assinment messae" MS com)ares t$e received
assinment command #it$ t$e in!ormation stored in its c$annel request and
Fudes #$et$er t$is messae is !or itsel!. >! t$is messae matc$es one o! its last
t$ree c$annel requests" MS #ill sto) +21%0 or +21%3 and s#itc$ to t$e assined
c$annel. +$en it starts to establis$ t$e sinalin lin? by usin Set Async$ronous
Balanced Mode -SABM. command.
1.10.2 >nitial Messae
A!ter receivin immediate assinment messae and decodin it" MS adFusts its
con!iuration o! transmission and rece)tion to t$e assined c$annel and
transmits sinalin accordin to t$e +A value s)eci!ied by BSS and t$e initial
ma:imum transmission )o#er broadcast in BCC' system in!ormation -see t$e
descri)tion o! ms+:@#rMa:CC'.. MS sends an SABM !rame on assined
SDCC'=+C' to establis$ t$e async$ronous balanced mode -SA@>N0. t$at is used
to establis$ sinalin messae lin? layer connection under ac?no#ledement
mode. Accordin to GSM )rotocol" SABM carries an initial messae t$at contains
layer 2 service request in!ormation.
9$en t#o MSs send t$e same c$annel requests -#$ic$ is )ossible in $i$ tra!!ic
volume area." t$e t#o MSs may res)ond to t$e same dedicated c$annel. in
order to save t$is )roblem" a!ter receivin SABM !rame" B+S ma?es no
modi!ication but sends a /A !rame -no !rame number ac?no#ledement.
containin t$e same in!ormation as t$at o! initial messae. >! t$e in!ormation
o! /A !rame is di!!erent !rom t$at o! SABM !rame" MS #ill abandon t$is c$annel
and start reaccess )rocess. Anly t$e ri$t MS can stay on t$is c$annel.
SABM !rame carries !our ?inds o! initial messaes* CM service request -suc$ as
call setu)" s$ort messae" and su))lementary service." location u)datin
request -eneric location u)datin" )eriodic location u)datin" and >MS>
attac$." >MS> detac$" and )ain res)onse. All t$ese messaes contain t$e
identity o! MS" detailed access cause" and MS classmar? -indicatin some ?ey
!eatures suc$ as transmission )o#er level" encry)tion alorit$m" s$ort messae
ca)acity" and !requency ca)acity..
A!ter receivin t$e initial messae" B+S sends an establis$ indication messae
to BSC. BSC receives t$is messae and sends com)lete layer 2 in!ormation to
MSC to request SCC@ connection to MSC. Gayer 2 in!ormation carries t$e causes
!or CM service request" #$ic$ includes mobile oriinated call" emerency call"
location u)datin" and s$ort messae service. +$is in!ormation also carries
ci)$er ?ey sequence number" MS identi!ication number" and some )$ysical
in!ormation o! t$e MS suc$ as transmit )o#er level" ci)$erin alorit$m"
)seudo,sync$roni(ation" and s$ort messae. A!ter receivin t$is in!ormation"
MSC sends connection con!irmed messae to BSC -i! t$e connection cannot be
establis$ed" MSC #ill send SCC@ re!used messae. to indicate t$at t$e sinalin
lin? bet#een MS and MSC $as been establis$ed. By t$is time" MSC can control
t$e transmission )ro)erties o! << manaement8 BSS monitors t$e transmission
quality and )re)ares !or $andover. +$en t$e MM connection beins.
Aut$entication or encry)tion is triered #$en required in t$e !ollo#in
)rocessin.
>n t$e immediate assinment )rocess" +2101 starts #$en BSC sends c$annel
active messae to B+S and ends #$en t$e establis$ indication is received. >!
+2101 times out be!ore sinalin c$annel is establis$ed" t$e activated c$annel
#ill be released.
1.10.4 >mmediate Assinment Failure
>! a !ailure occurs to t$e underlayin MS on t$e ne# c$annel be!ore t$e
establis$ment o! sinalin lin?" t$e net#or? releases t$e assined c$annel o!
MS. +$e !ollo#in )rocessin de)ends on t$e !ailure ty)e and )revious actions.
>! t$e !ailure is caused by t$e mismatc$ o! messae !ield in decision contention
and no re,assinment is initiated" t$e immediate assinment is restarted.
>! t$e !ailure is caused by ot$er reasons or i! t$e re,assinment triered by t$e
mismatc$ o! messae !ield in decision contention is carried out and t$e
assinment still !ails" MS turns into idle mode and triers cell re,selection.
>! t$e available in!ormation is not su!!icient to de!ine a c$annel a!ter t$e MS
receives immediate assinment messae" << connection !ails.
>! t$e assined !requencies o! MS belon to t#o or more t$an t#o !requency
bands" << connection !ails. >! t$e assined !requency o! MS is not consistent
#it$ t$e requested !requency but su))orted by MS" MS accesses t$e c$annel
#it$ t$e !requency used in c$annel request. >! MS does not su))ort t$e
assined !requency" << connection !ails.
>! +2101 times out be!ore t$e sinalin c$annel is establis$ed" net#or? releases
t$e assined c$annel. Eet#or? cannot tell #$et$er MS resends t$e access
attem)t or not.
1.11 Aut$entication and ;ncry)tion
GSM ta?es lots o! measures to )rotect t$e sa!ety o! system" suc$ as usin
+em)orary Mobile Subscriber >dentity -+MS>. to )rotect >MS>" usin @ersonal
>denti!ication Eumber -@>E. to )rotect S>M card" aut$entication t$rou$
aut$entication center -A/C. !or net#or? access" encry)tion" and equi)ment
identity reister.
Aut$entication and encry)tion require a rou) o! t$ree )arameters t$at
enerated in A/C. ;ac$ client is assined a Mobile Station >nternational >SDE
Eumber -MS>SDE. and >MS> #$en reisters in GSM net#or?. >MS> is )reserved
onto S>M card t$rou$ S>M )rinter and S>M )rinter #ill enerate a corres)ondin
client aut$entication value &i t$at is stored in S>M card and A/C as )ermanent
in!ormation. A/C $as a )seudo number enerator used to enerate a random
number <AED. GSM de!ines alorit$m A2" A8" and A1 t$at are used !or
aut$entication and encry)tion. >n A/C" <AED and &i toet$er )roduce a
res)onse number S<;S t$rou$ A2 aut$entication alorit$m and a &c t$rou$
A8 encry)tion alorit$m. <AED" &c" and S<;S !orm a t$ree,)arameter rou) o!
client. +$is rou) is stored in t$e data base o! t$is client in 'G<. Generally"
A/C trans!ers !ive rou)s o! )arameters to 'G< !or automatic storae. 'G< can
save ten rou)s o! suc$ )arameters. 9$en MSC=HG< requests !or t$ree,
)arameter rou) trans!er" 'G< sends !ive rou)s at t$e same time !or MSC=HG<
to use one by one. 9$en t$ere are t#o rou)s le!t" MSC=HG< #ill request !or
trans!er aain.
1.11.1 Aut$entication
Aut$entication is t$e )rocess t$at GSM net#or? c$ec?s #$et$er t$e >MS> or
+MS> !rom MS at radio inter!ace is valid or not. +$e )ur)ose o! aut$entication is
to avoid unaut$ori(ed access to GSM net#or? and t$e t$e!t o! )rivate
in!ormation by illeal users. Aut$entication also )rovides )arameters !or MS to
calculate ne# encry)tion ?ey.
+$e net#or? initiates aut$entication )rocedure in t$e !ollo#in situations*
MS requestin !or t$e c$ane o! in!ormation in HG< or 'G<8
Service access" includin MS oriinated call" MS terminated call" MS activation
and deactivation" and su))lementary services8
+$e !irst net#or? access a!ter MSC=HG< reboot8
Mismatc$in Ci)$er ?ey Sequence8
9$et$er to initiate aut$entication )rocedure de)ends on i! t$e &c value o! t$e
last service )rocessin stored in net#or? consistent #it$ t$at o! t$e )resent
access stored in MS. >! consistent" aut$entication )rocedure can be esca)ed
and t$is &c value is used directly !or encry)tion8 i! not" &c value needs to be
recalculated. MS does not send &c value to net#or? t$rou$ radio )at$ !or t$e
sa?e o! )rivacy. +$ere!ore" Ci)$er &ey Sequence Eumber -C&SE. is introduced.
C&SE is sent to MS by MSC=HG< t$rou$ aut$entication request messae durin
t$e last net#or? access. >t is stored in bot$ S>M card and MSC=HG<. Durin t$e
initial access o! MS" C&SE is sent to MSC=HG< t$rou$ t$e initial request
messae o! SABM !rame. MSC=HG< com)ares it #it$ t$e last C&SE. >! t$ey are
not consistent" aut$entication is required be!ore encry)tion. >! C&SEN0" it
means no &c is assined. Aut$entication )rocedure is initiates and controls by
net#or?. MSC=HG< sends an aut$entication request messae to MS to initiate
aut$entication )rocedure and +2%30.
>. Aut$entication Success
%. A/+';E+>CA+>AE <;Q/;S+ contains a <AED -1%8 bits. and a C&SE. +$e &i
and <AED toet$er enerate a S;<S -2% bits. t$rou$ alorit$m A2 and a &c -34
bits. t$rou$ alorit$m A8. +$e ne# &c re)laces t$e !ormer ?ey and is stored in
S>M card toet$er #it$ C&SE.
2. MS sends A/+';E+>CA+>AE <;S@AES; to net#or?. A!ter receivin t$is
messae" t$e net#or? sto)s +2%30 and c$ec?s its validity -net#or? com)ares it
#it$ t$e S;<S enerated by &i and <AED t$rou$ alorit$m A2 and c$ec?
#$et$er t$ey are consistent or not." and t$en enters t$e subsequent
)rocedures" suc$ as encry)tion.
>>. Aut$entication <eFect
>! aut$entication !ails" it means A/+';E+>CA+>AE <;S@AES; is invalid.
>! t$e MS uses +MS>" t$e net#or? #ill initiate identity )rocedure. >! t$e >MS>
)rovided by t$e MS is di!!erent !rom t$at in net#or?" t$e net#or? #ill restart
t$e aut$entication )rocedure8 i! t$e >MS> is correct" t$e net#or? #ill send
A/+';E+>CA+>AE <;X;C+ to t$e MS.
>! t$e MS uses >MS>" t$e net#or? #ill send A/+';E+>CA+>AE <;X;C+ directly to
MS. A!ter sendin A/+';E+>CA+>AE <;X;C+ messae" t$e net#or? releases all
t$e MM connections under establis$ment and restarts t$e )rocedure !or <<
connection release.
A!ter receivin A/+';E+>CA+>AE <;X;C+ messae" MS sets t$e roamin
disabled !la and deletes in!ormation suc$ as +MS>" GA>" and ci)$er ?ey.
>! MS receives A/+';E+>CA+>AE <;X;C+ messae in >MS> D;+AC' >E>+>A+;D
state" it sto)s +2%%0 a!ter << connection is released. >! )ossible" MS initiates
local release )rocedure a!ter t$e normal release )rocedure or +2%%0 timeout8
i! not -suc$ as t$e >MS> detac$ a!ter s#itc$ o!!." MS<< e:its abnormally.
>! MS receives A/+';E+>CA+>AE <;X;C+ messae in ot$er state" it e:its all MM
connections and call re,establis$ment )rocedures" sto)s +2%10 and +2%20" sets
and starts +2%40 to enter 9A>+ FA< E;+9A<& CAMMAED state and #ait !or t$e
release o! << connection8 >! << connection is not released a!ter +2%40 timeout"
MS #ill e:it << connection abnormally. /nder t$e t#o conditions above" MS
enters MM >DG; and EA >MS> state.
1.11.% ;ncry)tion
;ncry)tion occurs in service requests suc$ as location u)datin" service access"
and inter,o!!ice $andover. >t requires t$e su))ort o! GSM net#or? equi)ment
-es)ecially B+S." as #ell as t$e encry)tion ability o! MS.
>. Sinalin @rocedure
1. MSC sends BSC a Ci)$erin Mode CMD t$at contains encry)tion alorit$m" &c"
and #$et$er t$e MS is required to add >M;> in Ci)$erin Mode CM@.
%. BSC decides t$e !inal alorit$m based on t$e encry)tion alorit$m in
Ci)$erin Mode CMD" t$e encry)tion alorit$m t$at BSC allo#s" and t$e
encry)tion alorit$m t$at MS su))orts" and t$en in!orm B+S.
2. BSC sends MS Ci)$erin Mode CMD to in!orm MS o! t$e selected encry)tion
alorit$m.
4. A!ter receivin Ci)$erin Mode CMD" MS starts t$e transmission o! ci)$erin
mode and sends Ci)$erin Mode CM@ to t$e system.
1. A!ter receivin t$e Ci)$erin Mode CM@ !rom MS" BSC trans!er it to MSC.
>>. @rocedure Descri)tion
A1 alorit$m
GSM )rotocol s)eci!ies ei$t ?inds o! encry)tion alorit$m !rom A1=0 to A1=5.
A1=0 stands !or no encry)tion. +$e encry)tion )rocedure is initiated by t$e
net#or?. +$e encry)tion in!ormation o! Ci)$er Mode CMD s)eci!ies t$e
required encry)tion alorit$m. +$e alorit$m t$at enerates encry)ted code is
called A1 alorit$m. >t calculates by usin t$e &c -34 bits. and t$e current
!rame number -%% bits. to enerate a 114,bit encry)tion sequence and t$en
im)lements OA< o)eration #it$ t$e 114,bit burst. +#o encry)tion sequences
are used !or u)lin? and do#nlin?. For eac$ burst" one sequence is used !or MS
encry)tion and B+S decry)tion" t$e ot$er sequence is used !or B+S encry)tion
and MS decry)tion.
;ncry)tion alorit$m selection
9$en MS initiates call request" t$e SABM !rame carries Classmar? 1 or % to
indicate #$et$er t$e MS su))orts alorit$m A1=1" A1=%" or A1=2" and re)orts
Classmar? 2 in CGASS MA<& C'AEG; to !urt$er indicate #$et$er t$e MS
su))orts Alorit$m A1=4" A1=1" A1=3" or A1=5->n system in!ormation" i!
;CSCN1" MS re)orts Classmar? 2 immediately8 i! ;CSC N 0" t$e Classmar? 2 is
re)orted a!ter CGASSMA<& ;EQ/><R is initiated by t$e net#or?. +$ere!ore" t$e
con!iuration o! ;CSC N 1 is recommended #$en t$e encry)tion is used.. MSC
sends encry)tion command based on t$e con!iuration o! secret data. BSC
c$ooses t$e intersection o! t$e encry)tion alorit$m allo#ed in t$e command
sent by MSC" t$e encry)tion alorit$m allo#ed in BSC data con!iuration" and
t$e encry)tion alorit$m su))orted in t$e MS re)ort. >n t$e intersection" BSC
selects a )ro)er alorit$m based on t$e )riority level o! A1=5 S A1=3 S A1=1 S
A1=4 S A1=4 S A1=2 S A1=% S A1=1 S A1=0.
;ncry)tion in $andover
+$e 'AEDAH;< <;Q/;S+ contains t$e encry)tion in!ormation unit t$at
indicates t$e required encry)tion alorit$m and ?ey. >! one o! t$e t#o A
inter!aces o! BSS is in @'AS; >" due to t$e limitation o! ;+S> GSM @'AS; >
)rotocol -no ci)$erin mode settin in!ormation unit in $andover command."
t$e t#o A inter!aces matc$ only #$en t$ey s$are t$e same encry)tion
alorit$m -suc$ as A1=%. to ensure t$e normal inter,BSC $andover. At$er#ise"
s)ecial treatment $as to be made to t$e taret MSC or taret BSC -or t$e
source MSC or source BSC. to c$ane t$e $andover command !or inter,BSC
$andover.
For t$e interconnection o! A,inter!aces #$en t$e encry)tion is used" #$et$er
s)ecial data con!iuration is required !or BSC and MSC must be considered.
1.11.2 +MS> <eallocation
A!ter aut$entication and encry)tion" t$e system sends CM S;<H>C; ACC;@+ or
+MS> reallocation command to MS and initiates +2%10.
9$en MS reisters in t$e location area !or t$e !irst time" t$e net#or? allocates
a +MS> to it. 9$en t$e MS leaves t$is location area" it releases t$e +MS>. 9$en
t$e MS receives t$e +MS> reallocation command" it saves t$e +MS> and GA> and
sends +MS> reallocation com)lete messae. A!ter receivin t$is messae" t$e
net#or? sto)s +2%10.
>! t$e system cannot identi!y +MS> o! t$e MS" !or e:am)le" #$en t$e data base
error occurs" t$e MS must )rovide its >MS>. +$e identi!ication )roram is
initiated be!ore t$e +MS> reallocation to request !or t$e >MS>.
+$e identi!ication )roram sends identity request messae to t$e MS" a!ter
receivin t$is messae" t$e MS )rovides its >MS> by sendin identity res)onse
messae to t$e net#or?. 9$en t$is )rocedure is over" aut$entication"
encry)tion" and >MS> reallocation are im)lemented i! required.
1.11.4 ;:ce)tional Situations
>. Aut$entication
<< connection !ailure
>! t$e net#or? detects << connection !ailure be!ore receivin A/+';E+>CA+>AE
<;S@AES;" it releases all t$e MM connections and terminates all t$e active MM
)rocedures.
+2%30 timeout
+2%30 is started #$en MSC sends aut$entication request to BSC and sto)s #$en
MSC receives A/+';E+>CA+>AE <;S@AES;. >! t$e +2%30 times out be!ore t$e
A/+';E+>CA+>AE <;S@AES; is received" t$e net#or? releases << connection"
terminates t$e aut$entication )rocedure and all t$e active MM )rocedures" and
t$en releases all t$e MM connections and initiates << connection release
)rocedure.
/nreistered S>M card
>! t$e S>M card o! t$e MS is not reistered" t$e net#or? sends A/+';E+>CA+>AE
<;X;C+ messae directly to t$e MS.
>>. ;ncry)tion
;ncry)tion reFect
>! BSS does not su))ort t$e encry)tion alorit$m s)eci!ied in C>@';<>EG MAD;
CMD" it sends C>@';< MAD; <;X;C+ messae to MSC.
>! t$e encry)tion is initiated in BSS be!ore MSC requests !or t$e c$ane o!
encry)tion alorit$m" BSS also sends C>@';< MAD; <;X;C+ messae to MSC.
/n,encry)ted MS
+$e C>@';<>EG MAD; CAMMAED messae is valid #$en*
0+$e un,encry)ted MS receives C>@';<>EG MAD; CAMMMAED messae t$at
requires encry)tion.
0+$e un,encry)ted MS receives C>@';<>EG MAD; CAMMMAED messae t$at
requires non,encry)tion.
0+$e encry)ted MS receives C>@';<>EG MAD; CAMMMAED messae t$at
requires non,encry)tion.
>n ot$er cases" C>@';<>EG MAD; CAMMAED is considered #ron. +$e MS sends
<< S+A+/S messae #it$ t$e cause o! )rotocol error and )er!orms no action.
>>>. +MS> <eallocation
<< connection !ailure
>! << connection !ails be!ore +MS> reallocation com)lete messae is received"
all t$e MM connections are released and bot$ t$e old and ne# +MS>s are saved
durin a certain recovery time.
+2%10 timeout
+2%10 is started #$en MSC sends +MS>P <;AGGP CMD messae or GAC /@D ACC
messae #it$ t$e ne# +MS> and sto)s #$en MSC receives +MS> P<;AGGPCAM. >!
+2%10 times out be!ore t$e +MS> P<;AGGPCAM is received" MSC sends CG;A<
CAM messae to release << connection and terminate +MS> reallocation.
1.1% Gocation /)date
>n GSM" t$e )ain in!ormation cannot be sent in t$e #$ole net#or? due to t$e
ca)acity limit o! t$e )ain c$annel. +$ere!ore" t$e de!inition o! location area
-GA. is introduced. GAC contains many cells. +$e )ain !or t$e MS is carried
out t$rou$ t$e )ain in all t$e cells #it$in t$e GA o! t$e MS. +$e si(e o! t$e
GA is o! vital im)ortance to t$e system )er!ormance in net#or? desin.
+$e reistration manaement !or t$e GA is required since t$e )ain !or t$e MS
is carried out t$rou$ t$e )ain in all t$e cells #it$in t$e GA" #$ic$ brins
about t$e de!inition o! location u)date. Gocation u)date is divided into eneric
location u)date" )eriodic location u)date" and >MS> attac$.
1.1%.1 Generic Gocation /)date ->nter,GA Gocation /)date.
9$en t$e MS moves !rom one GA to anot$er GA" reistration is required. >! t$e
GA> stored in t$e MS is di!!erent !rom t$e GA> o! t$e current cell" t$e MS in!orms
t$e net#or? to c$ane t$e location in!ormation it stores. +$is )rocedure is
called eneric location u)date.
>n idle mode" i! cell re,selection occurs #$en t$e MS moves #it$in t$e GA" t$e
MS #ill not in!orm t$e net#or? immediately but im)lement cell re,selection
#it$out location u)date or net#or? involvement. >! t$e MS moves to anot$er GA
a!ter re,selection" t$e MS in!orms t$e net#or? o! t$is GA c$ane" #$ic$ is
called !orced reistration.
Accordin to #$et$er t$e HG< c$anes or >MS> involves" eneric location u)date
is divided into t$e !ollo#in ty)es*
>. >ntra Hl< Gocation /)date
>t is t$e sim)lest location u)date t$at requires no >MS>. >t $a))ens in t$e
current HG< #it$out in!ormin t$e 'G<.
>n t$e initial messae carried by SABM !rame" t$e access cause is MM GACA+>AE
/@DA+>EG <;Q/;S+ t$at carries t$e MS +MS> and GA>. +$e eneric location
u)datin is indicated. MSC receives t$is messae and !or#ards it to HG<. HG<
u)dates t$e MS location in!ormation and stores t$e ne# GA>" and t$en sends a
ne# +MS> to MS i! required -MS uses t$e !ormer +MS> i! no +MS> is carried in t$e
+MS> re,allocation command.. A!ter receivin t$e +MS> re,allocation com)lete
messae" MSC sends location u)datin acce)t messae and releases t$e
c$annel. Gocation u)datin com)letes.
>>. >nter,HG< Gocation /)datin" Sendin +MS>
A!ter t$e MS enters a cell" i! t$e current GA> is di!!erent !rom t$e GA> it stores"
it sends its GA> and +MS> to HG< t$rou$ MSC in location u)datin request. HG<
deduces t$e !ormer HG< based on t$e GA> and +MS> it received and sends a
MA@PS;EDP>D;E+>F>CA+>AE to t$e !ormer HG< to request !or >MS> and
aut$entication )arameter. +$e !ormer HG< sends t$e >MS> and aut$entication
)arameters to t$e current HG<. >! t$e current HG< cannot obtain t$e >MS>" it
sends MS an identity request messae to request !or t$e >MS>. A!ter receivin
t$e >MS>" HG< sends 'G< t$e location u)datin messae t$at contains t$e MS
identity in!ormation !or t$e data query and )at$ establis$ment o! 'G<. A!ter
receivin t$is messae" 'G< stores t$e number o! t$e current HG< and sends
MA@=DPCAEC;GPGACA+>AE to t$e !ormer HG< i! t$e current MSC=HG< $as t$e
normal service ri$ts. A!ter receivin t$is messae" t$e !ormer HG< deletes all
t$e in!ormation about t$is MS and sends t$e 'G< a
MA@=DPCAEC;GPGACA+>AEP<;S/G+ messae to con!irm t$e deletion. +$e 'G<
#ill send MA@P>ES;<+PS/BSC<>B;<PDA+A messae to )rovide t$e current HG<
#it$ t$e in!ormation it requires -includin aut$entication )arameters. a!ter t$e
)rocedure !or aut$entication" encry)tion" and +MS> reallocation is over" and
con!irm t$e location u)datin a!ter receivin t$e res)onse !rom t$e HG<.
>>>. >nter,HG< Gocation /)datin" Sendin >MS>
+$e )rocedure is similar #it$ t$e )rocedure above but easier because it
requests !or aut$entication )arameter !rom t$e 'G< t$rou$ >MS> directly.
1.1%.% @eriodic Gocation u)datin
+$e net#or? and t$e MS lose contact #$en*
+$e MS is s#itc$ed on but moves out o! t$e net#or? coverae area -dead
(one.. +$e net#or? lost contact #it$ t$e MS and reards it still in attac$
status.
+$e MS sends >MS> detac$ messae and t$e u)lin? quality is bad due to
inter!erence" t$e net#or? may not be able to decode t$is messae correctly.
+$e MS is still rearded in attac$ status.
+$e MS is )o#er o!!. >t cannot in!orm t$e net#or? o! its status and t$e contact
is lost.
>! t$e )ain !or MS $a))ens #$en t$e contact is lost" t$e system sends )ain
in!ormation in t$e GA t$at t$e MS reistered be!ore. +$e net#or? cannot
receive t$e res)onse !rom t$e MS. +$e system resource is #asted. +o solve t$is
)roblem" t$e im)licit detac$ timer is introduced in t$e HG< !or t$e >MS> status
manaement. >n addition" measures are ta?en in BSS to !orce t$e MS to re)ort
its location )eriodically. +$ere!ore" t$e net#or? is in!ormed o! t$e status o!
MS. +$is ?ind o! mec$anism is called )eriodic location u)datin. +$e net#or?
sends a )eriodic location u)datin time +2%1% to all t$e users in t$e cell
t$rou$ BCC' to !orce t$e MS to send location u)datin request #it$ t$e cause
o! )eriodic location u)datin a!ter +2%1% times out.
Be!ore t$e +2%1% times out" i! t$e timeout value is c$aned -!or e:am)le" t$e
service cell c$anes and t$e +2%1% timeout value is broadcast." t$e MS uses t$e
time #$en t$e c$ane $a))ens as t$e initial value and ?ee) on timin.
>! t$e +2%1% times out #$en t$e MS is in EA C;GG AHA>GABG;" G>M>+;D S;<H>C;"
@GME S;A<C'" or @GME S;A<C',EA<MAG S;<H>C; status" t$e location u)datin
is initiated a!ter t$e MS is out o! t$ese service status.
@eriodic location u)datin ensures t$e close contact bet#een net#or? and
mobile users. +$e s$orter u)datin )eriod leads to better net#or?
)er!ormance. But t$e !requent location u)datin #ill increase t$e sinalin
!lo# and reduce t$e utili(ation o! t$e radio resources" or even a!!ect t$e
)rocessin ability o! MSC" BSC" and B+S. An t$e ot$er $and" it #ill reatly
increase t$e )o#er consum)tion o! MS and reduce its standby time. +$e +2%1%
settin s$ould be based on com)re$ensive consideration.
+$e )rocedure !or )eriodic location u)datin is t$e same as t$at !or eneric
location u)datin.
1.1%.2 >MS> Attac$ and Detac$
>MS> attac$ and detac$ means to attac$ a binary mar? to t$e subscriber record
in MSC=HG<. +$e !ormer one is mar?ed as access ranted" and t$e latter one is
mar?ed as access denied.
9$en t$e MS is s#itc$ed on" it in!orms t$e net#or? o! its status c$ane by
sendin an >MS> A++AC' messae to t$e net#or? to in!orm. A!ter receivin t$is
messae" t$e net#or? mar?s t$e current user status in t$e system database !or
t$e )ain )roram.
>! t$e current GA> and t$e GA> t$e MS stores are t$e same" >MS> attac$ is
initiated. +$e )rocedure is similar to t$e intra HG< location u)datin only t$at
t$e location u)datin request messae is mar?ed as >MS> attac$ and t$e initial
messae contains >MS> o! t$e MS.
>! t$e current GA> is di!!erent !rom t$e GA> stored" eneric location u)datin is
initiated.
9$en t$e MS is s#itc$ed o!!" t$e >MS> detac$ is triered by a ?ey,)ress. Anly
one command is sent to MSC=HG< !rom t$e MS. +$is is an unac?no#leded
messae. A!ter receivin t$is messae" MSC in!orms HG< to do detac$ mar? to
t$is >MS> #$ile t$e 'G< is not in!ormed o! t$e no,radio o! t$is user. 9$en t$e
)ain !or t$is user occurs" 'G< requests !or t$e MS<E !rom t$e HG< and is
in!ormed o! t$e no,radio o! t$is user by t$is time. +$ere!ore" no )ain
)roram is im)lemented. +$e )ain messae is $andled directly" suc$ as
)layin t$e record* Y+$e subscriber is )o#ered o!!.Y
+$e )rocedure above is e:)licit >MS> detac$. +$ere is also im)licit detac$. +$e
im)licit detac$ $a))ens be!ore t$e im)licit detac$ timer times out. >! t$e
contact bet#een MS and net#or? is not establis$ed" t$e HG< sets t$e >MS>
status as detac$. +$e im)licit detac$ timer is set loner t$an t$e )eriodic
location u)datin timer +2%1% to avoid YabnormalY im)licit detac$. +$e im)licit
detac$ is denied durin t$e establis$ment o! radio connection. +$e im)licit
detac$ timer is reset a!ter t$e release o! radio connection. >m)licit detac$
timer is also called >MS> delete time.
HG< deletes t$e >MS> mar?ed as detac$ )eriodically -+$e )eriod is adFustable.
and re)orts t$e user status to t$e 'G<.
1.1%.4 ;:ce)tional Situations
>. MS
Access denied because o! access level limit
MS stays in t$e service cell and )er!orms t$e normal cell re,selection )rocedure
#it$out trierin location u)datin. 9$en t$e current cell allo#s access or
ot$er cell is selected" +$e MS initiates location u)datin immediately.
>MM;D>A+; ASS>GEM;E+ <;X;C+ messae is received durin random access
MS stays in t$e service cell and starts +21%% based on t$e value in t$e
immediate assinment reFect messae. +$e normal cell selection and re,
selection )rocedure is )er!ormed. >! t$e cell t$at t$e MS stays c$anes or
+21%% times out" t$e MS initiates location u)datin.
<andom access !ailure
>! t$e random access !ails" +2%12 is started. A!ter t$e +2%12 times out" t$e
random access )rocedure is initiated. >! t#o successive random accesses !ail"
t$e location u)datin is terminated. For t$e subsequent )rocessin" see t$e
!ollo#in descri)tion.
<< connection !ailure* Gocation u)datin )rocedure is terminated. For t$e
subsequent )rocessin" see t$e !ollo#in descri)tion.
+2%10 timeout* Gocation u)datin !ails. For t$e subsequent )rocessin" see t$e
!ollo#in descri)tion.
+$e com)letion o! << connection is abnormal* Gocation u)datin !ails. For t$e
subsequent )rocessin" see t$e !ollo#in descri)tion.
Gocation u)datin reFect due to reasons ot$er t$an Z%" Z2" Z3" Z11" Z1%" or
Z12* MS #aits !or t$e release o! << connection. For t$e subsequent )rocessin"
see t$e !ollo#in descri)tion.
Z % ->MS> un?no#n in 'G<.
Z 2 ->lleal MS.
Z 3 ->lleal M;.
Z 11 -@GME not allo#ed.
Z 1% -Gocation Area not allo#ed.
Z 12 -<oamin not allo#ed in t$is location area.
Subsequent )rocessin* >! t$e +2%10 is still timin" sto) it8 >! +2%10 times out"
<< connection !ails. Add 1 to t$e location u)datin attem)t timer. +$e
!ollo#in )rocessin de)ends on t$e GA> -stored and received !rom t$e service
cell. and t$e value o! t$e location u)datin attem)t timer.
>! t$e location u)datin status is /@DA+;D" t$e stored GA> and t$e received GA>
are t$e same" and t$e location u)datin attem)t timer is less t$an 4" MS ?ee)s
t$e /@DA+;D status. A!ter t$e release o! << connection" t$e sub status o! MM
>DG; becomes EA<MAG S;<H>C;. +$e MS also stores t$e in!ormation about t$e
!ormer location u)datin ty)e. +$e +2%11 is started a!ter << connection
release. A!ter it times out" t$e location u)datin )rocedure is started aain.
>! t$e location u)datin status is not /@DA+;D" or t$e stored GA> is di!!erent
!rom t$e received GA>" or t$e location u)datin attem)t timer is equal to or
less t$an 4" t$e MS deletes t$e ci)$erin ?ey sequence" GA>" +MS> stored in S>M
card and sets t$e location u)datin status as EA+ /@DA+;D. A!ter t$e release
o! << connection" t$e sub status o! MM >DG; becomes A++;M@+>EG +A /@DA+;.
A!ter t$e << connection release" i! t$e location u)datin attem)t is less t$an 4"
+2%11 is started. At$er#ise" +2%1% is started. A!ter t$e +2%11 or +2%1% times
out" t$e location u)datin )rocedure is started aain.
A!ter t$e sub status o! MM >DG; becomes A++;M@+>EG +A /@DA+;" t$e MS #ill
do t$e !ollo#in*
>! +2%11" +2%12" or +2%1% times out" )er!orm location u)datin.
>! GA c$anes" )er!orm eneric location u)datin
>! t$e cause !or t$e status c$ane is -2." -4." -3. -t$e cause is not t$e abnormal
release #it$ un?no#n reason." or -5. -cause 6retry in t$e ne# cell7." )er!orm
location u)datin #$en enterin t$e ne# cell.
>! t$e cause !or t$e status c$ane is -1." -3. -t$e cause is abnormal release #it$
un?no#n reason." or -5. -t$e cause is not 6retry in t$e ne# cell7." location
u)datin is not )er!ormed #$en enterin t$e ne# cell.
Eo >MS> detac$.
Su))ort emerency call request
<es)ond t$e )ain #it$ >MS>
@er!orm eneric location u)datin triered by t$e request !rom CM layer -i!
t$e location u)datin succeeds" t$e MMG connection request #ill be acce)ted.
For details" see section 4.1.1 o! t$e @rotocol 0408..
>>. Matc$in Bet#een >MS> Delete +ime and +2%1%
>! t$e )eriodic location u)datin !ails !or !our times" +2%1% #ill be started !or
t$e ne:t u)date. >n t$e bad coverae area" es)ecially in t$e area #$ere t$e
u)lin? and do#nlin? do not matc$ -do#nlin? is better t$an u)lin?." a!ter t$e
)eriodic location u)date !ails"
Anot$er location u)datin is initiated a!ter +2%1% times out. +$ere!ore" t$e
+2%1% is set to be s$orter in t$e bad coverae area. >n addition" i! t$e >MS>
delete time is less t$an t#ice o! t$e +2%1%" t$e users stay in t$e service area
but cannot be called. So t$e >MS> delete time s$ould be more t$an t#ice o! t$e
+2%1% and based on GAC.
>>>. Eet#or?
<< connection !ailure
Amon all t$e sub )rocedures attac$ed to t$e location u)datin )rocedure" i!
t$e << connection !ails" it is $andled accordin to t$e e:ce)tion $andlin o!
ot$er common )rocedures.
>! no ot$er common )rocedure is attac$ed to t$e location u)datin )rocedure"
t$e MS location u)datin is terminated.
@rotocol error
>! t$e net#or? detects )rotocol error a!ter receivin GACA+>AE /@DA+>EG
<;Q/;S+" it sends GACA+>AE /@DA+>EG <;X;C+ messae to t$e MS #it$ t$e
!ollo#in cause i! )ossible*
Z93 required >; error
Z99 >; error or no >; e:ists
Z100 Conditional >; error
Z111 @rotocol error" unde!ined
A!ter sendin GACA+>AE /@DA+>EG <;X;C+ to t$e MS" t$e net#or? initiates
c$annel release )rocedure.
1.12 MS Ariinatin Call Flo#
+$e MS needs to set u) a main sinalin lin? to connect to MSC !irst" and t$en
initiates t$e aut$entication" encry)tion" and +MS> reassinment !lo#.
1.12.1 Called Eumber Analysis
A!ter t$e aut$entication" encry)tion" and +MS> reassinment !lo# are over" t$e
MS starts t$e call setu) !lo#.
First" t$e MS sends a S;+/@ messae to t$e net#or? side. +$is messae
contains called number and t$e required services. +$e MSC im)lements t$e call
)roceedin accordin to t$e messae.
9$en receive t$e S;+/@ messae" t$e MSC sends t$e outoin call messae
S;EDP>EFAPFA<PA=CPCAGG to t$e HG<. A!ter receive t$e outoin call
messae" t$e HG< analy(es t$e items suc$ as called number" t$e callin )arty
ca)ability" and net#or? resources ca)ability accordin to t$e user in!ormation
obtained !rom t$e 'G< durin t$e location u)datin )rocess" to c$ec? #$et$er
to acce)t t$is call request. >! a certain item cannot be )assed" t$e HG< sends
t$e <;G;AS; CAM@G;+; messae to t$e MS. +$e call !ails. +$e MS t$en
)roceeds to release t$e bottom layer connection and s#itc$es to t$e idle state.
>! t$e above items can be )assed" t$e HG< sends t$e CAM@G;+;PCAGG messae
to t$e MSC. A!ter receive t$is messae" t$e MSC sends t$e CAGG @<AC;;D>EG
messae to t$e MS. >t means t$at t$e call request is acce)ted and t$e call is
set u).
1.12.% Hoice C$annel Assinment -Follo#,u) Assinment.
A!ter send t$e CAGG @<AC;;D>EG messae to t$e MS" t$e MSC activates t$e
!ollo#,u) assinment accordin to t$e service request. +$at is" assin t$e +C'
voice c$annel to t$e user. At t$is time" t$e MSC sends t$e ASS>GEM;E+
<;Q/;S+ messae to t$e BSC. +$is messae contains t$e in!ormation suc$ as
t$e requested c$annel ty)e to request t$e BSC to assin t$e +C' voice c$annel
!or t$e call.
A!ter receive t$e c$annel request !rom t$e MSC" t$e BSC sends t$e C$annel
Activation !or +C' messae to t$e B+S to activate corres)ondin terrestrial
resources and start a timer at t$e same time i! t$e +C' c$annel resources are
available. >! t$e B+S $as )re)ared t$e resources suc$ as circuit" t$e B+S sends
t$e C'AEE;G AC+>HA+>AE AC& messae to t$e BSC. >! t$e BSC $as no available
resources to assin" it sends t$e <;SA/<C; FA>G/<; messae to t$e MSC. But i!
t$e system allo#s queuin" t$e BSC sends t$e Q/;/>EG >ED>CA+>AE messae to
t$e MSC and )laces t$e assinment request in t$e queue and starts t$e timer
+11. >! t$e +11 times out" t$e BSC sends t$e CG;A< <;Q/;S+ messae to t$e
MSC.
+$e immediate assinment request" intra,BSC $andover" and inter,BSC
$andover do not su))ort queuin. Anly t$e +C' resource request -t$at is" t$e
assinment request and intra,cell $andover. allo#s queuin. +$e +C' resource
requests in t$e queue are assined #it$ relevant c$annels in t$e sequence o!
t$eir )riorities. >n t$e lent$ o! t$e queue reac$es its t$res$old or t$e timer
times out" t$e request is reFected.
9$en t$e BSC receives t$e C'AEE;G AC+>HA+>AE AC& messae !rom t$e B+S"
t$e BSC )uts t$e )$ysical in!ormation o! t$e c$annel )rovided by t$e B+S in t$e
ASS>GEM;E+ CAMMAED messae -t$is messae contains t$e in!ormation suc$ as
c$annel ty)e" voice=data indication" c$annel rate" voice decodin alorit$m
and trans)arent transmission indicator" assinment )riority and C>C.. +$e
ASS>GEM;E+ CAMMAED messae is sent to t$e MS t$rou$ t$e SDCC' c$annel.
A!ter receive t$e ASS>GEM;E+ CAMMAED messae !rom t$e B+S" t$e MS adFusts
t$e transceiver con!iuration to t$e +C' c$annel and t$en sends t$e SABM
messae to t$e B+S t$rou$ t$e FACC' c$annel in t$e #ay o! stolen !rame.
A!ter t$e B+S receives t$e SABM messae" t$e B+S sends t$e ;S+ABG>S'
>ED>CA+>AE messae to t$e BSC and t$en sends an /nnumbered Ac?no#lede
-/A. to t$e MS" Fust as t$e initial sinalin c$annel assinment does.
A!ter receive t$e /A" t$e MS sends t$e ASS>GEM;E+ CAMM@G;+; messae to
t$e B+S t$rou$ t$e FACC' c$annel. >! t$e MS !ails to identi!y t$e assinment
in!ormation and !ails to occu)y t$e s)eci!ied c$annel due to t$e radio inter!ace
!ailure" radio inter!ace messae !ailure or inter!erence" or $ard#are )roblems"
t$e MS returns to t$e oriinal c$annel and sends t$e ASS>GEM;E+ FA>G/<; to
t$e B+S. >! t$e MS does not receive t$e ASS>GEM;E+ CAMMAED sent !rom B+S
or t$e B+S does not receive t$e res)onse messae sent !rom MS due to
inter!erence or ot$er causes" t$e system starts t$e corres)ondin timers -suc$
as +2102 or +2105. and #$en t$e timer times out" t$e c$annel is released.
9$en receive t$e ASS>GEM;E+ CAM@G;+; messae" t$e BSC sends t$e
ASS>GEM;E+ CAM@G;+; messae to t$e MSC. At t$e same time" it also sends
t$e <F C'AEE;G <;G;AS; messae to t$e B+S to release t$e occu)ied SDCC'
sinalin c$annel. 9$en t$e B+S releases t$e sinalin c$annel" it sends t$e <F
C'AEE;G <;G;AS; AC& messae to t$e BSC. A!ter t$e BSC receive t$e messae"
it considers t$at t$e sinalin c$annel is in idle state and can be assined to
ot$er c$annel requests.
For di!!erent )ur)oses" t$e GSM $as t$ree di!!erent c$annel assinment !lo#s.
+$ey are initial c$annel assinment" !ollo#,u) c$annel assinment" and
$andover c$annel assinment.
>nitial c$annel assinment* is mandatory to establis$ t$e lin? transmission
bet#een t$e MS and t$e net#or?. For e:am)le" )rocess t$e location u)datin
request.
Durin t$e establis$ment o! t$e sinalin transmission" i! t$e +C' c$annel is
assined )re!erably" t$is assinment is called very early assinment -H;A..
A!ter t$e MSC sends t$e ASS>GEM;E+ <;Q/;S+ messae" t$e BSC does not a))ly
!or ne# c$annel but initiate t$e ModePModi!y !lo#. A!ter t$e ModePModi!y is
com)lete" t$e BSC re)orts t$e ASS>GEM;E+ CAM@G;+; messae to t$e MSC.
>! t$e SDCC' c$annel is assined !irst" and t$e +C' c$annel is assined #$en it
is needed" and t$en ASS>GEM;E+ <;Q/;S+ messae !rom MSC is sent be!ore t$e
Alertin messae" t$is assinment is called early assinment -;A..
>! t$e SDCC' c$annel is assined !irst and t$e +C' is assined a!ter t$e called
)arty sends t$e CAEE;C+ messae" Generally" it ado)ts t$e ;A mode.
>! t$e ;A mode is used in t$e initial assinment" #$en no SDCC' is available"
assin t$e +C' c$annel !or t$e c$annel request directly. +$e +C' c$annel
re)laces t$e SDCC' c$annel to send t$e sinalin messae. @lease note t$at
usin t$e +C' c$annel to transmit t$e sinalin #astes t$e resources a lot
because one +C' c$annel equals ei$t SDCC' c$annels. 9$en t$is situation is
quite serious" add more SDCC' to meet t$e requirement in time.
Follo#,u) c$annel assinment
A!ter t$e sinalin c$annel !inis$es t$e aut$entication and encry)tion )rocess"
i! t$ere is still voice or data request" t$e !ollo#,u) c$annel assinment is
triered to assin a +C' c$annel.
'andover c$annel assinment
+$is assinment is used to a))ly !or c$annels due to $andover durin t$e call
)rocess. +$e system Fudes #$et$er t$e $andover occurs in t$e SDCC' or in
t$e +C' to assin corres)ondin c$annels. +$e $andover !lo# and t$e
assinment !lo# in t$e cell are t$e same. +$e only di!!erence is t$at t$e
messae names are di!!erent. Similar to t$e immediate assinment !lo#" in t$e
MS assinment !lo#" t$e timer +2105 starts #$en t$e BSC sends t$e
ASS>GEM;E+ CAMMAED messae to t$e B+S. A!ter t$e BSC receives t$e
ASS>GEM;E+ CAM@G;+; messae !rom t$e B+S" t$e timer +2105 resets.
Generally" t$e timeout o! t$e timer is caused by t$e bad radio coverae. 9$en
t$e timer times out" t$e MS is considered disconnected #it$ t$e net#or? and
t$e resources are released !or ot$er MSs. Based on t$e statistics" t$e c$annel
assinment is enerally com)lete #it$in t#o seconds. >! t$e BSC does not
receive t$e ASS>GEM;E+ CAM@G;+; messae #it$in t#o seconds" t$e
assinment !ails. But sometime" t$e net#or? quality is bad" some messaes
needs to be sent several times" in t$is case" t$e assinment can be e:tended to
!ive seconds. Generally" i! t$e tra!!ic load o! t$e cell is $eavy" set t$e timer as
% seconds to 1 seconds. >! not $eavy" set t$e timer as 10 seconds.
1.12.2 Call Connection
A!ter receivin t$e ASS>GEM;E+ CAM@G;+; messae !rom t$e BSC" t$e MSC
sends t$e >nitial Address Messae ->AM. t$at includes t$e in!ormation used to
establis$ t$e route to t$e called net#or?. +$e MSC #ill receive t$e call setu)
re)ort soon. >! succeeds" t$e MSC receives an ADDD<;SS CAM@G;+; messae
-ACM.8 i! !ails because o! certain reason -suc$ as busy line or conestion." t$e
MSC receives a <;G;SAS; messae !rom t$e called end.
>! MSC receives t$e ACM" MSC sends t$e AG;<+>EG messae to t$e MS -MS
translates it into rin bac? tone.. +$is messae is a D+A@ messae. >! no ans#er
is received !rom t$e called )arty and t$e callin )arty does not terminate t$e
connection" t$e net#or? #ill terminate t$e call or )er!orm no ans#er call
trans!er a!ter a #$ile.
>! t$e called )arty )ic?s u) t$e )$one" MSC receives an AES9;< messae. +$e
lin? bet#een t$e callin )arty and t$e called )arty is connected. MSC sends a
CAEE;C+ messae in t$e CC )rotocol to t$e MS. A!ter receivin t$is messae"
t$e MS sends a CAEE;C+ AC&EA9G;DG; messae in t$e CC )rotocol to t$e
system. +$e system starts c$arin a!ter receivin t$is messae. >! t$e called
end is data device" it enters CAEE;C+ status directly a!ter receivin t$e S;+/@
indication. +$e call connection )rocedure is over and t$e t#o )arties start t$e
conversation or data transmission service.
1.12.4 Call <elease
>! t$e callin )arty $ans u) !irst" t$e MS sends disconnect messae to MSC
t$rou$ FACC'. A!ter receivin t$is messae" t$e MSC sends release messae to
in!orm t$e called )arty to terminate t$e communication. +$e end,to,end
connection is over. But t$e call is not com)lete" because certain tas?s suc$ as
sendin c$are indication are )er!ormed. 9$en t$e connection to t$e MS is no
loner necessary" t$e system sends a <;G;AS; messae to t$e MS and starts
+208. A!ter receivin t$is messae" t$e MS sends a <;G;AS; CAM@G;+; messae
to t$e system and t$e call is over. +$e MS sto)s t$e +208 a!ter receivin t$e
<;G;AS; CAM@G;+; messae. Similarly" i! t$e called )arty $ans u) !irst" it
sends a <;G;AS; messae to t$e callin )arty. +$e MSC sends t$e callin )arty
a D>SCAEE;C+ messae a!ter receivin t$e <;G;AS; messae. >! t$e call is
terminated in an abnormal #ay" t$is messae !urt$er indicates t$e cause !or
t$at.
9$en t$e MSC receives t$e <;G;AS; CAM@G;+; messae !rom t$e MS" it sends a
CG;A< CAMMAED messae to BSC to release all t$e sinalin lin?s. +$is messae
contains t$e cause !or t$e call clearance" suc$ as $andover com)lete or
location u)datin com)lete. +$e call connection release is over. >! t$e
abnormal release occurs because o! radio lin? !ailure or device !ailure" t$e BSC
sends a CG;A< <;Q/;S+ messae to t$e MSC.
A!ter receivin t$e CG;A< <;Q/;S+ messae" BSC sends a C'AEE;G <;G;AS;
messae to t$e MS and starts +2109 to s$o# t$at all t$e lo#er layer lin?s are
released. Mean#$ile" it requires t$e MS to enter t$e idle mode. 9$en t$e MS
receives t$e C'AEE;G <;G;AS; messae" it removes t$e u)lin? sinalin lin?
-to sto) sendin t$e measurement re)ort o! u)lin? c$annel associated sinalin
on SACC'.. +$e MS sends D>SC messae to B+S and starts +2110. A!ter receivin
t$is messae" +$e B+S sends /A to MS and t$e <;G;AS; >ED>CA+>AE to t$e BSC.
9$en t$e +2110 times out or t$e MS receives t$e /A !rame" it enters t$e idle
mode.
>n order to ensure t$e timely removal o! t$e u)lin? and do#nlin?" #$en t$e BSC
sends t$e C'AEE;G <;G;AS; messae to t$e MS !or t$e u)lin? removal" it also
sends a deactivate SACC' -SACC'. to t$e B+S requirin !or t$e release o! t$e
do#nlin? sinalin -to sto) t$e sinalin connection bet#een t$e t#o )arties..
A!ter receivin t$is messae" t$e B+S sto)s t$e transmission o! t$e do#nlin?
SACC' !rame and sends t$e deactivate SACC' ac?no#ledement to t$e MSC.
A!ter receivin t$e <;G;AS; >ED>CA+>AE messae" BSC resets t$e +2109 and
starts t$e +2111" and sends <F C'AEEG; <;G;AS; to t$e B+S -t$e +2111 is reset
at t$e same time." requirin !or t$e release o! +C' resources. 9$en t$e BSC
receives t$e <F C'AEEG; <;G;AS; ac?no#ledement messae !rom t$e B+S" it
sends a CG;A< CAM@G;+; messae to t$e MSC" indicatin t$at t$e radio lin?
clearance is over and t$e c$annel is available !or reallocation.
A!ter receivin t$e CG;A< CAM@G;+; messae" t$e MSC releases t$e SCC@
connection by sendin <GSD and receivin <GC. +$e #$ole MS oriinatin call
!lo# is over.
1.12.1 ;:ce)tional Situations
>. Eo ;stablis$ >ndication Messae >s <eceived A!ter C$annel Activation
+$e main causes are*
+$e MS may send many c$annel requests even i! t$e BSS #or?s #ell" #$ic$
activates many sinalin c$annels. But t$e MS only occu)ies one o! t$em. At$er
c$annels are released by t$e BSC a!ter t$e +2101 times out as t$ey cannot
receive t$e establis$ indication !rom t$e MS. >! t$e +:Pinterer is )ro)er" t$e
cause !or t$is )roblem is t$at t$e u)lin? rece)tion is normal but t$e do#nlin?
sinal cannot be received by t$e MS. /nder suc$ circumstances" t$e received
level and t$e received quality o! u)lin? and do#nlin? s$ould be c$ec?ed. >! t$e
MS is not !ar a#ay !rom t$e B+S but t$e received level and t$e received quality
are bad" c$ec? t$e antenna !eeder and t$e +<O in B+S.
>m)ro)er con!iuration o! +:,inteer in BSC
+$e +:,inteer a!!ects t$e interval o! c$annel request re,sendin. >m)ro)er +:,
inteer only leads to t$e activation o! many c$annels by BSS" but no call #ill be
a!!ected.
>>. BSC Sendin >mmediate Assinment <eFect
>! t$e BSC sends immediate assinment reFect to t$e MS a!ter receivin t$e
c$annel required messae" t$e usual causes are*
Eo )ro)er sinalin c$annel is available !or t$e MS because o! all c$annels are
busy or t$e c$annels are bloc?ed.
B+S sends c$annel activation neative ac?no#lede a!ter receivin t$e c$annel
activation messae.
>! t$e B+S sends lots o! c$annel activation neative ac?no#lede messaes to
t$e BSC" it is usually because t$e transmission at Abis inter!ace is not stable"
#$ic$ leads to t$e inconsistent c$annel status o! t$e BSC and B+S" or because
errors occur in certain board o! B+S.
>>>. MSC Sendin Disconnect Messae >nstead o! Assinment <equest to
+erminate t$e Call
>n t$e call connection )rocess" t$e immediate assinment is !ollo#ed by t$e
assinment )rocedure. But due to certain reasons" t$e MSC sends a disconnect
messae instead o! t$e assinment request messae to t$e MS and t$en
terminates t$e call. /nder suc$ circumstances" many com)laint )$ones !rom
users cannot et t$rou$. C$ec? t$e !ollo#in*
+$e A inter!ace circuit o! MSC
+$e data consistencies o! t$e A inter!ace bet#een t$e MSC and BSC" es)ecially
t$e circuit )ool data.
>H. Assinment Failure
A!ter receivin t$e assinment request" t$e BSC sends assinment !ailure
messae instead o! assinment com)lete. +$e usual causes are*
Eo )ro)er voice c$annel is available !or t$e MS.
BSC $as no )ro)er voice c$annel !or t$e MS because all t$e voice c$annels are
busy or t$e c$annels are bloc?ed.
+$e cause value carried by t$e assinment !ailure messae is no radio resource.
+$e MS voice c$annel access !ails.
/nder t$is condition" t$e assinment !ailure is re)orted !rom t$e MS.
Due to t$e s)ecial !eatures o! t$e radio transmission" t$is ?ind o! assinment
!ailure occurs most !requently and is unsolvable. >! t$e occurrence rate is too
$i$" c$ec? t$e antenna !eeder" t$e B+S board" and t$e )arameters related to
c$annel access in BSC data con!iuration.
+$e A inter!ace circuit o! BSC !ails" !or e:am)le" t$e C>C in t$e assinment
request is not available.
+$e $ard#are o! BSC !ails.
+$e cause value in t$e assinment !ailure messae sent by BSC is equi)ment
!ailure.
+$e transmission at A inter!ace !ails.
H. Directed <etry
A!ter receivin t$e assinment request messae !rom t$e MSC" i! no +C' is
available and t$e BSC allo#s directed retry" t$e BSC im)lements t$e $andover
#it$ t$e cause value o! directed retry to c$ane t$e service cell o! t$e MS.
H>. ;:ce)tional @rocedure Due to Call Dro)
Call dro) may occur any time durin t$e call !lo#" #$ic$ a!!ects t$e !ollo#in
)rocedures. For e:am)le" t$e call dro) occurs #$en t$e BSC receives t$e
assinment request messae !rom t$e MSC. +$e assinment )rocedure may be
not com)lete -t$e c$annel may be Fust assined and no assinment command
messae is sent.. /nder t$is condition" BSC may send clear request messae
instead o! assinment com)lete messae or assinment !ailure messae to t$e
MSC.
H>>. ;:ce)tional @rocedure Due to 'anu)
'an u) o! t$e callin )arty or t$e called )arty may occur any time durin t$e
call !lo#" #$ic$ a!!ects t$e !ollo#in )rocedures. For e:am)le" t$e $anu)
occurs #$en t$e BSC receives t$e assinment request !rom t$e MSC. /nder t$is
condition" t$e call !lo# may be terminated be!ore t$e BSC sends assinment
com)lete or assinment !ailure to t$e MSC. +$is assinment )rocedure neit$er
succeeds -BSC sends assinment com)lete. nor !ails -BSC sends assinment
!ailure..
H>>>. ;:ce)tional )rocedure because MSC sends clear command
A!ter t$e A inter!ace connect is establis$ed" MSC may send clear command or
disconnect messae to t$e BSC durin t$e call !lo#" #$ic$ a!!ects t$e !ollo#in
)rocedures. For e:am)le" t$e $an u) occurs #$en t$e BSC receives t$e
assinment request !rom t$e MSC. /nder t$is condition" t$e call !lo# may be
terminated be!ore t$e BSC sends assinment com)lete or assinment !ailure to
t$e MSC. +$is assinment )rocedure neit$er succeeds -BSC sends assinment
com)lete. nor !ails -BSC sends assinment !ailure.
>! it $a))ens many times" analysis t$e !ollo#in t#o !actors*
+$e cause value carried in t$e clear command
+$e cause value is usually t$e call control i! t$e call is terminated in a normal
#ay. At$er#ise" t$e cause value may be )rotocol error" equi)ment !ailure" or
ot$ers.
+$e interval bet#een t$e clear command or disconnect messae and t$e last
messae
+$e interval bet#een t$e clear command or disconnect messae and t$e last
messae indicates #$et$er t$e e:ce)tional )rocedure is triered by timeout.
1.14 MS Ariinated Call Flo#
1.14.1 ;nquiry
A!ter t$e sinalin lin? !or t$e callin end is establis$ed" t$e >nitial Address
Messae #it$ >n!ormation ->A>. is send !rom t$e callin end to t$e GMSC. +$e >A>
contains t$e MS>SDE o! t$e called )arty. GMSC analy(es t$e identi!ication
number o! t$e CCS5 o! t$e 'G< and sends t$is 'G< t$e
S;EDP<A/+>EGP>EFA<MA+>AE messae. A!ter receivin t$is messae" t$e 'G<
c$ec?s t$e user record" and t$en )er!orms di!!erent )rocedures and res)onds
t$e GMSC as !ollo#s*
/nder normal circumstances" t$e 'G< only $as t$e )artial in!ormation about
t$e identi!ication o! t$e current HG<" suc$ as t$e CCS5 address or t$e universal
mar?. +o et t$e routin in!ormation !or t$e call" t$e 'G< sends t$e HG< a
@<AH>D; <AAM>EGP E/MB;< messae t$at contains t$e user >MS> in!ormation"
requirin t$e HG< to )rovide a MS<E !or t$is call. 9$en t$e MSC=HG< receives
t$is messae" it selects a roamin number !rom t$e idle numbers to tem)orarily
connect it to t$e >MS>" and sends t$e @<AH>D;P<AAM>EGPE/MB;<P<;S/G+
messae #it$ t$e MS<E assined to t$is call in it to t$e 'G<. 9$en t$e 'G<
receives t$e MS<E" it trans!ers t$e in!ormation by sendin a
S;EDP<A/+>EGP>EFA<MA+>AEP<;S/G+ messae to t$e call oriinatin GMSC.
+$en t$e GMSC can !ind t$e HG< #it$ t$e obtained MS<E and sends t$e >A> to
it. A!ter receivin t$is messae" t$e MSC restores t$e >MS> o! t$is user in its
memory record #it$ t$e MS<E and starts t$e )ain !or t$e MS. A!ter t$e call is
establis$ed" t$is roamin number is released !or anot$er user.
>! t$e record o! t$e called )arty is set as Barrin o! All >ncomin Calls -BA>C. or
Barrin o! >ncomin Calls #$en roamin is outside t$e $ome @GME country
-B>CProam. accordin to t$e messae sent by t$e HG< and t$e user is in roamin
no#" t$e 'G< reFects t$is call.
>! t$e user record is set as Call For#ardin /nconditional -CF/." t$e 'G< sends
t$e MS<E to t$e oriinal GMSC to analy(e t$is number and rede!ine t$e routin.
>! no HG< number o! t$e user is !ound and no call !or#ardin is set" ;rror
messae #ill be sent to t$e GMSC.
1.14.% @ain
A!ter receivin t$e >A> !rom t$e GMSC" t$e called MSC sends a
S;EDP>EFAP>=CPCAGG messae to t$e HG< and t$e HG< #ill analy(e t$e called
number and t$e net#or? resource ca)acity to c$ec? #$et$er t$is requirement
is acce)table. >! certain item is not acce)ted" it in!orms t$e callin end t$at
t$e call establis$ment !ails. /nder normal circumstances" t$e HG< sends t$e
MSC a @AG>EG MA@ messae t$at contains t$e location area identi!ication -GA>.
and t$e >MS> or +MS> o! t$e called )arty" in!ormin t$e MSC to )er!orm t$e
)ain )rocedure.
9$en t$e MSC obtains t$e GA in!ormation o! t$e MS !rom t$e HG<" it sends all
t$e BSCs in t$is GA t$e )ain messae t$at contains t$e cell list and t$e +MS>
and >MS> in!ormation required !or )ain. +$e >MS> can be used in t$e )ain !or
t$e MS t$rou$ t$e cell )ain c$annel. >n addition" it is also used to con!irm
t$e )ain subc$annel in t$e discontinuous rece)tion )rocessin.
BSC sends t$e @AG>EG CAMMAED to all t$e cells in t$e GA. +$is command
messae contains t$e )ain c$annel rou) number and t$e timeslot number
-obtained by t$e calculation o! t$e last t$ree numbers o! t$e >MS>" t$e total
number o! t$e )ain c$annels" and t$e total number o! t$e )ain timeslots..
9$en t$e cell receives t$is )ain command" it sends t$e @AG>EG <;Q/;S+
messae on t$e )ain c$annel. +$e messae contains t$e >MS> or +MS> o! t$e
user )aed.
>! t$e called MS detects t$e )ain by decodin t$e )ain in!ormation" it sends
a c$annel request to initiate t$e c$annel allocation )rocess. A!ter receivin t$e
immediate assinment command !rom t$e net#or?" t$e MS sends t$e initial
messae o! @AG>EG <;S@AS; on t$e c$annel assined t$rou$ t$e SABM !rame"
and t$en im)lements t$e aut$entication" encry)tion" +MS> reallocation" and
!inally beins t$e call establis$ment )rocess.
1.14.2 Call ;stablis$ment !or t$e Called @arty
A!ter t$e +MS> reallocation is over" t$e MSC sends t$e MS a S;+/@ messae t$at
includes all t$e details required suc$ as t$e service ty)e and t$e callin
number. A!ter receivin t$is messae" t$e called MS con!irms t$e in!ormation
and sends a CAGG CAEF><M;D messae bac? i! t$e service is available. +$e call
con!irmed messae carries t$e )arameters t$at t$e MS selects" suc$ as t$e
c$annel ty)e -!ull rate +C' or $al! rate +C'. and t$e service ty)e.
A!ter receivin t$e call con!irmed messae" t$e MSC sends t$e assinment
command to t$e BSC !or t$e voice c$annel allocation. A!ter t$e assinment
)rocedure is over" t$e called MS sends an AG;<+>EG messae to t$e net#or?
and a rinin )rom)t occurs to t$e called MS. #$en t$e MSC receives t$is
messae" it sends an Address Com)lete Messae -ACM. to t$e callin end. A!ter
receivin t$is messae" t$e callin end ma?es a rin bac? tone as t$e
oriinatin user )rom)ter.
+$e called user $ears t$e rinin and res)onds" and t$en sends a CAEE;C+
messae to t$e MSC. A!ter receivin t$is messae" t$e MSC connects all t$e
transmission lin?s. +$e end,to,end transmission is establis$ed.
1.14.4 +$e >n!luence o! Call +rans!er to <outin
>n t$e su))lementary services" call trans!er $as t$e reatest in!luence on call
routin. +$e call trans!er is mainly caused by Call For#ardin /nconditional
-CF/." Call For#ardin Busy -CFB." Call For#ardin on mobile subscriber Eot
<eac$able -CFE<c." and Call For#ardin on Eo <e)ly -CFE<y.. +$e routin
selection !or eac$ !unction is as !ollo#s*
>. CF/
9$en t$e GMSC sends t$e S;EDP<A/+>EGP>EFA<MA+>AE messae to t$e 'G<" i!
t$e CF/ !unction is available" t$e 'G< sends t$e
S;EDP<A/+>EGP>EFA<MA+>AEP<;S/G+ messae #it$ t$e trans!er number in it
bac? to t$e GMSC !or it to rede!ine t$e routin.
>>. CFB
9$en t$e GMSC !inds t$e HMSC=HG< #it$ t$e MS<E obtained !rom t$e 'G<" but
t$e called end is busy and t$e CFB !unction is available" t$e HMSC=HG<
im)lements t$e call trans!er o! t$e trans!er number and sends it to t$e t$ird
)arty. >! t$e CFB !unction is not available" t$e GESC $andles t$e call directly"
suc$ as )layin t$e user bus$ record.
>>>. CFE<c
+$e routin selection !or t$is !unction is based on $o# t$e net#or? decides t$e
called )arty is not reac$able. +$e )rocessin is di!!erent !or di!!erent criteria.
>! t$e last location reistration o! t$e called user !ails" and t$e 'G< ?ee)s t$e
record o! t$is situation and ?no#s t$e MS is unreac$able" it ma?es t$e CFE<c
decision by itsel!.
>! t$e 'G< does not ?ee) t$e record o! t$is situation" t$e call !lo# continues
until t$e MSC )er!orms t$e )ain !or t$e user and ets no res)onse !rom t$e
user in due time. +$e user is decided not reac$able. +$e MSC !or#ards t$is
call. +$is ?ind o! situation $as many causes. Ane o! t$em is t$at t$e user enters
t$e dead (one or t$e MS is )o#er,o!!" but t$e HMSC $as not made t$e )eriodic
c$ec? on t$e >MS> attac$ed user yet" so it cannot Fude t$e MS status and t$e
)ain !ails. Anot$er cause is t$at t$e MS is in !requent location u)datin on
t$e ede o! t$e GA and cannot res)ond t$e )ain or t$e c$annel request !ails"
#$ic$ leads to )ain timeout.
>! t$e MS is in >MS> detac$ -t$e MS is s#itc$ed o!! or out o! t$e service area !or
a lon time." because t$e detac$ ta is in t$e HG< instead o! t$e 'G<" t$e call
!or#ardin can only be initiated by t$e HMSC=HG<. 9$en t$e HG< )eriodically
deletes t$e lon,term detac$ed >MS> and in!orms t$e 'G<" t$e 'G< need not
contact t$e HG<.
>H. CFE<y
>! t$e )ain o! t$e HMSC !or t$e user succeeds and t$e called end sends t$e
AG;<+>EG messae to t$e system" but t$e called user ma?es no res)onse in due
time and t$e CFE<y !unction is activated" t$e call !or#ardin )rocedure is
initiated.
H. C9 and 'AGD
Call 9aitin -C9. is a su))lementary service. 9$en t$e MSC receives t$e >A>
!rom t$e callin end" i! t$e called user is in anot$er conversation and t$e C9
!unction is enabled" t$e MSC s?i)s t$e )ain )rocedure and directly sends a
S;+/@ messae to t$e MS by usin t$e current sinalin mode. 9$en t$e C9
!unction is enabled" t$e $andover o! t$e t#o calls can be )er!ormed.
9$en t$e CFB and t$e C9 are enabled at t$e same time" t$e C9 is initiated
!irst i! anot$er call is comin. +$e CFB #ill be initiated #$en a t$ird call is
comin.
1.14.1 ;:ce)tional Situations
+$is section only analy(es t$e common abnormal )rocedures. For ot$er
abnormal )rocedures" see YMobile Ariinatin Call ;stablis$ment @rocedure.Y
/)on )ain !ailure" t$e MSC )rom)ts voice in!ormation to t$e callin )arty"
indicatin t$e called MS is outside t$e servin area or cannot be connected. >n
t$is case" trace t$e sinalin on inter!aces A and Abis to c$ec? #$et$er t$e
)ain !ailure is caused by*
Eo @AG>EG CAMMAED at A inter!ace
Eo @AG>EG CAMMAED at Abis inter!ace
Eo @AG>EG <;S@AES; at Abis inter!ace
Eo @AG>EG <;S@AES; at A inter!ace
>. Eo @ain Command at A >nter!ace
+$rou$ sinalin tracin over inter!ace A" t$e MSC is detected t$at it $as not
sent a @AG>EG messae to t$e BSC. >n t$is case" c$ec? t$e data con!iuration
and MS in!ormation in t$e MSC=HG< and 'G< on t$e ESS side. Additionally"
)o#er o!! t$e called MS" )o#er it on and ma?e a test call to c$ec? #$et$er t$e
MS is normal.
C$ec?in user data in HG<
9$en an MS is )aed" t$e MSC Fudes t$e current state o! t$e MS by t$e user
data -includin MS active state" reistered GA" cell in!ormation." and decides
#$et$er or $o# to send t$e @AG>EG messae.
>! t$e MS state $as c$aned -!or e:am)le" t$e MS is s#itc$ed o!!" or $as
entered a di!!erent GA. and $as not reistered in t$e net#or? normally or
u)dated user data in HG<" t$e MS may )robably be unable to be )aed.
>n t$at case" t$e MS only need to initiate a location u)datin )rocedure to
ensure t$at t$e user data in HG< is correct. +$e )eriod o! )eriodic location
u)datin is indicated in system in!ormation. An MSC side" t$ere is also a
location u)datin )eriod -See YGocation u)datin @rocedureY.. +$e t#o
)arameters o! BSC and MSC must satis!y a certain relations$i)" #$ic$ requires
t$at MS must initiate a location u)datin )rocedure #it$in t$e )eriod s)eci!ied
in MSC.
C$ec?in <A, or Cell,<elated )arameter settins in MSC
>! a routin area or cell related )arameter is incorrectly set in t$e MSC" t$e
transmission o! t$e @AG>EG messae may !ail. For e:am)le" i! a #ron taret
BSC is selected" t$e @AG>EG messae t$at s$ould $ave been sent to t$e local
BSC #ill be sent to anot$er BSC.
>>. Eo @ain Command at Abis >nter!ace
/)on receivin t$e @AG>EG messae !rom t$e MSC" t$e BSC detects t$at t$e
MSC $as not sent @AG>EG CAMMAED to t$e B+S over inter!ace Abis. >n t$is case"
c$ec? t$e o)erations and data con!iuration in t$e BSC
C$ec?in i! !lo# control is enabled
C$ec? i! t$e system load suddenly increases due to centrali(ed transmission o!
s$ort messaes or mass access bursts.
C$ec?in relevant data con!iuration
C$ec? i! t$e CG> in!ormation in BSC data con!iuration is consistent #it$ t$e
GAC in!ormation in t$e @AG>EG messae over A inter!ace. Additionally" i! <A, or
cell,related )arameter is not correctly set in t$e MSC" !or e:am)le" a #ron
taret BSC is selected" t$e @AG>EG CAMMAED messae cannot be success!ully
sent over Abis inter!ace.
C$ec? #$et$er t$e !ollo#in )arameters in t$e [System in!ormation table\ are
correctly set* YBSPAGPBG&SP<;SY" YCCC',CAEFY and YBSP@APMF<MSY.
>>>. Eo @ain <es)onse at Abis >nter!ace
+$rou$ sinalin tracin over Abis inter!ace" t$e BSC is detected t$at it $as
not received t$e ;stablis$ment >ndication -@AG>EG <;S@AES;. a!ter sendin
@AG>EG CAMMAED to t$e B+S. >n t$is case" c$ec? t$e relevant data
con!iuration and radio sinal coverae.
C$ec? i! t$ere is @C' or AGC' overload due to centrali(ed s$ort messae
transmission or mass access bursts.
C$ec? t$e called MS or S>M in it.
C$ec? B+S by ma?in test calls in a di!!erent cell.
C$ec? data con!iuration in BSCC$ec? #$et$er t$e !ollo#in )arameters in t$e
[System in!ormation table\ are correctly con!iured* YBSPAGPBG&SP<;SY" YCCC',
CAEFY" YBSP@APMF<MSY" Y+:,inteer"Y and YMS MAO retransY. C$ec? t$e settin
!or Ylocation u)datin )eriodY in BSC and t$at in MSC
C$ec? radio sinal coverae
Due to t$e )roblem o! radio sinal coverae" t$ere mi$t be some blind
coverae areas. +$e MS t$at $as entered a blind coverae area cannot receive
t$e @AG>EG <;Q/;S+ messae. >n t$at case" t$e MS cannot be )aed. Suc$
cases" i! any" only e:ist in )artial areas.
>H. Eo @ain <es)onse at A >nter!ace
+$rou$ sinalin tracin at Abis inter!ace" t$e BSC is detected t$at it $as
received an ;stablis$ment >ndication -@AG>EG <;S@AES;. messae !rom t$e B+S
but t$is messae is not re)orted over inter!ace A.
1.11 'A
As a ?ey tec$noloy in t$e cellular mobile telecommunication system" $andover
-'A. can reduce t$e call dro) rate and t$e net#or? cross inter!erence. +$e
$andover )rocedure consists o! $andover trier" $andover )re)aration and
decision" and $andover e:ecution.
'A can be divided into sync$ronous 'A and async$ronous 'A based on +imin
Advance -+A.. Sync$ronous 'A means t$e t#o cells are sync$roni(ed #it$ eac$
ot$er and t$e MS can calculate t$e ne# +A -t$e 'A command indicates
#$et$er t$e 'A is sync$ronous or not.. Async$ronous 'A requires t$e B+S to
calculate t$e ne# +A. 9$en t$e MS receives t$e 'A command and requests !or
t$e ne# B+S access" t$e ne# B+S in!orms t$e MS o! t$e calculated +A. +$e MS
access to t$e ne# c$annel can also be divided into !our ty)es* sync$ronous"
async$ronous" )re,sync$ronous" and )seudo,sync$ronous. +$e !irst t$ree ty)es
are required in MS and t$e last one is o)tional. +$e )seudo,sync$ronous 'A can
be )er!ormed only #$en t$e MS su))orts t$is !unction. >n t$e )seudo,
sync$ronous 'A" t$e $andover command !rom t$e B+S o! t$e oriinal service
cell contains t$e <+D value -t$e +A di!!erence bet#een t$e source B+S and t$e
taret B+S.. +$e MSC calculates t$e +A required !or t$e access to t$e ne# B+S
based on t$e <+D value.
+$e 'A )rocess involves MS" B+S" BSC" and MSC. Accordin to t$e location
#$ere t$e 'A $a))ens" t$e 'A can be divided into intra,cell 'A and inter,cell
'A. +o be more s)eci!ic" intra,cell 'A" intra,B+S 'A" intro,BSC 'A" intra,MSC
'A" and inter,MSC 'A. +$e !unction o! eac$ unit is* MS measures t$e do#nlin?
)er!ormance and t$e sinal strent$8 B+S monitors t$e received sinal level
and quality o! t$e u)lin? and t$e inter!erence level o! t$e idle tra!!ic c$annel8
BSC $andles t$e measurement re)ort and ma?es t$e 'A decision8 MSC decides
t$e taret cell o! t$e inter,BSC 'A.
1.11.1 'A @re)aration
>. Measurement <e)ort
+$e 'A decision de)ends on t$e measurement re)ort -M<. sent by MS t$rou$
u)lin? SACC' to t$e net#or? and t$e M< o! t$e u)lin? sent by B+S. +$ese t#o
re)orts are sent to BSC at t$e same time !or decision. +$e system in!ormation
t$at includes t$e )arameters o! t$e current cell and t$e nei$bor cell are sent
to t$e MS under t$e dedicated mode t$rou$ t$e do#nlin? SACC'. +$e MS
re)orts t$e <OG;H and quality" +A value" )o#er control" and D+O usae to t$e
net#or? accordin to t$e system in!ormation. >n addition" t$e MS also )er!orms
t$e )seudo,sync$roni(ation #it$ t$e nei$bor cell de!ined by t$e system !or 'A
and measures t$e <OG;H !rom t$e BCC'. +$e MS measures all t$e !rames
e:ce)t t$e idle !rames t$at are used to sync$roni(e t$e nei$bor cell and
decode SC'. +$e MS re)orts t$e condition o! t$e cell and t$e si: nei$bor cells
#it$ t$e stronest <OG;H it measures durin t$e measurement )eriod to t$e
system !or t$e 'A decision.
Measurement )eriod
+$e SACC' measurement )eriod is di!!erent i! t$e MS occu)ies di!!erent
c$annel under t$e dedicated mode.
0>! t$e SACC' is associated #it$ SDCC'" t$e measurement )eriod is 450ms"
because a com)lete SACC' messae bloc? occu)ies t#o 11 multi!rames o!
SDCC'.
0>! t$e SACC' is associated #it$ +C'" t$e measurement )eriod is 480 ms"
because a com)lete SACC' messae bloc? occu)ies !our %3 multi!rames o!
+C'.
A com)lete M< consists o! !our continuous SACC' bursts. An t$e SDCC'" t$e
!our bursts are transmitted continuously. An t$e +C'" eac$ %3 multi!rame $as
only one SACC' burst" so a com)lete M< requires !our %3 multi!rames.
Fiure 1,1 Measurement )eriod
9$et$er to use D+O or not" t$e M< $as t#o values* !ull measurement value and
sub measurement value. For details" see t$e D+O descri)tion in C$a)ter %.
M< )rocessin
B+S $andles t$e u)lin? M< it ma?es and t$e do#nlin? M< it collects !rom t$e
MS. >t obtains t$e sam)le values o! t$e <OG;H" <OQ/AG" and +A" and t$en
calculates t$e arit$metical mean value and t$e #ei$ted mean value based on
t$e related )arameters. 9$en t$e time is u)" t$e system decides #$et$er to
)er!orm t$e level $andover" quality $andover" or distance $andover.
>>. Eei$bor Cell Monitorin
+o establis$ t$e 'A relation #it$ t$e nei$bor cells" t$e MS must listen to t$e
standard !requency o! t$e nei$bor cells de!ined in t$e system messae. +$e
standard !requency carries t$e sync$ronous c$annel and !requency correction
c$annel. Ane #ay to decide t$e received c$annel is t$e standard !requency
c$annel is to con!irm t$at t$e !requency carries a FCC'. +$e MS also decodes
t$e SC' t$at carries t$e +DMA !rame number and BS>C. +$e MS can only
analy(e t$e BCC' standard !requency o! t$e nei$bor cell in t$e idle timeslot
o! t$e +C' multi!rame. >n !act" durin t$e data e:c$ane" t$e interval bet#een
t$e end o! t$e rece)tion and t$e beinnin o! t$e transmission -about 1 ms.
can be used to measure t$e <OG;H and t$e <OQ/AG" but it is not su!!icient to
measure t$e level o! t$e nei$bor cell. +$e interval bet#een t$e end o! t$e
transmission and t$e beinnin o! t$e rece)tion -about % ms. is su!!icient to
measure t$e level o! t$e nei$bor cell" but not su!!icient to !ind t$e FCC'. >n
t$e %3 muli!rame o! +C'" t$ere is al#ays an idle !rame -about 3 ms. available
!or MS to decode t$e FCC' and SC'. But t$e FCC' o! t$e nei$bor cell may not
be !ound durin t$is timeslot. +$ere!ore" t$e use o! t$e arit$metic !eature o!
t$e t#o numbers %3 and 11 is required. Because t$ese t#o numbers $ave no
common !actor" t$e FCC' can be !ound durin t$e 11 )eriods. 9$en SACC' is
associated #it$ SDCC'" alt$ou$ its )eriod is also 11 multi!rame" t$e SDCC'
c$annel assined to t$e MS only occu)ies 1=8 o! t$e 11 multi!rame. Since t$ere
are lots o! idle timeslots" t$e MS can sync$roni(e t$e nei$bor cell.
9$en t$e MS receives t$e SC'" t$e sync$roni(ation is establis$ed. +o translate
t$e messae on t$e do#nlin? CSC'" t$e MS must ?no# t$e trainin sequence o!
t$e CSC'. +$e trainin sequence is o! ei$t ty)es" matc$in t$e BCC 0 to BCC 5
o! BS>C res)ectively. +$e BS>C carried by t$e SC' can in!orm t$e MS o! t$e
trainin sequence number o! its service cell.
BS>C also enables t$e MS to di!!erentiate t$e cells usin t$e same BCC'
!requency. +$e t#o cells #it$ t$e same BCC' !requency and BS>C must be !ar
!rom eac$ ot$er. +$e MS re)orts t$e si: nei$bor cells #it$ t$e stronest
sinals" but di!!erentiates t$em accordin to t$e BS>C and !requency it obtains
to ac$ieve t$e )re,sync$roni(ation. +$e M< only contains t$e sequence number
o! t$e !requencies in t$e BA list. +$ere!ore" i! a cell s$ares t$e same !requency
and BS>C #it$ t$e nei$bor cell and its sinal is stron enou$" error re)ort and
decision o! MS may occur" leadin to 'A !ailure and call dro).
>>>. Conditions <equired !or Eei$bor Cells to Xoin in 'A Decision Queue
9$en t$e B+S receives t$e re)ort on t$e nei$bor cell !rom t$e MS" it c$ec?s
#$et$er t$is nei$bor cell is quali!ied to Foin in t$e 'A decision queue. +$e
!ollo#in conditions must be met*
<OG;H-n. S <:GevMinCell-n.J MAO-0"@a-n.. J AFFS;+ -%,4.
@a-n.NMSP+O@9<PMAO-n. MAOP@A9;<PAFPMS
<OG;H-n. is t$e <OG;H o! t$e nei$bor cell8 <:GevMinCell-E. is t$e minimal
access level o! t$e nei$bor cell8 AFFS;+ is t$e o!!set o! t$e minimal access
level8 MSP+O@9<PMAO-n. is t$e ma:imal transmit )o#er o! MS de!ined by t$e
system8 MAOP@A9;<PAFPMS is t$e ma:imal transmit )o#er t$e MS can
ac$ieve. +$e unit is dBm.
<:GevMinCell-n. and MSP+O@9<PMAO-n. are de!ined by t$e 'A cell )arameters.
/nder t$e dedicated mode" t$e system in!orms t$e MS by sendin t$e system
messae t$rou$ SACC'. +$e nei$bor cell can be listed in t$e 'A candidate
cells only #$en its <OG;H is quali!ied accordin to t$e !ormula above.
+$e de!ined <:GevMinCell -n. must be $i$er t$an t$e <OG;HPACC;SSPM>E. >! it
is too lo#" t$e t$res$old !or t$e candidate cells is reduced" #$ic$ may lead to
'A !ailure. +$e )ur)ose to de!ine t$e @a is to ensure t$e lo# )o#er MS can
access t$e nei$bor cell only #$en t$e <OG;H is $i$ enou$" t$us im)rovin
t$e quality o! conversation.
1.11.% 'A +y)es
'A must be )er!ormed on time under di!!erent conditions to ensure t$e quality
o! communication. Accordin to t$e cause o! t$e 'A" it can be divided into
@o#er Budet -@BG+. 'A" ede 'A" bad quality -BQ. 'A" direct retry" and
timin advance -+A. 'A.
>. @BG+ 'A
@BG+ 'A is based on )at$ loss. @BG+ 'A alorit$m loo?s !or a cell #it$ less
)at$ loss to decide #$et$er 'A is necessary. +$e biest di!!erence bet#een
t$e @BG+ 'A and ot$ers is t$at t$e trierin condition is )at$ loss but not
receivin )o#er.
+$e !ormula o! @BG+ 'A is as !ollo#s*
@BG+ -n. S @GB+P'oPMarin -n. -%,1.
@BG+-n. N - BS+OPMAO , <OG;HPDG , @9<PCPD . , - BS+OPMAO-n.,
<OG;HPEC;GG-n. ., - <OG;HPDG , <OG;HP/G , S;ES>PCA<<;C+., ma:
- BS+OPMAO-n., min-MS+OPMAO-n."@. , BS+OPMAO J min -MS+OPMAO"@. "0 .
BS+OPMAO* +$e ma:imum transmit )o#er o! BS in service cell
BS+OPMAO -n.* +$e ma:imum transmit )o#er o! BS in nei$bor cell
<OG;HPDG* +$e do#nlin? received sinal level in service cell
<OG;HP/G* +$e u)lin? received sinal level in service cell
S;ES>PCA<<;C+* +$e correct !actor o! MS=BS receiver sensitivity
<OG;HPEC;GG -n.* t$e received sinal level o! MS !rom nei$bor cell n
@9<PCPD* t$e decrease o! t$e transmission )o#er in B+S )o#er control
@* Ma: MS +ransmission )o#er
MS+OPMAO -n.* Ma: MS transmit )o#er allo#ed o! t$e nei$borin cell n
MS+OPMAO* Ma: MS transmit )o#er allo#ed o! t$e service cell
+$e nei$bor cell #it$ t$e biest @BG+ -n. is selected as t$e taret cell !or
'A. +$e @GB+P'oPMarin is t$e de!ined <OG;H di!!erence value bet#een t$e
service cell and t$e nei$bor cell #$en t$e 'A is initiated. >! t$is value is too
lo#" it may lead to )in,)on $andover8 i! it is too $i$" 'A $ysteresis may
occur and t$e 'A e!!iciency is reduced. Since t$e @GB+P'oPMarin is de!ined
!or t$e s)eci!ic nei$bor cell" t$e tra!!ic load can be adFusted accordinly. For
e:am)le" #$en cell A and cell B are adFacent" A is t$e $i$,tra!!ic cell and B is
t$e lo#,tra!!ic cell" t$e call distribution can be balanced by reducin t$e
@GB+P'oPMarin !rom A to B and increasin t$at !rom B to A. >n !act" t$is #ay
to balance t$e call distribution equals t$e decrease o! t$e coverae area !or
cell A and t$e increase o! t$e coverae area !or cell B.
@BG+ 'A only $a))ens bet#een t$e )eer cells. .
>>. ;de 'A
+$e u)lin?=do#nlin? ede 'A marin is de!ined in t$e 'A )arameters. 9$en
BSC !inds in t$e M<s !rom t$e MS and B+S t$at t$e u)lin? or do#nlin? <OG;H is
lo#er t$an t$e ede 'A marin de!ined" it selects a )ro)er nei$bor cell !rom
t$e M<s as t$e taret cell to initiate 'A" t$us avoidin t$e call dro).
>n t$e ede 'A" t$e <OG;H o! t$e nei$bor cell s$ould be $i$er t$an t$at o!
t$e service cell by a certain value. +$is value is called t$e ede 'A marin.
+$is alorit$m is also used to avoid )in,)on $andover. +$e ede 'A marin
s$ould be $i$er t$an t$e minimal access level o! t$e MS.
>>>. BQ 'A
+$e decision mec$anism o! BQ 'A is similar to t$at o! t$e ede 'A. 9$en BSC
!inds in t$e M<s !rom t$e MS and B+S t$at t$e bit error rate o! t$e u)lin? or
do#nlin? is $i$er t$an t$e BQ 'A marin de!ined" t$e BQ 'A is initiated. +o
!urt$er di!!erentiate t$e BQ 'A" t$e inter!erence 'A is introduced. >! t$e
<OG;H is $i$er t$an t$e de!ined <OG;H marin o! t$e inter!erence 'A and t$e
<OQ/AG is $i$er t$an t$e quality 'A marin" t$e !requency inter!erence
e:ists. +$e inter!erence 'A #ill trier t$e intra,cell 'A -#$en t$e intra,cell
'A is available. !irst to im)rove t$e bad conversation quality due to
inter!erence" and t$en trier t$e inter,cell 'A. +$e intra,cell 'A is not
e!!ective #$en t$e !requency $o))in is used. By im)rovin t$e inter!erence
'A marin" t$e BQ 'A #ill be mainly )er!ormed bet#een cells.
>H. Direct <etry
Durin t$e call establis$ment" t$e SDCC' is assined !irst and t$en is t$e +C'.
>! t$e service cell $as no idle +C'" t$e call attem)t usually !ails because o! +C'
conestion. +o !ully utili(e t$e radio resources and reduce t$e conestion" t$e
direct retry !unction is introduced. 9$en t$e SDCC' is assined" but no +C' is
available" t$e assinment request is sent in t$e !orm o! M< and t$e call is
accessed to t$e idle s)eec$ c$annel. A!ter t$e direct retry !unction is enabled"
t$e queuin !unction can be activated to )rovide enou$ time !or t$e system to
select t$e nei$bor cell available !or direct retry.
H. +A 'A
+A 'A can be used to control t$e coverae area o! t$e B+S. 9$en t$e BSC !inds
t$e +A value re)orted by t$e MS is $i$er t$an t$e de!ined marin" t$e +A 'A
is initiated. >! t$e +A marin is relatively lo#" t$e !requent )in,)on $andover
may be triered. +$ere!ore" s)ecial attention s$ould be )aid to t$e matc$in
o! di!!erent ?inds o! 'A.
1.11.2 'A @rocess Analysis
>. >ntra,Cell 'A
>n t$e real net#or?" sometimes t$e inter!erence may occur to certain
!requency or a certain +<O !ails" leadin to t$e $i$ <OG;H but lo# <OQ/AG or
t$e remar?ably lo# sinal level o! +<O. +o im)rove t$e conversation quality
and avoid t$e call dro)" t$e intra,cell 'A is used.
+$e intra,cell 'A is initiated by t$e <OG;H marin or <OQ/AG quality. Durin
t$e conversation" BSC analy(es t$e M< !rom t$e MS and B+S. >! t$e requirement
!or intra,cell 'A marin is satis!ied" it sends a C'AEE;G AC+>H; messae to B+S
to initiate t$e intra,cell 'A. +$e connection )rocess is similar to t$e +C'
assinment durin t$e call establis$ment. Because t$e +C' is also assined
#it$in t$e cell" t$e B+S can indicate t$e MS to )er!orm t$e intra,cell 'A
t$rou$ 'A command or assinment command. 9$en t$e BSC receives t$e
ASS>GEM;E+ CAM@G;+;='AEDAH;< CAM@G;+; messae !rom t$e B+S" it sends
MSC t$e 'A @;<FAM;D messae t$at contains t$e 'A ty)e. +$en t$e BSC sends
a <F C'AEE;G <;G;AS; messae to B+S. A!ter receivin t$e messae" t$e B+S
releases t$e +C' resource and sends a <F C'AEE;G <;G;AS; AC& messae
bac?.
9$en t$e intra,cell 'A is enabled" intra,cell 'A increases a lot" and t$e system
load also increases. +$ere!ore" i! t$e tra!!ic load is already $eavy" t$e intra,
cell 'A !unction is not recommended.
>>. >ntra,BSC 'A
>ntra,BSC 'A is )er!ormed by BSC and no MSC $as to be involved. +o in!orm
MSC t$at t$e 'A is com)lete" BSC #ill send a 'A @;<FAM;D messae to MSC.
1. +$e MS sends M< to B+S1 on SACC' at /m inter!ace" and B+S1 !or#ards t$e
messae to t$e BSC.
%. BSC receives t$e M<. >! it decides t$at t$e MS s$ould be $anded over to
anot$er cell" it sends C$annel Activation to B+S% o! t$e taret cell to activate
t$e c$annel.
2. B+S% receives t$e C'AEE;G AC+>HA+;. >! t$e c$annel ty)e is correct" it turns
on t$e )o#er am)li!ier on t$e s)eci!ied c$annel to receive in!ormation in t$e
u)lin? direction" and send C'AEE;G AC+>HA+; AC& to t$e BSC.
4. A!ter receivin t$e C'AEE;G AC+>HA+; AC& !rom B+S%" t$e BSC sends
'AEDAH;< CAMMAED to t$e MS t$rou$ B+S1 and starts +2102. +$e $andover
command contains all t$e !eature in!ormation o! t$e transmission on t$e ne#
c$annel and t$e data required !or MS access. >t also indicates #$et$er t$is 'A
is sync$ronous or async$ronous.
1. A!ter receivin t$e 'AEDAH;< CAMMAED" t$e MS decides t$e ty)e o! it. >! it
is sync$ronous 'A" t$e MS sends t$e taret cell !our continuous 'AEDAH;<
ACC;SS messaes on t$e assined +C'" and t$en starts t$e transmission based
on t$e calculated. For t$e sync$ronous 'A" t$e !ormer +A can be used8 !or )re,
sync$ronous 'A" t$e +A in t$e $andover command is used ->! t$e +A is not
)rovided in t$e $andover command" t$e de!ault value is used.8 !or )seudo,
sync$ronous 'A -MS re)orted #$et$er t$is 'A is su))orted or not be!ore." t$e
+A is calculated based on t$e di!!erence value )rovided in t$e $andover
command. @lease note t$at t$e 'AEDAH;< ACC;SS is send by t$e access burst.
>t is t$e only time #$en t$e access burst is used on t$e DC'. >t only contains
t$e 8,bit 'A re!erence number obtained !rom t$e $andover command. Since
t$is re!erence number is ?no#n to t$e taret cell" t$e taret cell can c$ec?
#$et$er t$e access request is !rom t$e e:)ected MS #it$ t$is number.
+$e 'A re!erence number is not !ully de!ined in t$e )rotocol. Durin t$e 'A
access" i! t$e assined +C' is on t$e BCC'" due to sync$roni(ation error and
delay or ot$er reasons" t$e access burst may o!!set to t$e BCC' <AC' timeslot.
>! t$e 8,bit re!erence number is t$e same as a service a))lication number" t$e
system #ill reard it as a random access by mista?e and assin t$e SDCC'
t$rou$ AGC'" leadin to a #aste o! AGC' and SDCC'. But as t$e access burst
contains t$e BS>C in!ormation" only t$e 'A access cell #ill be a!!ected.
Since t$ere are more t$an !our 'A access bursts" and a!ter t$e ne# BSS assins
a c$annel to t$e MS" it #ill no re,assin t$is c$annel to ot$er MS" even i! no
re!erence number is used" t$e net#or? can !ind t$e MS to access and t$e 'A
#ill not be a!!ected.
+o !urt$er avoid t$e #aste o! radio resources" t$e re!erence number is assined
a !i:ed value t$at is di!!erent !rom t$e a))lication number !or service ty)e in
random access.
3. B+S% receives t$e 'AEDAH;< ACC;SS !rom t$e MS" and send 'AEDAH;<
D;+;C+ to t$e BSC noti!yin t$at t$e 'AEDAH;< ACC;SS messae is received.
5. For async$ronous 'A" a!ter t$e B+S% c$annel o! t$e taret cell is activated"
it #aits !or t$e MS access on t$e assined DC' -until t$e +2102 times out..
9$en it detects t$e $andover access !rom t$e MS" t$e B+S% sends t$e 'A
D;+;C+ messae to t$e BSC and t$e @'RS>CAG >EFA t$at contains t$e
calculated +A to t$e MS. Durin t$e @'RS>CAG >EFA transmission" t$e net#or?
initiates +2101. Be!ore receivin t$e SABM !rame res)onse !rom t$e MS" t$e
B+S% re,enables t$e +2101 a!ter timeout and resends t$e @'RS>CAG >EFA ER1.
For async$ronous 'A" a!ter receivin t$e @'RS>CAG >EFA" t$e MS sends t$e
SABM to t$e B+S%8 !or sync$ronous 'A" t$e MS sends t$e SABM to t$e B+S%
immediately a!ter sendin t$e 'AEDAH;< ACC;SS.
8. For async$ronous 'A" t$e MS starts t$e +21%4 #$en sendin t$e 'AEDAH;<
ACC;SS messae !or t$e !irst time and sto)s t$e +21%4 a!ter receivin t$e
@'RS>CAG >EFA. For details" see t$e )arameter descri)tion section.
9. A!ter receivin t$e !irst SABM" B+S% sends BSC t$e ;S+ >ED to in!orm it o! t$e
radio lin? establis$ment. 9$en t$e net#or? receives t$is messae" it sends an
;S+ABG>S'; >ED>CA+>AE messae to t$e BSC to s$o# t$at t$e data lin? layer is
establis$ed. Mean#$ile" it also sends t$e /A res)onse !rame to t$e MS. a!ter
receivin t$e /A res)onse" t$e MS reards t$at t$e sinalin ans#er mode is
establis$ed #it$ t$is cell.
10. +$e MS sends 'AEDAH;< CAM@G;+; to t$e B+S%" and B+S% !or#ards it to
t$e BSC. +$en it sends t$e taret cell a 'AEDAH;< CAM@G;+; messae t$at
only contains t$e $andover com)lete indication but no ot$er in!ormation. +$e
MS sto)s considerin t$e )ossibility to return to t$e !ormer c$annel only #$en
t$is messae is sent. >! t$e MS does not receive t$e @'RS>CAG >EFA !rom t$e
taret cell or t$e /A res)onse !rame" it sends a 'AEDAH;< FA>G/<; messae on
t$e source c$annel.
11. A!ter receivin t$e 'AEDAH;< CAM@G;+; messae" t$e BSC sto)s t$e +2102
and sends MSC t$e 'AEDAH;< @;<FA<M;D t$at contains t$e $andover ty)e.
Mean#$ile" t$e BSC initiates t$e local release !or t$e !ormer c$annel o! B+S1.
9$en t$e taret cell receives t$e $andover com)lete messae !rom t$e MS" it
!or#ards it to t$e BSC. A!ter receivin t$is messae" t$e BSC sends t$e <F
C'AEE;G <;G;AS; messae to in!orm t$e source cell to release t$e !ormer
+C'. 9$en t$e source cell receives t$is re)ort" it sends a <F C'AEE;G <;G;AS;
AC& to indicate t$e radio c$annel is released and available !or anot$er
assinment.
>>>. >ntra MSC 'A
Com)ared #it$ t$e intra,BSC 'A )rocedure" t$e )rocedure !or t$e inter,BSC
'A only $as several A inter!ace sinalin added.
1. 9$en t$e MS $as to be $anded over to t$e cell #$ere t$e BSC% belons to"
t$e BSC1 sends a 'A <;Q/><;D messae t$at contains cell >D o! t$e taret cell
rou) and t$e source cell and t$e 'A cause to t$e MSC and starts +5 at t$e
same time.
%. A!ter t$e MSC receives t$is messae" i! it s$ares t$e same GAC #it$ t$e
taret cell" it searc$es t$e BSC o! t$e taret cell -BSC%. and sends t$e BSC% a
'AEDAH;< <;Q/;S+ messae t$at contains t$e in!ormation o! t$e taret cell
and t$e source cell" transmission mode" encry)tion mode" classmar?" and t$e
c$annel ty)e required. 9$en t$e BSC% receives t$is messae" it sends MSC a CC
messae to indicate t$at t$e connection bet#een t$e MSC and its SCC@ is
establis$ed !or transmission o! t$e in!ormation !rom t$e A inter!ace.
2. A!ter t$e ne# c$annel is activated" t$e BSC% sends t$e MSC a 'A <;Q/;S+
AC& to indicate t$at t$e c$annel is available. +$is messae carries t$e 'A
command #it$ t$e in!ormation about t$e resource allocation in it to s$o# t$at
t$e local end is ready !or 'A.
4. A!ter receivin t$e 'A <;Q/;S+ AC&" t$e MSC sends a 'A CAMMAED to t$e
BSC1. BSC1 sto)s t$e +5 and starts t$e +8" and !or#ards t$e 'A CAMMAED to
t$e MS and starts +2102" in!ormin t$e MS to access t$e ne# c$annel. +$is
command contains t$e cell >D" c$annel ty)e" and 'A re!erence.
1. A!ter receivin t$e 'A CAM@G;+; !rom t$e BSC%" MSC sends a CG;A<
CAMMAED to t$e BSC1. +$is command contains t$e clear cause -suc$ as 'A
clear.. BSC1 sto)s +8 and +2102" and releases t$e !ormer c$annel. Mean#$ile"
it sends a CG;A< CAM@G;+; messae to t$e MSC.
+2102 is started #$en BSC sends t$e 'A command and cleared #$en t$e BSC
receives t$e 'A CAM@G;+; ->E+<A BSC. or CG;A< CAMMAED ->E+;< BSC.. +$e
+2102 s$ould be set less t$an +8. Durin t$e 'A" t$e BSC )rovides t$e time !or
+C' bot$ in t$e source cell and t$e taret cell accordin to t$e +2102. 9$en
t$e +2102 is timin" t#o c$annels are reserved. +$e lonest 'A ->E+;< MSC.
may ta?e about !ive seconds" so t$e +2102 can be set to !ive seconds. >! it is
set too lon" t$e system resources #ill be #asted.
>! t$e taret cell and t$e source cell are not in t$e same GA" a location
u)datin #ill be )er!ormed at t$e end o! eac$ call.
>H. >nter,MSC 'A
+$e )rocedure !or inter,MSC 'A is s$o#n in Fiure 1,%3.
1. 9$en MSCa receives t$e 'AEDAH;< <;Q/><;D messae !rom t$e BSC" i! it
!inds t$at t$e GAC o! t$e )re!erred taret cell is not in t$e local GAC list" it
queries t$e remote GAC list t$at contains t$e routin address o! t$e nei$bor
MSC=HG<.
%. 9$en t$e taret MSCb is !ound" t$e MSCa sends a @<;@A<; 'AEDAH;<
messae t$at contains t$e 'AEDAH;< <;Q/;S+ to it.
2. A!ter receivin t$e @<;@A<; 'AEDAH;< messae" t$e MSCb sends t$e HG<b
an AGGACA+;P'APE/MB;< messae to request !or 'A number -'AE.
assinment. +$e 'AE indicates t$e routin bet#een MSCa and MSCb.
4. HG<b selects an idle 'AE and sends it to MSCb t$rou$ t$e S;ED 'A <;@A<+
messae.
1. MSCb establis$es a SCC@ lin? to t$e taret BSC and sends a 'AEDAH;<
<;Q/;S+ messae to BSCB. +$en t$e BSC activates t$e c$annel o! t$e taret
cell. A!ter receivin t$e c$annel activation res)onse !rom t$e taret cell" t$e
BSC sends MSCb a 'AEDAH;< <;Q/;S+ AC& messae t$at contains t$e 'A
command.
3. A!ter receivin t$is messae" MSCb sends a @<;@A<; 'AEDAH;< AC&
messae t$at contains t$e 'AEDAH;< <;Q/;S+ AC& and t$e 'AE to t$e MSCa.
5. MSCa receives t$is messae and sends an >AM to MSCb. +$e >AM contains t$e
'AE assined by HG<b !or MSCb to identi!y #$ic$ s)eec$ c$annel is reserved
!or t$e MS. MSCb sends a S;ED 'A <;@A<+ <;S@ messae to t$e HG<b anytime
a!ter it receives t$e >AM.
8. A!ter MSCa receives t$e ACM !rom t$e MSCb" it sends t$e 'A command to
t$e MS. +$en t$e MS #ill )er!orm t$e 'A access to t$e taret cell.
9. A!ter receivin t$e 'A access messae !rom t$e MS" MSCb sends MSCa a
@<AC;SS ACC;SS S>GEAGG>EG messae to indicate t$at t$e 'A is detected.
10. 9$en t$e taret cell receives t$e 'AEDAH;< CAM@G;+; messae !rom t$e
MS" it in!orms t$e MSCb. +$en t$e MSCb sends a S;ED ;ED S>GEAG <;Q messae
to MSCa to in!orm it t$e 'A is com)lete. A!ter t$e 'A,D;+;C+ or 'A,
CAM@G;+; is received" t$e connection bet#een MSCa and MSCb is establis$ed.
MSCb #ill release t$e 'AE.
11. 9$en MSCa receives t$e 'A com)lete messae" it sends a clear command
to t$e !ormer BSC to release t$e c$annel resource. +$e inter,MSC 'A is
com)lete. +o avoid t$e @S+E=>SDE contradiction o! t$e MSCa and MSCb" MSCb
must send an ans#er sinalin #$en receivin t$e 'A,D;+;C+=CAM@G;+;.
1%. MSCa controls t$e call until it is cleared. 9$en MSCa clears t$e MS call" it
also clears t$e call control !unction o! MSCa and sends a MA@,S;ED,;ED,S>GEAG
messae to release t$e MSCb MA@ resource.
MSCb sends a 'A !ailure indication to t$e MSCa i! t$e MSCb cannot identi!y t$e
taret cell" t$e 'A to t$e taret cell is not allo#ed" t$e taret cell $as no radio
c$annel available" or t$e data error occurs. +$e MSCa #ill )er!orm t$e 'A to
t$e secondary cell or terminate t$e 'A.
H. Subsequent >nter,MSC 'A
A!ter t$e MSCb receives t$e 'A request" it c$ec?s t$is taret cell belons to
MSCb and )er!orms t$e inter,MSC 'A. A!ter t$e 'A is com)lete" it in!orms t$e
MSC.
+$e subsequent 'A is t$e $andover o! MSCb to ot$er MSC a!ter an inter,MSC 'A
is com)lete. +$e taret MSC can be t$e !ormer MSCa or t$e ne# MSCbD. +$e
circuit s#itc$ $a))ens in t$e MSCa !or bot$ situations. A!ter t$e subsequent 'A
is com)lete" t$e connection bet#een MSCa and MSCb is released. +$e
)rocedure !or t$e subsequent 'A #it$ circuit s#itc$ is as !ollo#s*
MSCb is $anded over bac? to MSCa
1. MSCb sends MA@ @<;@A<; S/BS;Q/;E+ 'AEDAH;< request to MSCa. +$is
messae contains MSCa number" taret cell >D" and all t$e in!ormation in 'A
<;Q/;S+.
%. MSCa is t$e call control MSC. >t can searc$ t$e idle c$annel immediately
#it$out taret 'A number routin.
2. A!ter t$e radio c$annel is assined" MSCa sends a MA@ @<;@A<; S/BS;Q/;E+
'AEDAH;< res)onse bac?.
4. >! t$e +C' is busy" BSSa sends a Q/;/>EG >ED>CA+>AE to MSCb -o)tional..
MSC sends MSCb t$e MA@ FA<9A<D ACC;SS S>GEAGG>EG request t$at contains
t$e subsequent +C' assinment result -'A <;Q/;S+ AC& or 'A FA>G/<;.. >! t$e
radio c$annel cannot be assined or t$e error occurs to t$e taret cell >D" or
t$e taret cell >D does not matc$ t$e taret MSC number accordin to t$e 'A
<;Q/;S+" a MA@ @<;@A<; S/BS;Q/;E+ 'AEDAH;< res)onse t$at contains t$e
'A FA>G/<; in!ormation in it is sent to t$e MSCb. MSCb ?ee)s t$e connection to
t$e MS.
1. >! t$e MSCa is success!ully assined" and t$e MA@ @<;@A<; S/BS;Q/;E+
'AEDAH;< res)onse is sent to MSCb. +$e MSCb requests t$e $andover o! t$e MS
to t$e ne# cell o! t$e MSCa by sendin a 'A command.
3. A!ter receivin t$e 'A com)lete messae" MSCa releases t$e circuit
connection to MSCb.
5. MSCa must send a )ro)er MA@ messae to terminate t$e MA@ )rocedure !or
MSCa and MSCb durin t$e basic 'A. 9$en MSCb receives t$e MA@ S;ED ;ED
S>GEAG res)onse messae" it releases t$e BSSb resources.
MSCb is $anded over to MSCb]
Eote 1* +$is messae can be sent anytime a!ter t$e >AM is received.
1. MSCb receives t$e 'A request and !inds t$at t$e taret cell does not belon
to t$e MSCb. >t sends a @<;@A<; S/BS 'AEDAH;< to t$e MSCa. +$is messae
contains t$e MSCbD >D" taret cell >D" and all t$e in!ormation in 'A <;Q/;S+.
MSCa #ill initiate a basic 'A to MSCbD.
%. >! t$e MSC can be !ound in t$e MSCa GAC list and remote GAC list -it contains
in!ormation about ot$er MSC." a!ter t$e 'AE is )rovided by t$e HG<bD and t$e
MSCbD c$annel is activated"
2. MSCa sends a MA@ @<;@A<; S/BS;Q/;E+ 'AEDAH;< res)onse messae to
t$e MSCb. +$is messae contains t$e 'A <;Q/;S+ AC& !rom t$e BSSbD and t$e
BSSMA@ in!ormation t$at may be s)ecial.
4. A!ter receivin t$is messae" MSCb sends t$e 'A command to t$e MS. A!ter
t$e access succeeds" i! t$e MSCa receives t$e MA@ S;ED ;ED S>GEAG <;Q/;S+
-it contains t$e 'A CAM@G;+; in!ormation o! t$e BSSbD. !rom t$e MSCbD" t$e
'A is com)lete and t$e connection bet#een MSCa and MSCb is released. MSCa
also sends t$e MA@ S;ED ;ED S>GEAG res)onse to MSCb to end t$eir MA@
conversation. MSCb receives t$is messae and releases t$e radio resources.
1. A!ter t$e subsequent 'A is com)lete" t$e MSCbD re)laces t$e MSCb. Any
subsequent inter,MSC 'A is t$e same as described above.
+$e remote GAC list o! MSCa must be com)lete and contain as many MSCs as
)ossible besides t$e nei$bor MSC. For e:am)le" i! a user in )lace A calls
anot$er user in )lace B" t$e MSC in )lace A must contains all t$e data o! t$e
MSCs and cells #it$in t$e area bet#een A and B. At$er#ise" t$e 'A cannot be
)er!ormed and t$e call dro)s.
1.11.4 ;:ce)tional Situations
+$e !ollo#in are some e:tra e:ce)tional situations on t$e basis o! #$at $as
described be!ore.
>. 'A Failure Due to C>C ;:ce)tion
>! t$e C>C allocated in t$e 'andover <;Q received by BSC is mar?ed as BGAC&"
BSC #ill res)ond to MSC #it$ 'andover Failure due to Yrequested terrestrial
resource unavailableY.
>>. 'A Failure Due to MS Access Failure
>! t$e B+S cannot decode 'andover Access or 'andover Com)leted correctly
#$en a MS accesses t$e ne# c$annel" t$e 'A #ill !ail. +$e MS returns to t$e
old c$annel" and res)onds #it$ a 'anover Failure messae.
For t$e intra,BSC $andover" i! t$e BSC $as not received t$e 'andover CM@
messae on t$e ne# c$annel" or 'andover Failure messae on t$e old c$annel
at e:)iry o! timer +2102A" it #ill consider t$e call as dro))ed and send a Clear
<;Q messae to t$e MSC on t$e old c$annel. /)on receivin t$e Clear CMD
messae !rom t$e MSC" t$e BSC releases t$e old c$annel and noti!ies t$e taret
cell to release t$e ne# c$annel. >! timer +2102B1 or +2102B% times out" t$e
taret cell #ill release t$e ne# c$annel.
For t$e inter,BSC $andover" i! BSC1 $as not received t$e 'andover CM@
messae at e:)iry o! timer +2102B%" it #ill send a Clear <;Q messae to t$e
MSC to release t$e call. >! BSC% $as not received t$e 'andover D;+ or 'andover
CM@ messae" it #ill send a Clear <;Q messae to t$e MSC !or t$e same
)ur)ose.
1.13 Call <e,;stablis$ment
1.13.1 >ntroduction
+$e re,establis$ment )rocedure allo#s MS to resume a connection in )roress
a!ter a radio lin? !ailure" )ossibly in a ne# cell or in a ne# location area -re,
establis$ment in a ne# location area initiates no location u)datin..
9$et$er call re,establis$ment is allo#ed de)ends on t$e callin status" t$e
cell]s allo#ance o! call re,establis$ment" and activated MM connection -MM is in
status 3 YMM connection activatedY or status %0 Y 9aitin !or additional MM
connectionY Call re,establis$ment can only be initiated by MS. GSM )rotocol
does not s)eci!y t$e im)lementation mode !or t$e s$ort messae service and
t$e inde)endent call su))lementary service. >n t$e ot$er end" no voice is $eard
durin t$e call re,establis$ment.
Durin t$e radio transmission" a connection may be bro?en suddenly because o!
t$e reat transmission loss due to obstructions suc$ as brides" buildins" or
tunnels. 9$en t$e call re,establis$ment is used" t$e MS can maintain t$e
conversation by usin anot$er cell in a s$ort time" t$us im)rovin t$e net#or?
quality. Call re,establis$ment can be rearded as t$e 'A initiated by MS to
save t$e interru)ted call in t$e current cell.
Call re,establis$ment is o! t#o ty)es accordin to t$e entity t$at $as t$e radio
lin? !ailure !irst.
>. <adio Gin? Failure Accurs to MS First
+$e MS sends a call re,establis$ment request in t$e selected cell -source cell or
taret cell.. +$e !ormer c$annel resource is released a!ter t$e B+S timer times
out.
>>. <adio Gin? +imeout Accurs to BSS First
A!ter t$e radio lin? timer in B+S times out" t$e B+S sends a radio lin? !ailure
messae to t$e BSC and BSC activates t$e SACC'. Accordin to t$e )rotocol"
t$e net#or? must $andle t$e conte:t !or a #$ile a!ter detectin t$e lo#er
layer !aults !or t$e success!ul call re,establis$ment. +$e im)lementation mode
and duration are decided by t$e equi)ment )rovider. A!ter detectin t$e radio
lin? !ailure" t$e MS selects a nei$bor cell #it$ t$e $i$est <OG;H #it$in !ive
seconds and sends t$e c$annel request in t$e selected cell. +$is cell s$ould not
be barred and t$e C1 is over 0. >n addition" t$is cell must )ermit t$e call re,
establis$ment. >! all t$e nei$bor cells are not quali!ied" t$e call re,
establis$ment is abandoned.
Durin t$e call re,establis$ment" t$e MS cannot return into t$e idle mode. >!
t$e MS selects a cell in di!!erent GA as t$e taret cell !or call re,establis$ment"
it cannot )er!orm location u)datin until t$e call ends.
/nder normal circumstances" t$e call re,establis$ment )rocedure lasts about 4
to %0 seconds. Most users $ave $un u) t$e )$one be!ore t$e )rocedure is over.
+$ere!ore" t$e call re,establis$ment cannot ac$ieve its oal but #astes a lot o!
radio resources. For t$e areas #it$ limited c$annel resources" t$e activation o!
t$is !unction is not recommended.
1.13.% Call <e,;stablis$ment @rocedure
1. A!ter t$e MM connection !ailure indication is re)orted to t$e CM entity" i!
t$e MS receives at least one request !or MM connection re,establis$ment !rom
CM" it #ill initiate t$e call re,establis$ment )rocedure. >! several CM entities
request !or re,establis$ment" only one re,establis$ment )rocedure #ill be
initiated.
%. A!ter t$e CM sends t$e request !or t$e re,establis$ment o! MM connection"
MM sublayer sends a request !or t$e establis$ment o! << connection and enters
t$e 9A>+ FA< <;;S+ABG>S' state. +$is request includes an establis$ment cause
and a CM re,establis$ment request. 9$en t$e << sublayer indicates a <<
connection is establis$ed -t$e CM re,establis$ment request messae $as been
sent t$rou$ t$e /m inter!ace." t$e MM sublayer starts +2%20 and indicates to
all t$e CM entities t$at t$e MM connection is under construction. +$e MM
sublayer stays in 9A>+ FA< <;;S+ABG>S' state.
+$e CM <e,establis$ment <equest messae contains t$e MS identity ->MS> or
+MS>." Classmar? %" and encry)ted sequence number.
9$et$er t$e CM entity can request !or re,establis$ment de)ends on )rotocol
discriminator -@D..
2. A!ter receivin t$e CM re,establis$ment request" t$e net#or? analy(es t$e
request ty)e and starts t$e MM )roram or << )roram. +$e net#or? can start
t$e classmar? enquiry )roram to obtain more in!ormation about t$e MS
encry)tion ability. +$e net#or? can also decide to )er!orm t$e aut$entication
)rocedure or ci)$erin mode settin )rocedure.
4. 9$en t$e << sublayer indicates t$e ci)$erin mode settin )rocedure is over
or t$e CM S;<H>C; ACC;@+ messae is received" t$e MM connection is re,
establis$ed. +$e +2%20 sto)s and in!orms all t$e CM entities related to t$e re,
establis$ment to enter t$e MM CAEE;C+>AE AC+>H; state.
1. >! t$e net#or? cannot connect t$e re,establis$ment request to t$e current
MS call" it sends t$e CM S;<H>C; <;X;C+ #it$ t$e reFect cause to t$e MS.
+$e reFect cause -value. includes unidenti!iable call -Z28." unidenti!iable >MS>
-Z 4." unaut$ori(ed M; -Z 3." net#or? !ailure -Z15." conestion -Z%%."
unsu))orted service -Z2%." and tem)orary service !ailure -Z24.
3. A!ter receivin t$e CM S;<H>C; <;X;C+" t$e MS sto)s +2%20 and releases all
MM connections and << connections. >! t$e reFect cause i! Z4" t$e MS deletes
t$e +MS>" GA>" and C&SE in S>M card" and c$anes t$e status !rom 6u)datin7
into 6no u)datin7" and t$en enters t$e 69A>+ FA< E;+9A<& CAMMAED7
state. +$e location u)datin #ill be initiated a!ter t$e << release.
>! t$e reFect cause is Z3" t$e MS deletes t$e +MS>" GA>" and C&SE in S>M card"
and c$anes t$e status !rom Yu)datinY into 6roamin in$ibit7. +$e S>M is
rearded invalid until t$e MS is s#itc$ed o!! or t$e S>M card is )ulled out.
1.13.2 ;:ce)tional Situations
>. <e,;stablis$ment @ro$ibition or Failure
9$en MM connection is establis$ed" t$e MM layer may send an indication to t$e
CC layer. >! t$e MM layer is disconnected" t$e connection may be re,establis$ed
t$rou$ CC request.
>! t$e re,establis$ment is not allo#ed" and t$e call is initiated #it$in t$e
establis$ment or clearin )eriod" t$e CC layer s$all release MM connections.
>! re,establis$ment is unsuccess!ul" MM connections s$all be released" and a
release indication s$all be sent to t$e CC layer.
>>. << Connection Failure
>! random access !ailure or << CAEE;C+>AE FA>G/<; is detected by t$e MS" t$e
MS #ill sto) timer +2%20" abort t$e call re,establis$ment )rocedure" and
release all MM connections.
>! << CAEE;C+>AE FA>G/<; is detected by t$e MSC" t$e MSC #ill abort t$e call
re,establis$ment )rocedure and release all MM connections.
>>>. +2%20 +ime,out
>! t$e +2%20 times out" t$e MS #ill sto) call re,establis$ment and release MM
and << connections.
1.13.4 SM @rocedure
S$ort messaes can be transmitted eit$er on SDCC' or SACC'. A s$ort messae
)rocedure can be classi!ied into s$ort messae callin )rocedure and called
)rocedure. For details" see GSM02.40 )rotocol.
1.13.1 S$ort Messae @rocedure on SDCC' 9$en MS is callin
>. Sinalin @rocedure
>>. @rocedure Descri)tion
+$e random access" immediate assinment" aut$entication" and encry)tion
)rocedures o! s$ort messae )rocedure on SDCC' #$en MS is callin are t$e
same as eneral )rocedures. A!ter encry)tion" t$e MS sends SABM aain"
noti!yin t$e net#or? side t$at t$is user needs s$ort messae service -SMS..
+$en" BSC )rovides a trans)arent,transmission c$annel !or MS to e:c$ane
s$ort messae in!ormation #it$ MSC. >n t$is )rocedure" t$e MSCs o! some
manu!acturers are ca)able to send ASS <;Q to BSC" requestin it to assin
c$annel !or s$ort messae transmission. +$e time !or sendin ASS <;Q is t$e
same as t$at !or a common call. BSC can )rovide SMS eit$er by allocatin ot$er
c$annels or by usin t$e oriinal SDCC'.
@oint to @oint s$ort messaes )rotocol is divided into connection manaement
layer -CM." relay layer -<G." trans)ort layer -+G. and a))lication layer -AG..
C@PDA+A and C@PAC& are t$e messaes on CM layer" C@PDA+A is used to
transmit t$e content o! <G and AG messae" and C@PAC& is t$e
ac?no#ledement messae o! C@PDA+A.
+$e release )rocedure a!ter messae is sent is t$e same as eneral ones.
1.13.3 S$ort Messae @rocedure on SDCC' 9$en MS is called
>. Sinalin @rocedure
>>. @rocedure Descri)tion
+$e )ain res)onse and immediate assinment )rocedures o! s$ort messae
)rocedure on SDCC' #$en MS is called are t$e same as eneral )rocedures. For
t$e s$ort messae )rocedure #$en MS is called" a!ter encry)tion" t$e BSC
sends ;S+ <;Q to MS to establis$ s$ort messae connection. 9$en ;S+ CEF is
received !rom MS" t$e connection is success!ully establis$ed. BSC trans)arently
transmits t$e s$ort messae till t$e end o! t$e transmission.
+$e release )rocedure a!ter messae is sent is t$e same as eneral ones.
1.13.5 S$ort Messae @rocedure on SACC' 9$en MS is callin
>. Sinalin @rocedure
>>. @rocedure Descri)tion
+$e MS sends CM S;<H <;Q t$rou$ FACC'. +$e MSC res)onds #it$ t$e CM S;<H
ACC messae and establis$es CC layer connection. +$en" it establis$es << layer
connection on SACC'" and sends t$e s$ort messae.
1.13.8 S$ort Messae @rocedure on SACC' #$en MS is called
>. Sinalin @rocedure
>>. @rocedure Descri)tion
+$e BSC receives t$e C@ DA+A messae !rom MSC" and establis$es an << layer
connection !or SMS. /)on rece)tion o! C@ AC& !rom MS" MSC sends t$e s$ort
messae.
1.15 CBS
Cell Broadcast Service -CBS. is similar to )ain station broadcast in!ormation.
>t means t$e mobile net#or? o)erator broadcasts t$e )ublic in!ormation to t$e
mobile users #it$in a certain area. +$e in!ormation t$at t$e users can read is
called CBS messae. >t is enerated by t$e Cell Broadcast ;ntity -CB;. and sent
to t$e Cell Broadcast Center -CBC. !or )rocessin. A!ter t$e )rocessin" it is
!or#arded to t$e BSC and broadcast to t$e users t$rou$ CBC'. +$e MS can
only receive t$e CBS messae in idle mode. /nli?e t$e @oint to @oint S$ort
Messae service" t$e CBS messae is broadcast #it$out t$e ac?no#ledement
o! t$e user terminal.
CBS includes*
, Common )ublic in!ormation service" suc$ as #eat$er" ne#s" stoc? mar?et"
e:c$ane rate" and lottery.
, S)ecial )ublic in!ormation service" suc$ as )eo)le searc$" tra!!ic naviation"
and call c$are )rom)t.
, Advertisin service" suc$ as in!ormation about stores" restaurants" and
t$eaters.
1.15.1 CBS Mec$anism
A)erators or in!ormation )roviders can de!ine t$e cell broadcast area t$rou$
CB;. +$e minimal area is a cell and t$e ma:imal area can be all t$e cells o! t$e
BSCs t$at t$e CBC connects #it$. Features suc$ as intervals" duration" and
)riority levels can also be s)eci!ied to meet di!!erent requirements. +$e !ield
lent$ o! t$e CBS messae sent to BSC !rom CBC must be 8% bytes. >! t$e lent$
is s$orter t$an 8% bytes" !ill codes are added to it. >! t$e lent$ e:ceeds 8%
bytes" t$e messae is bro?en to a ma:imum o! 11 )aes. >! t$e sendin !ails"
t$e messae may be sent aain and t$e messae #it$ $i$ )riority level is sent
!irst. +$e CBS in!ormation is sent to t$e )ro)er cells t$rou$ !our continuous
SMS B<AADCAS+ <;Q/;S+ messaes or one SMS B<AADCAS+ CAMMAED
messae. ;ac$ CBS messae contains 8%,byte user in!ormation and 3,byte
$eader. +$e CBS messae can be sent to B+S in t$e !orm o! SMS B<AADCAS+
<;Q/;S+ or SMS B<AADCAS+ CAMMAED. For details" see 1.15.%
B+S can send t$e CBC' Goad >ndication messae to BSC and t$e system #ill
s)eed u) or delay t$e messae sendin accordin to t$is messae. Alt$ou$ t$e
BSC considers t$e CBC' ca)acity #$en sendin t$e messae and t$e B+S can
indicate t$e status o! t$e current CBC'" #$en t$e CBC' GAAD >ED>CA+>AE
mode is enabled" t$e B+S can send CBC' GAAD >ED>CA+>AE to request !or
immediate broadcast o! t$e m-111. SMSCB timeslot messae #$en t$e C'CB
is idle. A!ter t$e BSC sends t$e m timeslot messae" it sends messaes
accordin to its o#n sc$edule. >! t$e messae volume t$at t$e B+S requests
e:ceeds t$e volume t$at t$e BSC can )rovide" t$e BSC only sends t$e messaes
#it$in its volume limit. 9$en t$e CBC' GAAD >ED>CA+>AE mode is enabled" t$e
B+S can send CBC' GAAD >ED>CA+>AE to sto) t$e sendin o! t$e m-111.
timeslot messae i! overload occurs. +$en t$e BSC #ill continue t$e sendin
accordin to its o#n sc$edule.
CBC' GAAD >ED>CA+>AE is only used in D<O mode.
+$e CBC' is o! t#o ty)es* basic CBC' and e:tended CBC'. +$ey are !our
continuous multi!rames. +$e +B o! basic CBC' is 0" 1" %" or 28 +$e +B o!
e:tended CBC' is 4" 1" 3" or 5. +B N -FE D>H 11. mod -8..
For t$e basic CBC'" t$e CBS messae $ead is sent on t$e multi!rame #it$ +B
bein 08 !or t$e e:tended CBC'" it is sent on t$e multi!rame #it$ +B bein 4.
+$e system messae on BCC' indicates #$et$er t$e CBS is available or not.
9$en SMSCB is used" t$e BSPAGPBG&SP<;S is set as 1 or above. 9$en t$e CBC'
is ma))ed to t$e CCC'JSDCC'=4" t$e number o! BSPAGPBG&SP<;S #ill not be
limited by SMSCB.
MS recom)oses t$e CBS messae and dis)lays it !or t$e user.
MS obtains t$e CBS messae !rom t$e CBC'. B+S in!orms MS o! t$e s$ort
messae in!ormation durin t$e sc$edule in t$e !orm o! bitma) by sendin
sc$edule messae. +$ere are t$ree rece)tion modes !or MS on CBC'*
, Eon,D<O mode. MS reads t$e !irst bloc? o! all messae timeslots. +$e rest
bloc?s #ill be read i! t$e messae $ead indicates t$at t$e !ollo#in timeslots
are used. >! t$e MS does not su))ort ot$er rece)tion mode" or it does not
receive t$e sc$edulin !or t$e ne:t messae timeslot" Eon,D<O mode is used.
, First D<O mode. >! MS receives t$e sc$edulin !or t$e ne:t messae timeslot"
but t$e !irst sc$edulin messae o! t$e last sc$edulin )eriod" or all t$e
in!ormation o! t$e last )eriod or even earlier )eriod is not received" !irst D<O
mode is used.
, Second D<O mode. >! MS receives t$e im)ortant in!ormation o! t$e last
sc$edulin )eriod and reads t$e !irst sc$edulin messae o! t$e current )eriod"
second D<O mode is used.
9$et$er t$e net#or? uses D<O to receive t$e broadcast s$ort messae can be
set t$rou$ t$e maintenance console in BSC.
1.15.% BSC,B+S Messae +ransmission Mode
A CBS messae consists o! ei$ty ei$t 8,bit bytes. +$ese bytes are divided into
!our messae bloc?s #it$ eac$ bloc? containin t#enty t#o 8,bit bytes. ;ac$
bloc? is added by an 8,bit bloc? ty)e" and t$e lent$ o! t$e bloc? is t#enty
t$ree 8,bit bytes. A CBS messae contains !our continuous bloc?s* !irst bloc?"
second bloc?" t$ird bloc?" and !ourt$ bloc?.
9$en t$e SMS B<AADCAS+ <;Q/;S+ mode is used" t$e messae is sent to B+S
!rom BSC. +$e BSC $andles t$e queuin" re)etition" and s$ort messae sendin.
>t also considers t$e CBC' ca)acity and ta?es c$are o! t$e SMS sementation
at radio inter!ace. >n t$e SMS B<AADCAS+ <;Q/;S+ messae" eac$ SMSCB
>n!ormation cell carries a com)lete !rame t$at can be transmitted on CBC' and
t$e layer % in!ormation t$at indicates t$e radio )at$. SMSCB C$annel >ndicator
cell indicates t$e C'C' used !or broadcast. >! t$is cell does not )rovide t$e
in!ormation" t$e basic CBC' #ill be used.
9$en t$e SMS B<AADCAS+ CAMMAED mode is used" SMS B<AADCAS+ CAMMAED
messae is sent to B+S !rom BSC. BSC requires t$e immediate messae sendin
durin t$e ne:t CBC' time. +$e de!ault broadcast mode !or B+S can also be set
t$rou$ t$is messae. >n t$e de!ault broadcast mode" i! t$ere is no ot$er
messae to broadcast" B+S #ill send t$e de!ault messae.
>n t$e SMS B<AADCAS+ CAMMAED messae" t$e SMSCB messae cell contains
t$e in!ormation to be broadcast on CBC'. >t $as !our continuous bloc?s #it$ a
ma:imum o! 88 bytes. B+S sements t$e messae and establis$es t$e bloc?
!ormat. >t also adds bytes to t$e bloc? i! required. SMSCB C$annel >ndicator cell
indicates t$e C'C' used !or broadcast. >! t$is cell does not )rovide t$e
in!ormation" t$e basic CBC' #ill be used.
% GSM <adio Eet#or? @lannin
%.1 Avervie#
+$e desin o! radio net#or? )lannin -<E@. is t$e basis o! t$e construction o! a
#ireless mobile net#or?. +$e desin level o! net#or? )lannin decides t$e
!uture layout o! a net#or?.
Durin net#or? )lannin" t$e documents concernin base station distribution"
c$annel assinment" and cell data must be out)utted. And t$e maFor tas?s
involved are as !ollo#s*
1. Analy(e carriersD requirements on net#or? coverae" ca)acity and quality.
%. Analy(e t$e coverae and ca)acity !eatures o! t$e candidate mobile
communication systems and bands" and t$en analy(e t$e investment !easibility
t$rou$ estimatin t$e net#or? scale.
2. Decide t$e net#or? structure and base station ty)e based on !urt$er
analysis.
First analy(e #$et$er to construct a layerin net#or? accordin to user
distribution" )ro)aation conditions" city develo)ment )lan and e:isted
net#or? conditions" and t$en analy(e t$e sites #it$in t$is area to decide
#$et$er to use omni antennas or directional antennas to meet t$e
requirements on coverae and ca)acity.
4. ;stimate t$e number o! base stations
Be!ore estimatin t$e number o! base stations" estimate t$e coverae distance
o! base stations o! various ty)es in various coverae areas. +$e !actors decidin
t$e e!!ective coverae area o! a base station include*
, Halid transmit )o#er o! t$e base station
, 9or?in bands to be used -900 M'( or 1800 M'(.
, Antenna ty)e and installation )osition
, @o#er budet
, <adio )ro)aation environment
, CarriersD inde:es on coverae
+$en t$rou$ calculatin t$e coverae distance and dividin t$e coverae
areas" you can obtain a rou$ number o! base stations !or various coverae
areas.
1. @lan an ideal base station address accordin to cellular structures.
Accordin to eora)$ic ma)s or administrative ma)s and #it$ t$e $el) o! on,
t$e,s)ot surveys" you can $ave a !ull understandin o! t$e areas to be )laned"
and t$en mar? t$e area #$ere t$e number o! users is lare as a taret address.
A!ter t$at" mar? t$e addresses o! ot$er base stations accordin to t$e ideal
cellular structure and t$e result o! lin? budet.
3. Calculate t$e number o! c$annels o! t$e cells o! eac$ base station
, ;stimate t$e tra!!ic o! a base station accordin to its ideal location" and t$en
obtain t$e number o! carriers and c$annels needed by eac$ base station by
c$ec?in ;rl table accordin to t$e inde:es o! call loss rate.
, Decide t$e !requency reuse mode accordin to band #idt$" net#or? quality
requirement" and equi)ment su))ortability.
, ;stimate t$e ma:imum base station con!iuration ty)e accordin to t$e
!requency band#idt$ and reuse mode )rovided by t$e construction carriers. >!
t$e system ca)acity in some areas cannot be met" you need to add more base
stations or cells to t$e system accordin to cell s)littin )rinci)les and actual
conditions. A!ter t$at" reselect an ideal base station address on t$e ma) and
re,estimate t$e number o! c$annels required by t$e base station.
5. @redict t$e coverae area and decide t$e )roFect data" namely" )er!orm t$e
)reliminary emulation. +$e s)eci!ic tas?s are as !ollo#s*
, Select t$e desin inde:es
Select t$e minimum received )o#er and t$e )enetration ratio inde: at t$e
coverae area ede.
, Select t$e desin )arameters" #$ic$ includes*
Antenna $ei$t -above t$e round." antenna a(imut$ anle" antenna ain"
antenna tilt anle" base station $ei$t above sea level" base station ty)e"
!eeder lent$" antenna !eeder system loss" combinin and distribution modes"
transmitter out)ut )o#er" receiver sensitivity" base station diversity rece)tion"
and diversity ains.
, @redict t$e coverae area o! eac$ cell accordin to t$e )ro)aation models in
di!!erent areas" and t$en ive t$e o)inions on adFustin t$e base station
address" antenna direction" antenna tilt anle" and antenna $ei$t in t$e areas
#$ere dead (ones may be )resent and sinals are )oor. Finally" )rovide t$e
)roFect data.
8. Select actual base station address and decide base station ty)e*
@er!orm !iled e:amination accordin to t$e ideal base station addresses" and
t$en record t$e )ossible addresses accordin to various construction conditions
-includin )o#er su))ly" transmission" electromanetic bac?round" and land
ta?en over.. Finally" recommend a suitable address based on interated
consideration o! t$e deviation !rom t$e ideal base station address" t$e e!!ect
on !uture cell s)littin" economic bene!its" and coverae )rediction.
A!ter t$e base station address is selected" decide t$e actual base station ty)e
accordin to t$e number o! base station c$annels.
A!ter t$e base station ty)e is decided" you need to ma?e a sc$eme !or antenna
con!iuration. For movin a net#or?" i! you intend to )rovide a best
combination sc$eme !or t$e antenna !eeders" you must !ully investiate t$e
combination o! t$e antenna !eeders o! t$e oriinal carriers" )lan t$e !uture
e:)ansion o! t$e base station" and desin t$e combination o! t$e antenna
!eeders su))orted by current equi)ments.
9. @lan !requency and adFacent cell
Decide t$e !requency and adFacent )lannin accordin to t$e actual base
station distribution and ty)e.
10. Ma?e cell data
+o ensure t$at t$e net#or? runs stably" you must desin t$e )arameters
relative to )er!ormance !or eac$ cell. +$ese )arameters include system
in!ormation )arameters" $andover )arameters" )o#er control alorit$m
)arameters" and so on.
, Eote*
For t$e selection o! $andover bands" t$e $andover alorit$ms to be enabled"
and #$et$er to use !requency $o))in" )o#er control" and D+O" t$ey must be
decided in coverae )rediction and !requency )lannin" because t$e related
)arameters #ill be used in emulation.
>n addition" sections %.9 and t$at later introduce t$e solutions to t$e )lannin
o! dual,band net#or? and t$e )lannin in s)ecial occasions.
%.% @lannin Foundation
%.%.1 Coverae and Ca)acity +aret Con!irmation
Be!ore )lannin a net#or?" you must con!irm t$e net#or? coverae and
ca)acity taret and relative s)eci!ications !rom carriers. +$ey are s)eci!ied as
!ollo#s*
, De!inition o! coverae areas
, S)eci!ic division o! t$e service quality in coverae areas
, Grade o! service -GoS. at /m inter!ace
, @rediction o! net#or? ca)acity and subscriber ro#t$ rate
, Available bands and restrictions on usin bands
, <estrictions on base station address and t$e number o! carriers
, @enetration loss in cars or indoor environment
, @er!ormance and sensitivity o! base stations
, <ules on base station namin and numberin
, >n!ormation o! t$e base stations in t$e e:istin net#or?
;nineers )er!orm t$e net#or? )lannin and uide t$e subsequent construction
#or? accordin to t$e )revious tec$nical s)eci!ications. Because any c$ane o!
t$ese s)eci!ications #ill a!!ect net#or? construction" you must discuss t$ese
s)eci!ications #it$ carriers and et t$eir con!irmation.
%.%.% @er!ormance +aret Con!irmation
Carriers em)$asi(e muc$ on t$e !uture net#or? quality. +$ere!ore" net#or?
)lannin enineers must Fude t$e inde:es concernin net#or? )er!ormance
accordin to construction di!!iculty and e:)erience" and t$en coo)erate #it$
carriers to desin a reasonable solution.
Generally" t$e )er!ormance o! voice services can be Fuded accordin to &@>
inde:es. +$e &@> inde:es vary sli$tly #it$ carriers.
+$e mean o)inion score -MAS. is divided into !ive levels.
, +$e call #$ose quality is above level 2 can access t$e mobile communication
net#or?.
, +$e call #$ose quality is above level 4 can access t$e )ublic net#or?.
%.2 Coverae Analysis
%.2.1 Area Division
>. +y)es o! coverae area
+$e sinal )ro)aation models are a))lied in accordance #it$ t$e )ro)aation
environments in areas o! di!!erent ty)es. +$e sinal )ro)aation models decide
t$e desin )rinci)les" net#or? structures" rade o! services and !requency
reuse modes !or t$e radio net#or?s in coverae areas. >n order to decide t$e
cell coverae area" you can t$e radio coverae areas into t$e !ollo#in !our
ty)es*
, Bi city
, Middle,si(ed city
, Small to#n
, Countryside
Bi city
, Dense )o)ulation
, Develo)ed economy
, Gare tra!!ic
, Dense $i$ buildins and mansions distributed in center areas
, Flouris$in s$o))in centers
Middle,si(ed city
, <elatively dense )o)ulation
, <elatively develo)ed economy
, <elatively lare tra!!ic
, Dense buildins distributed in center areas
, Active and )romisin s$o))in centers
Small to#n
, <elative lare )o)ulation
, @romisin economic develo)ment
, Moderate tra!!ic
, <elative dense buildins distributed in center areas
, A certain scale o! s$o))in centers but #it$ reat )otentiality
Countryside
, Scattered )o)ulation
, Develo)in economy
, Go# tra!!ic
>n addition" you must consider t$e coverae o! t$e areas at t$e intersections
and various trans)ort arteries" includin*
, ;:)ress #ay
, Eational $i$ #ay
, @rovincial $i$#ay
, <ail#ay
, Sea,route
, <oads in mountain areas
Generally" it is recommended to a))ly omni base stations in t$e countries
)lains and t$e areas #it$ restricted land!orms. >n bi cities" middle,si(ed
cities" and alon e:)ress#ays" it is recommended to a))ly directional base
stations.
>>. De!ine t$e !ield strent$ at coverae area edes
9$en de!inin t$e !ield strent$ o! t$e u)lin? edes o! a service area" you must
consider t$e !actors*
Mobile station sensitivity ,10% dBm
Fast !adin )rotection 4 dB -2 dB !or countryside.
Slo# !adin )rotection 8 dB -3 dB !or countryside.
Eoise -environmental noise and inter!erin noise. )rotection 1 dB
<emar?*
, +o ensure t$e indoor coverae in bi and middle,si(ed cities" you can consider
11dB !or t$e averae )enetration loss bet#een buildins and consider addin
1dB to t$e )rotection marin.
, Generally" t$e )ro)aation loss o! GSM 1800M'( sinals is 8 dB reater t$an
t$at o! t$e GSM 900M'( sinals in averae.
, <adio lin?s $ave t#o directions" namely" u)lin? direction and do#nlin?
direction" and t$e coverae area is de!ined by t$e direction in #$ic$ t$e
sinals are )oor" so you must consider t$e u)lin? and do#nlin? balance.
+$ere!ore" i! you intend to )lan an ideal net#or?" you must ma?e a ood )o#er
control budet so t$at t$e u)lin? and do#nlin? can be as balance as )ossible.
>>>. De!ine coverae )robability
+$e de!inition o! coverae )robability varies #it$ t$e coverae areas" and t$e
coverae )robability is radually im)roved alon #it$ t$e construction o! t$e
net#or?.
Generally" a call must be ensured to access t$e net#or? at 90L o! t$e )laces
and 99L o! t$e time #it$in t$e coverae area.
, For t$e outdoor environment in bi cities" t$e t#o ratios must be reater.
, For t$e areas in countryside" t$e t#o ratios can be lo#er.
, For trans)ort arteries" di!!erent standards are a))lied" and t$e coverae
)robability can be de!ined in accordance #it$ t$e ty)es o! t$e arteries.
%.2.% <adio ;nvironment Survey
+$rou$ surveyin radio )ro)aation environments" you can et !amiliar #it$
t$e overall land!orms" estimate t$e rou$ antenna $ei$t" and select t$e
)ro)er radio )ro)aation model" amon #$ic$ t$e radio )ro)aation model
$el)s you estimate t$e number o! base station #$en )redictin t$e coverae. >!
necessary" you must adFust t$e )ro)aation model.
%.4 Eet#or? Structure Analysis
9$en considerin t$e layout o! base stations" you must dee)ly analy(e net#or?
structure. Generally" accordin to net#or? layers" a net#or? can be divided
into middle,layer" $i$,layer" and lo#,layer. +$e base stations at t$e middle,
layer bear t$e reatest tra!!ic in a net#or?
%.4.1 Middle,Gayer Station
>. De!inition and a))lication
A middle,layer station in bi and middle,si(ed cities is de!ined as !ollo#s*
, +$e antenna is installed on buildin to)s.
, +$e antenna $ei$t ranes !rom %1 to 20 meters" #$ic$ is reater t$an t$e
averae $ei$t o! t$e buildins.
, >t covers several bloc?s.
>n small to#ns and countryside areas" e:ce)t t$e $i$,layer stations are
desined !or controllin tra!!ic !lo# or !or land!orm reasons" most o! t$e base
stations are middle,layer stations.
>>. Advantaes
Com)ared #it$ $i$,layer stations" middle,layer stations can utili(e !requency
resources more e!!iciently. Com)ared #it$ lo#,layer stations" middle,layer
stations can absorb tra!!ic more e!!iciently. +$ere!ore" t$e middle,layer
stations bear t$e reatest tra!!ic in a net#or?.
>>>. Distance bet#een stations
+$e averae distance bet#een most middle,layer stations rane !rom 0.3 to 1
?m e:ce)t in countryside areas. >n bi cities" t$e distance bet#een some
middle,layer stations is s$orter t$an 0.3 ?m. 'o#ever" it is suested t$at t$e
distance bet#een middle,layer stations in bi cities cannot be s$orter t$an 0.4
?m. >! t$is distance is too s$ort" t$e buildins #ill )roduce stron inter!erence
aainst t$e sinals o! t$e base stations. >n t$is case" to control t$e coverae
area is quite demandin.
>H. C$allenes
Because no suitable round obFective is available" to ensure t$e quality o!
service o! a net#or? is quite demandin. Accordin to t$e e:)erience on
)roFect construction and maintenance" reat c$allene is )resent in t$e
selection o! base station address" station desin" )roFect construction" net#or?
maintenance" and net#or? quality.
%.4.% 'i$,Gayer Station
>. De!inition and a))lication
A $i$,layer station in bi and middle,si(ed cities is de!ined as !ollo#s*
, +$e antenna $ei$t ranes !rom 10 to 10 meters" #$ic$ is !ar reater t$an t$e
averae $ei$t o! t$e buildins.
, >ts coverae areas contain t$e areas covered by multi)le middle,layer
stations.
Because t$e $i$,layer stations ma?e )oor use o! t$e !requency resources" t$ey
are mainly a))lied to t$e tra!!ic net#or?s #$ere )eo)le move !ast in bi and
middle,si(ed cities.
>n addition" to control construction cost and meet coverae requirements" you
can install some $i$,layer stations in suburban areas" $i$roads" small to#ns"
and countryside areas.
>>. Functions
+$e $i$,layer stations must be as !e#er as )ossible but be as e!!ective as
)ossible. +$ey mainly )rovide services to t$e !ast,movin subscribers in cities.
^ Eote*
+$e coverae o! $i$ buildins is reali(ed by indoor distribution systems.
%.4.2 Go#,Gayer Station
>. De!inition and a))lication
A lo#,layer station is de!ined as !ollo#s*
, +$e antenna $ei$t is s$orter t$an %0 meters" #$ic$ is s$orter t$an t$e
averae $ei$t o! t$e buildins.
, +$e antenna can be installed on t$e outer #alls o! t$e lo#er !loors o! a
buildin" on t$e to) o! lo#er roo!s" or in t$e rooms o! a buildin.
Generally" at t$e early stae o! t$e net#or? construction" sinal net#or?
desin is a))lied" so most o! t$e base stations are middle,layer stations. A!ter
t$e basic net#or? is establis$ed" you must adFust t$e base stations and add
ne# base stations accordin to tra!!ic and coverae requirements.
For )o)ulated commercial areas #$ere t$e tra!!ic is $eavy" you can use lo#,
layer stations" #$ic$ are constructed #it$ micro cell layer and distributed
antenna system. >n t$is case" not only t$e requirements on indoor coverae are
met" but also t$e inter!erence and di!!iculties o! base station selection caused
by s$ort distance bet#een stations are avoided. 9it$ t$e develo)ment o! t$e
net#or?" t$e lo#,layer stations #ill develo) into t$e layerin net#or?
structure.
>>. At$er considerations
+$e coverae area o! a lo#,layer station is small" so it can !ully use !requency
resources but cannot absorb t$e tra!!ic e!!iciently. As a result" ideal tra!!ic
cannot be ensured i! t$e base station deviates !ar a#ay !rom t$e areas #$ere
t$e tra!!ic is $eavy.
+$ere!ore" #$en constructin a lo#,layer station" you must consider #$et$er
t$e base station is used to ma?e u) coverae or solve t$e )roblem o! $eavy
tra!!ic" because t$e construction )ur)ose is directly related to t$e selection o!
t$e address and ty)e o! t$e base station.
^ Eote*
A layerin net#or? cost muc$ !requency resource" so it is not recommended !or
t$e net#or?s #$ere t$e !requency resource is inadequate.
%.1 +ra!!ic Analysis
%.1.1 +ra!!ic @rediction and Cell S)littin
>. +ra!!ic )rediction
+$e net#or? construction requires t$e consideration o! economic !easibility
and rationality. +$ere!ore" a reasonable investment decision must be based on
t$e )rediction o! t$e net#or? ca)acity o! t$e early and late stae.
9$en )redictin net#or? ca)acity" you must consider t$e !ollo#in !actors*
, @o)ulation distribution
, Family income
, Subscri)tion ratio o! !i:ed tele)$one
, Develo)ment o! national economy
, City construction
, Consum)tion )olicy
A!ter )redictin t$e total net#or? ca)acity" you must )redict t$e density o!
subscriber distribution. Generally" base stations are constructed in urban areas"
suburban areas" and trans)ort arteries. +$ere!ore" you can use t$e )ercentae
o! )rediction met$od.
At t$e early stae o! construction" t$e subscribers in cities account !or a larer
)ercentae o! t$e total )redicted subscribers. 9it$ t$e develo)ment o! t$e
net#or? construction" t$e )ercentae o! t$e subscribers in suburban areas and
trans)ort arteries ro#s. +$e tra!!ic o! eac$ subscriber is 0.0%1 ;rl in urban
areas and 0.0%0 ;rl in suburban areas.
+$e !ormula calculatin tra!!ic is*
A N -n C +. = 2300
'ere"
, 6n7 is t$e call times in busy $our
, 6+7 is t$e duration o! eac$ call" in t$e unit o! second.
>n t$is #ay" t$e number o! voice c$annels needed !or a base station can be
obtained t$rou$ )redictin t$e tra!!ic.
^ Eote*
9$en estimatin t$e number o! voice c$annels needed !or a base station in t$e
!uture" you must consider t$e e!!ect caused by cell s)littin.
>n a GSM system" you can use ;rl model to calculate t$e tra!!ic density t$at t$e
net#or? can bear. +$e call loss can be %L or 1L de)endin on actual
conditions.
Because restrictions on cell coverae area and t$e #idt$ o! t$e available
!requencies are )resent" you must )lan t$e cell ca)acity reasonably. >! ood
voice quality is ensured" you must en$ance t$e c$annel utili(ation ratio as
muc$ as )ossible.
>n actual net#or?in" i! t$e net#or? quality is ensured at a certain level" t#o
ca)acity solutions are available" namely" a !e# stations #it$ $i$,level
con!iuration and multi)le stations #it$ lo#,level con!iuration. Bot$ t$e
advantaes and disadvantaes o! t$e t#o solutions are a))arent" so #$ic$ one
s$ould be used de)endin on t$e actual conditions o! an area.
For net#or? construction" you can e:)and t$e ca)acity eit$er t$rou$ addin
base stations or t$rou$ e:)andin t$e base station ca)acity. +$e e:)ansion
strateies ado)ted must be in accordance #it$ t$e tra!!ic density in an area.
For e:am)le" t$e strateies suc$ as addin 1800 M'( base stations" e:)andin
sector ca)acity" addin micro cells" or im)rovin indoor coverae can be used
to e:)and net#or? ca)acity.
>>. Cell s)littin
Cell s)littin is quite e!!ective !or t$e e:)ansion o! net#or? ca)acity. An omni
base station can s)lit into multi)le sectors" and a sector can s)lit into multi)le
smaller cells. >n ot$er #ord" you must )lan cell radius in accordance #it$ t$e
tra!!ic density o! an area.
Cell s)littin means more base station and reater cost are needed. +$ere!ore"
#$en )lannin a net#or?" you must consider t$e !ollo#in !actors*
, +$e rules and diarams o! !requency reuse are re)eatable.
, +$e oriinal base stations can still #or?.
, +$e transition cells must be reduced or avoided.
, +$e cell can s)lit #it$out e!!ect.
Cell s)littin is quite im)ortant in a net#or?. +$e !ollo#ins !urt$er describe
t$e cell s)littin based on 1,to,4 s)littin.
Cell s)littin is used to s)lit a conested cell into multi)le smaller cells.
+$rou$ settin t$e ne# cells #$ose radiuses are smaller t$an t$e oriinal cells
and )lacin t$em amon t$e oriinal cells" you can increase t$e number o!
c$annels in a unit area" t$us increasin c$annel reuse times. >n t$is case"
system ca)acity is e:)anded.
+$rou$ adFustin t$e )roFect )arameters relative to antenna !eeders and
reducin transmitter )o#er" you can narro# t$e coverae area o! a cell. ;rror_
<e!erence source not !ound. s$o#s t$at a cell s)lits into !our smaller cells by
$al! o! its radius.
Smaller cells are added #it$out c$anin t$e !requency reuse mode. +$ey are
s)lit )ro)ortional to t$e s$a)e o! t$e oriinal cell clusters.
>n t$is case" t$e coverae o! a service area de)ends on t$e smaller cells" #$ic$
are 4 times outnumber o! t$e oriinal cells. +o be more s)eci!ically" you can
ta?e a circle #it$ t$e radius < as an e:am)le" t$e coverae area o! t$e circle
#it$ t$e radius < is 4 times t$at o! a circle #it$ t$e radius <=%.
A!ter cell s)littin" t$e number o! cell clusters in t$e coverae area increases.
+$us t$e number o! c$annels in t$is coverae area increases and t$e system
ca)acity is e:)anded accordinly.
Rou can adFust t$e coverae area o! t$e ne# cells t$rou$ reducin t$e
transmit )o#er. For t$e transmit )o#er o! t$e ne# cells #$ose radiuses are
$al! o! t$at o! t$e oriinal cell" you can c$ec? t$e )o#er 6@r7 received at t$e
ne# cell ede and at t$e oriinal cell ede" and ma?e t$em equal. 'o#ever"
you must ensure t$at t$e !requency reuse sc$eme o! t$e ne# micro cells is t$e
same as t$at o! t$e oriinal cell. As !or Fiure 1,1"
, @r [at t$e ede o! t$e oriinal cell\ N @t1<,n" and"
, @r [at t$e ede o! t$e ne# cell\ N @t% -<=%.,n
'ere"
@t1 and @t% are t$e transmit )o#er o! t$e base stations o! t$e oriinal cell and
t$e ne# cell" and n is )at$ !adin e:)onent. >! ma?e n N 4" ma?e t$e received
)o#er at t$e ede o! t$e ne# and oriinal cell equal" t$e !ollo#in equation
can be obtained*
@t% N @t1=13
+$at is to say" i! t$e micro cells are used to cover t$e oriinal coverae area
and t$e requirement o! S=> is met" t$e transmit )o#er must be reduced by 1%
dB.
Eot all cells need s)littin. >n !act" it is quite demandin !or carriers to !ind out
a )er!ect cell s)littin sc$eme. +$ere!ore" many cells o! di!!erent scales e:ist
in a net#or? simultaneously. As a result" t$e minimum distance amon intra,
!requency cells must be maintained" #$ic$ !urt$er com)licate !requency
allocation.
>n addition" you must )ay attention to t$e $andover because success $andover
ensure t$e all subscribers to enFoy ood quality o! service reardless o! movin
s)eed.
9$en t#o layers o! cells are )resent #it$in an area but t$eir coverae scale is
di!!erent" accordin to t$e !ormula @t% N @t1=13" neit$er all ne# cells can
sim)ly a))ly t$e oriinal transmit )o#er" nor all oriinal cells can sim)ly a))ly
t$e ne# transmit )o#er.
>! all cells a))ly reat transmit )o#er" t$e c$annels used by smaller cells
cannot be se)arated !rom t$e intra,!requency cells. >! all cells a))ly lo#er
transmit )o#er" $o#ever" some bi cells #ill be e:clusive !rom t$e service
areas.
For t$e )revious reason" t$e c$annels in t$e oriinal cells can be divided into
t#o rou)s. Ane rou) meets t$e reuse requirement o! t$e smaller cells" and
t$e ot$er rou) meets t$e reuse requirement o! t$e bier cells. +$e bier
cells are a))lied to t$e communication o! !ast,movin subscribers" #$ic$
requires a !e#er $andover times.
+$e )o#er o! t$e t#o c$annel rou)s decides t$e )roress o! cell s)littin. At
t$e early stae o! cell s)littin" t$e c$annels in t$e lo#,)o#er rou) are !e#er.
As t$e requirement ro#s" more c$annels are needed in lo#,)o#er rou). +$e
cell s)littin does not sto) until all c$annels #it$in t$is area are a))lied in t$e
lo#,)o#er rou). >n t$is case" all cells in t$is area $ave s)lit into multi)le
smaller cells" and t$e radius o! eac$ cell is quite small.
^ Eote*
Commonly" you can restrict cell coverae area t$rou$ adFustin t$e )roFect
)arameters o! t$e base station.
%.1.% Hoice C$annel Allocation
>. Hoice c$annel decision
+$e base station ca)acity re!ers to t$e number o! c$annels t$at must be
con!iured !or a base station or a cell. +$e calculation o! t$e base station
ca)acity is divided into t$e calculation o! t$e number o! radio voice c$annels
and t$e calculation o! t$e number o! radio control c$annels.
Accordin to t$e in!ormation o! base stations and cells and t$e density
distribution o! subscribers" you can calculate t$e total number o! t$e
subscribers. +$en accordin to t$e radio c$annel call loss ratio and tra!!ic" you
can obtain t$e number o! voice c$annels t$at must be con!iured by c$ec?in
;rl B table.
Generally" you can decide t$e number o! voice c$annels as !ollo#s*
1. Accordin to t$e band#idt$ and t$e reuse mode allo#ed by current GSM
net#or?s #it$in t$e areas to be )lanned" you can obtain t$e ma:imum number
o! carriers t$at can be con!iured !or a base station.
%. ;ac$ carrier $as 8 c$annels. Rou can obtain t$e ma:imum number o! voice
c$annel numbers t$at can be con!iured !or a base station by detractin t$e
control c$annels !rom t$e 8 c$annels.
2. Accordin to t$e number o! voice c$annels and call loss ratio -enerally %L
dense tra!!ic areas and 1L !or ot$er areas." you can obtain t$e ma:imum tra!!ic
-;rl number. t$at t$e base station can bear t$rou$ c$ec?in ;rl B table.
4. +$rou$ dividin t$e ;rl number by t$e averae busy,$our tra!!ic o!
subscribers" you can obtain t$e ma:imum number o! subscribers t$at t$e base
station can accommodate.
1. Accordin to t$e data o! subscriber density" you can obtain t$e coverae
area o! t$e base station.
3. A!ter t$e areas are s)eci!ied based on t$e subscriber density" accordin to
t$e area o! an area and t$e actual coverae area o! t$e base station" you can
calculate t$e number o! needed base stations.
5. For im)ortant areas" you must consider bac? u) stations and t$e coo)eration
bet#een carriers. For e:am)le" an im)ortant county needs at least t#o base
stations and t$ree im)ortant carriers.
8. For t$e areas #$ere burst tra!!ic is )ossible" suc$ as t$e )lay round and
seasonal tourism s)ots" you must )re)are t$e equi)ments -suc$ as carriers and
micro cells. and !requency resources !or !uture use.
9. +$e dynamic !actors" suc$ as roamin ratio" subscriber mobility" service
develo)ment" industry com)etition" c$arin rate c$ane" one,#ay c$are" and
economic ro#t$" must be considered.
10. +o con!iure a base station" you must consider t$e transmission at t$e Abis
inter!ace so t$at t$e ca)acity can be met #$ile savin transmission. For
e:am)le" t$e a))lication and concatenation o! t$e Abis inter!ace 11*1 and 1%*1
s$ould be considered.
11. For indoor coverae and ca)acity" you can use micro cells and distributed
antenna systems. For t$e coverae in countryside areas and $i$roads" you can
use economical micro base stations. For t$e transmission in countryside areas
and $i$roads" you can use 'DSG because it is cost e!!ective.
1%. @re)are t$e some carriers" micro cells" and micro base stations !or ne#
coverae areas and !uture o)timi(ation.
12. >n some s)ecial areas" you can use t$e base stations consistin o! omni and
directional cells" but you must consider t$e isolation bet#een omni antennas
and directional antennas. For tra!!ic control" you can use t$e alorit$m in
terms o! net#or? layers.
14. For some $i$roads #$ic$ require a little tra!!ic by lare coverae" you can
use t$e t#o net#or?in modes. +$ey are*
, -A micro base station #it$ sinle carrier. J -0.1 J 0.1 cell #it$ t#o set o!
directional antennas.
, A micro base station #it$ sinle carrier J 8,s$a)ed antenna
>>. <elations$i) bet#een carrier number and bearable tra!!ic
;rl tra!!ic model can calculate t$e tra!!ic t$at a net#or? can bear. +$e call loss
ratio can be %L or 1L accordin to actual conditions. +able 1,5 describes t$e
relations$i) bet#een t$e number o! carriers and t$e tra!!ic t$at a net#or? can
bear accordin to ;rl B table.
Accordin to ;rl B table" t$e larer t$e number o! carriers and t$e call loss
ratio are" t$e reater t$e tra!!ic t$at eac$ +C' bear" and t$e reater t$e +C'
utili(ation ratio is -t$e c$annel utili(ation ratio is an im)ortant indicator o! t$e
quality o! net#or? )lannin and desin.. >! t$e number o! subscribers o! a base
station is small" you can consider delayin t$e construction.
Because restrictions on t$e coverae area o! a cell and t$e band#idt$ o! t$e
available !requencies" you must )lan a reasonable ca)acity !or t$e cell. >! ood
voice quality is ensured" you must ta?e measures to en$ance t$e c$annel
utili(ation ratio as muc$ as )ossible.
For t$e construction o! t$e dual,band net#or?" you can use t$e !requencies
#it$ #ider bands to en$ance c$annel utili(ation ratio" #$ic$ is $el)!ul !or
tra!!ic s$arin.
>n actual a))lications" #$en t$e tra!!ic on eac$ +C' accounts !or 80,90L o!
total iven by ;rl B table -t$e call loss ratio is %L." t$e conestion ratio in t$is
cell rise reatly. +$ere!ore" #e enerally calculate t$e tra!!ic t$at a net#or?
can bear by ta?in t$e 81L o! t$e tra!!ic iven by ;rl B table as a re!erence.
>>>. ;:am)le
+$e ca)acity o! a local net#or? needs to be e:)anded. Accordin to t$e service
develo)ment" )o)ulation ro#t$ and mobile )o)ularity" t$e subscribers in t$is
area are e:)ected to reac$ 100"000 in % years.
>! only t$e !ollo#ins are considered*
, <oamin !actor -accordin to t$e develo)ment trend o! tra!!ic statistics. N
10L.
, Mobile !actor -t$e subscriber moves sli$tly #it$in t$e local net#or? instead
o! roamin. N 10L.
, Dynamic !actor -#it$ burst tra!!ic considered. N 11L.
+$e net#or? ca)acity N 100000 M -1 J 10L J 10L J 11L. N 121"000.
'o#ever" because t$e conestion is )resent" #e enerally calculate t$e tra!!ic
t$at a net#or? can bear by ta?in t$e 81L o! t$e tra!!ic iven by ;rl B table as
a re!erence. As a result" t$e net#or? ca)acity must be desined as !ollo#s*
+$e net#or? ca)acity N 121" 000=81L N 118"800" about 130"000.
%.1.2 Control C$annel Allocation
>. SDCC' allocation
Stand,alone dedicated c$annel -SDCC'. is an im)ortant c$annel in a GSM
net#or?. Mobile station activities" suc$ as location u)date" attac$ and detac$"
call setu) and s$ort messae" are )er!ormed on SDCC'. +$e SDCC' is used to
transmit sinalin and data.
>t is di!!icult to induce a tra!!ic model !or t$e SDCC'8 es)ecially it even
becomes im)ossible a!ter t$e lare,scale a))lication o! layerin net#or?s and
s$ort messaes. Moreover" t$e equi)ments o! some carriers su))ort SDCC'
dynamic allocation !unction. As a result" t$e tra!!ic model !or SDCC' must be
adFusted accordin to actual conditions.
+$e advantaes o! t$e SDCC' dynamic !unction are as !ollo#s*
, AdFustin SDCC' ca)acity dynamically
, <educin SDCC' conestion ratio
, <educin t$e e!!ect o! initial SDCC' con!iuration aainst system
)er!ormance
, Ma?in SDCC' and +C' con!iuration more ada)tive to t$e c$aracteristics o!
cell tra!!ic
, A)timi(in t$e )er!ormance o! t$e systems under t$e same carrier
con!iuration.
>n conclusion" t$e SDCC' dynamic allocation !unction is divided into t#o ty)es"
namely"
, Dynamic allocation !rom SDCC' to +C'
, Dynamic recovery !rom SDCC' to +C'
>>. CCC' allocation
Common control c$annels -CCC'. contain access rant c$annel -AGC'." )ain
c$annel -@C'. and random access c$annel -<AC'.. +$e !unction o! a CCC' is
sendin access rant messae -immediate assinment messae. and )ain
messae.
All tra!!ic c$annels in eac$ cell s$are t$e CCC'. +$e CCC can s$are a )$ysical
c$annel -a timeslot. #it$ SDCC'" or it can solely occu)y a )$ysical c$annel.
+$e )arameters relative to t$e CCC' include CCC' Con!iure" BS AG BG&S @;S"
and BS @A MF<MS.
'ere"
, CCC' Con!iure desinates t$e ty)e o! CCC' con!iuration" namely" #$et$er
t$e CCC' s$ares one )$ysical c$annel #it$ t$e SDCC'. >! t$ere are 1 or % +<O
in a cell" it is recommended t$at t$e CCC' occu)ies a )$ysical c$annel and
s$are it #it$ t$e SDCC'. >! t$ere are 2 or 4 +<Os" it is recommended t$at t$e
CCC' solely occu)ies a )$ysical c$annel. >! t$ere are more t$an 4 +<O" it is
recommended to calculate t$e ca)acity o! t$e )ain c$annels in t$e CCC'
accordin to actual conditions !irst" and t$en you can )er!orm t$e
con!iuration.
, BS AG BG&S @;S indicates t$at t$e number o! CCC' messae bloc?s reserved
to t$e AGC'. A!ter CCC' con!iuration is done" t$is )arameter" in !act" decides
allocates t$e ratio o! AGC' and @C' in CCC'. Some carriers can set sendin
)riority !or t$e 6access rant messae and 6)ain messae7. 9$en t$e !ormer
messae set to be )rior to t$e later one" t$e BS AG BG&S @;S can be set to 0.
, BS @A MF<MS indicates t$e number o! multi,!rames t$at can be ta?en as a
cycle o! )ain sub,c$annels. >n !act" t$is )arameter decides t$e number o!
)ain sub,c$annels t$at a cell can be divided into.
^ Eote*
>n CCC' con!iuration" t$e location area )lannin" )ain modes and system
!lo# control must be considered.
%.3 Base Station Eumber Decision
A!ter tra!!ic and coverae analysis" accordin to t$e selected base station
equi)ments and )arameters" you can obtain t$e coverae areas o! various base
stations t$rou$ lin? budet. +$e coverae area $el)s you calculate t$e
number o! base stations required by eac$ area. +$en you decide t$e base
station con!iuration accordin to tra!!ic distribution. Finally" you must
)er!orm emulation usin relative )lannin so!t#are so t$at coverae" ca)acity"
carrier,to,inter!erence ratio can be assured and inter!erence can be avoided.
%.3.1 C$aracteristics o! 2,sector base stations in urban areas
Cellular communication is named because t$e coverae areas o! base stations
are e:truded t$rou$ small cellular,s$a)ed bloc?s. >n urban areas" !or t$e
)ur)ose o! ca)acity e:)ansion and radio !requency o)timi(ation" mainly 2,
sector base stations are used. +$is section e:)lains some basic conce)ts o! a 2,
sector base station.
+$is is a standard 2,sector cellular layout. +$edistance bet#een t#o 2,sector
base stations is < J r" $ere < N %r. 'o#ever" 6<7 is mainly used in cell radius
estimation because t$e direction alon 6<7 is t$e direction o! t$e maFor lobe o!
t$e directional antenna. >n t$e desin !or cellular layout" $o#ever" 6r7
indicates t$e cell radius.
>n a cellular cell" i! t$e included anle bet#een a direction and t$e direction o!
t$e maFor lobe o! t$e antenna" t$e coverae distance alon t$is direction is r N
<=%" and t$e )at$ loss alon t$is direction is about 10dB less t$an t$at alon
t$e direction o! t$e maFor lobe o! t$e antenna -!or t$e deduction" it is
introduced in t$e !ollo#in." namely" t$e equivalent isotro)ic radiated )o#er
-;><@. alon t$is direction can be about 10dB less t$an t$at alon t$e maFor
lobe.
Accordin to t$is !eature" in t$e cellular layout o! t$is ?ind" you can ado)t t$e
directional antenna #$ose a(imut$ beam #idt$ ranes !rom 30 to 31 derees
because t$eir $ori(ontal lobe ain diaram also meets t$is !eature.
>! 6<7 is t$e cell radius" t$e cell area is S N 0.3491 C < C <. Sometimes t$e 6r7
is used as cell radius" so t$e cell area is S N % 1981CrCr. +$ere!ore" #$en
calculatin t$e cell area" you must ma?e clear #$et$er 6r7 or 6<7 is used.
+$e !ollo#ins deduce t$e ;><@ required alon 6<7 direction and 6r7 direction.
As s$o#n in Fiure 1,2" t$e coverae distance alon 6r7 direction is $al! o! t$at
alon 6<7 direction" namely" r N <=%. +o ?ee) even coverae" you must ma?e
t$e !ield intensity at t$e edes o! t$e cell equal" namely" <:lvelB N <:levelC.
Su))ose t$at t$e ;>@< transmitted !rom cell A is ;><@< and ;><@r alon 6<7
direction and 6r7 direction res)ectively" and t$e city 'A+A mode is used !or
)at$ loss" t$e )at$ loss !rom )oint A and B is e:)ressed as equation -1. *
;><@< 0 <OG;HB N 39.11 J %1.33l! , 12.8%l$1 J -44.9 , 3.11l$1. l< -1.
And t$e )at$ loss !rom )int A to )oint C is e:)ressed as equation -%.*
;><@r, <OG;Hc N 39.11 J %1.33l! , 12.8%l$1 N -44.9 , 3.11l$1. lr -%.
Subtract -%. !rom -1." t$e equation -2. is e:)ressed as !ollo#s*
;><@< , ;><@r N-44.9 , 3.11l$1.C-l< 0 lr. N-44.9 , 3.11l$1. C l -<=r. -2.
>ntroduce < N %r" t$e equation -4. is obtained as !ollo#s*
;><@< , ;><@r N 0.2 C -44.9 , 3.11l$1. -4.
9$en t$e antenna $ei$t 6$17 increases !rom 1m to 100m" t$e values o! -;><@<
, ;><@r. decrease !rom 1% to 9.1" #$ic$ can be rou$ly treated as 10dB.
1.1.% <e!erences !or Desin o! Base Station @arameters
9$en estimatin t$e number o! base stations" you must )er!orm u)lin? and
do#nlin? budet. Based on t$e coverae division and )ro)aation environment
survey" you can obtain some )roFect )arameters and a))ly t$em to lin? budet.
%.3.2 /)lin? and Do#nlin? Balance
A!ter base station )arameters are s)eci!ied" you can )er!orm lin? budet to
estimate t$e coverae area o! t$e base station. >n addition" you must consider
t$e sensitivity o! t$e base station equi)ments at t$is time.
>n a mobile communication system" radio lin?s are divided into t#o directions"
namely" u)lin? and do#nlin?. For an e:cellent system" you must )er!orm a
ood )o#er budet so t$at t$e balance is )resent bet#een u)lin? sinals and
do#nlin? sinals. At$er#ise" t$e conversation quality is ood !or one )arty but
bad !or t$e ot$er )arty at t$e edes o! t$e cell. >! u)lin? sinals are too bad"
t$e mobile station cannot start a call even i! sinals are )resent.
'o#ever" t$e because t$e !adin !or u)lin? c$annels and do#nlin? c$annels is
not totally t$e same and t$e ot$er !actors suc$ as t$e di!!erence o! t$e
)er!ormances o! receivers are )resent" t$e calculated u)lin? and do#nlin? are
not absolute" but t$e t$ere a !luctuation o! % to 2 dB.
+$e measurement re)ort on u)lin?s and do#nlin?s at t$e Abis inter!ace can tell
#$et$er t$e u)lin? and do#nlin? reac$ a balance. >n addition" dialin tests in
actual net#or? can also tell #$et$er t$e balance bet#een u)lin?s and
do#nlin?s are reac$ed. >! t$e conversation quality on do#nlin?s u)lin?s
becomes )oor simultaneously" it means t$at t$e do#nlin?s and u)lin?s are
balance.
^ Eote*
Some carriers )rovide t$e tra!!ic statistics on u)lin? and do#nlin?
measurement" #$ic$ can also tell #$et$er t$e balance bet#een u)lin?s and
do#nlin?s are reac$ed.
>. Gin? budet model
9$en calculatin u)lin? and do#nlin? balance" you must consider t$e !unctions
o! t$e to#er am)li!ier !irst. >n a base station receivin system" t$e t$ermal
movement o! t$e active )arts and radio !requency -<F. conductors cause
t$ermal noise" #$ic$ reduces t$e sinal,to,noise ratio o! t$e receivin system.
>n t$is case" t$e receivin sensitivity o! t$e base station is restricted and t$e
conversation quality is reduced. +o im)rove t$e receivin )er!ormance o! t$e
base station" you can add a lo#,noise am)li!ier under t$e receivin antenna.
And t$is is t$e )rinci)le o! t$e to#er am)li!ier.
+$e contributions o! t$e to#er am)li!ier to u)lin?s and do#nlin?s are Fuded
accordin to t$e )er!ormance o! its lo#,noise am)li!ier and ain. >n !act" it is
t$e to#er am)li!ier t$at reduces t$e noise coe!!icient o! t$e base station
receivin system. +$e )o#er am)li!ier can im)rove t$e coe!!icients !or t$e
u)lin? receivin system -start !rom t$e out)ut end o! t$e receivin antenna..
'o#ever" i! t$e !unctions o! t$e to#er am)li!ier are quanti!ied by t$is" t$e
u)lin? im)roved value can be re)resented by t$e EFDelta -it is t$e reduced
value o! t$e noise coe!!icient o! t$e receivin system. a!ter a to#er am)li!ier is
added to t$e system.
-1. Eo to#er am)li!ier
9$en t$ere is no to#er am)li!ier" t$e sensitivity o! t$e equi)ments at t$e
du)le:er in)ut inter!ace at t$e to) o! t$e base station cabinet are ta?en as a
re!erence.
For do#nlin? sinals" i!"
Mobile station receiver out)ut )o#er N @outm
Base station diversity received ain N Gdb
Base station receivin level N @inb
Base station side noise deterioration N @bn
Antenna receivin ain N antenna transmittin ain -accordin to reci)rocity
t$eorem.
+$e !ollo#in equation can be obtained*
@inb J M! N @outm J Gam 0 Gd J Gab J Gdb 0 G!b 0 @bn
Generally" @mn is almost equal to @bn" so t$e !ollo#in equation can be
obtained*
@outb N @outm J Gdb J -@inm 0 @inb. J Gcb
-%. 9it$ to#er am)li!ier
>! a to#er am)li!ier is )resent" t$e im)roved value o! t$e noise coe!!icients o!
t$e u)lin? receivin system can be re)resented by EFDelta" so t$e equation
@outb N @outm J Gdb J -@inm 0 @inb. J Gcb can be develo)ed into t$e !ollo#in
equation*
@outb N @outm J Gdb J -@inm , @inb. J Gcb J EFDelta
+$e t#o equations" @outb N @outm J Gdb J -@inm 0 @inb. J Gcb and @outb N
@outm J Gdb J -@inm , @inb. J Gcb J EFDelta are used to calculate base station
transmit )o#er #$en t$e u)lin?s and do#nlin?s are balance. 'ere"
@inb is t$e base station receivin sensitivity
@inm is t$e mobile station receivin sensitivity
Gdb -antenna diversity receivin ain. is 2.1dB
Accordin to t$e requirements in )rotocols GSM01.01" t$e mobile station
transmit )o#er and t$e re!erence receivin sensitivity o! t$e mobile station
and base station are s)eci!ied in +able 1,10. At )resent" $o#ever" t$e
sensitivities in actual systems are reater t$an t$e re!erence values listed in
t$e !ollo#in table.
>>. Bass station sensitivity
+$is section !urt$er introduces t$e base station sensitivity and t$e !unctions o!
t$e to#er am)li!ier.
<eceiver sensitivity re!ers to t$e minimum sinal level needed to by t$e in)ut
end o! t$e receiver #$en t$e certain bit error rate -B;<. is met. +$e receiver
sensitivity detects t$e )er!ormances o! t$e !ollo#in com)onents*
<eceiver analo <F circuit
>ntermediate !requency circuit and demodulation
Decoder circuit
+$ree )arameters are used to measure t$e receiver bit error )er!ormance.
+$ey are !rame e:)uration rate -F;<." residual bit error rate -<B;<." and bit
error rate -B;<.. 9$en a !ault is detected in a !rame" t$is !rame is de!ined as
deleted one.
'ere"
F;< indicates t$e ratio o! t$e deleted !rames to t$e total received !rames. For
!ull rate voice c$annels" t$e F;< is )resent #$en t$e 2,bit cyclic redundancy
c$ec? -C<C. detects errors or bad error indication -BF>. is caused. For sinalin
c$annels" t$e F;< is )resent #$en t$e !ire code -F><;. or ot$er )ac?et codes
detect errors. +$e F;< is not de!ined in data services.
FB;< indicates t$e B;< t$at are not announced as deleted !rames" namely" it is
t$e ratio o! t$e bit errors in t$e !rame detected as 6ood7 to t$e total number
o! bits transmitted in 6ood7 !rames.
B;< indicates t$e ratio o! t$e received error bits to all transmitted bits.
Because B;< occurs at random" t$e statistical measurement is mainly a))lied
to measure receiver error rate. +$at is" sam)le multi)le measurin )oints on
eac$ c$annel and #$en t$e number o! measurin )oints is certain" i! t$e B;< o!
eac$ measurement is #it$in t$e required limit" t$e B;< o! t$is c$annel meets
t$e B;< as required.
'o#ever" t$e number o! sam)led measured )oints and t$e limit value o! t$e
B;< must meet t$e !ollo#in conditions*
For eac$ inde)endent sam)led measurin )oint" t$e times !or it to )ass a
6bad7 unit must be as !e#er as )ossible" t$at is" t$e )robability must be
smaller t$an %L.
For eac$ inde)endent sam)led measurin )oint" t$e times !or it to )ass a
6ood7 unit must be as more as )ossible" t$at is" t$e )robability must be
reater t$an 99.5L.
+$e measurement $as vivid statistical !eatures.
+$e measurin time must be reduced to t$e minimum.
As a result" you can measure t$e receiver sensitivity t$rou$ measurin
#$et$er t$e receiver B;< $as reac$ed t$e requirement #$ile enterin
sensitivity level to t$e receiver.
;nter t$e re!erence sensitivity level to t$e receiver in various )ro)aation
environments. For t$e data )roduced a!ter receiver demodulation and c$annel
decodin" t$e inde:es !or F;<" <B;<.
+$e requirements on BCC'" AGC'" @C'" and SACC' are t$e same as t$at on
SDCC'.
+$e value o! 6a7 in t$is table de)ends on t$e c$annels. >t is 1 !or base stations"
and 1 to 1.3 !or mobile stations.
>>>. Contributions o! to#er am)li!ier to base staiton sensitivity
>n terms o! tec$nical )rinci)les" t$e to#er am)li!ier reduces t$e noise
coe!!icients o! t$e base station receivin system" #$ic$ is $el)!ul !or im)rovin
t$e sensitivity o! t$e base station receivin system.
>n an actual system" to im)rove t$e receivin )er!ormance o! t$e base station"
you can add a lo#,noise am)li!ier near t$e !eeder o! t$e receivin antenna.
>n a mobile communication system" t$e receiver sensitivity N noise s)ectrum
intensity -dBm='(. J band#idt$ -dB'(. J noise coe!!icient -dB. J C=> -dB..
'ere t$e noise s)ectrum intensity" band#idt$" and noise coe!!icient are system
t$ermal noise. C=> is t$e sinal,to,noise ratio required at t$e /m inter!ace. >n a
narro# band system" C=> indicates t$e modulation )er!ormance required by t$e
receiver baseband" and it is a )ositive number.
>n a s)readin communication system" because s)read s)ectrum ain is
)resent" t$e value o! C=> is !ar beyond t$e requirement o! t$e modulation
)er!ormance o! t$e receiver baseband" and it is a neative number.
9$en t$ere are nM cascaded receivers" t$e equivalent noise coe!!icient is as
!ollo#s*
'ere"
Gn indicates t$e receivers ain at eac$ level -includin t$e loss at eac$ level..
Fn indicates t$e noise coe!!icient o! t$e receivers at eac$ level.
+$e noise coe!!icient o! t$e )assive device is equal to its loss" and t$e ain o!
t$e )assive device is t$e reci)rocal o! t$e loss.
Accordin to t$e )revious equation" t$e noise coe!!icient o! t$e cascadin
system is determined by t$e receivers at t$e !irst level.
>t must be )ointed out t$at t$e linear values o! t$e )arameters must be a))lied
in t$e )revious equation" so t$e 6F7 is a linear value" #$ic$ must be converted
into a loarit$m. Moreover" accordin to t$is equation" t$e noise t$e cascaded
receivers are determined by t$e noise coe!!icient -F1. o! t$e receivers at t$e
!irst level.
'o#ever" #$en t$e to#er am)li!ier sto)s #or?in" because t$e loss is )resent
on du)le:er and by)ass connectors" about %dB o! redundant loss is introduced
on reverse lin?.
Accordin to t$e equation " t$e !ollo#in t#o assum)tions conclude t$e
reularity o! t$e e!!ect o! to#er am)li!ier on t$e base station system.
-1. Assum)tion 1
'ereunder is a series o! assum)tions*
F1 N %.1 dB -1.5582." noise coe!!icient o! t$e to#er am)li!ier
F% N 4.1 dB -%.8184." noise coe!!icient o! t$e base station
G N % -11.849. dB" to#er am)li!ier ain
Goss o! t$e !eeder and ot$er )assive devices N 2 dB -%.
Gain o! t$e !eeder and ot$er )assive devices G0 N 02 dB -1=%.
Eoise coe!!icient o! t$e !eeder and ot$er )assive devices F0 N 1=G0
9$en t$e to#er am)li!ier is not added" t$e noise coe!!icient o! t$e base
station receivin system #it$ t$e antenna out)ut end as re!erence )oint is as
!ollo#s*
F N F0 J -F%01.=G0 N 10Mlo -% J -%.818401.=0.1. N5.1dB
9$en t$e to#er am)li!ier is added" t$e noise coe!!icient o! t$e base station
receivin system #it$ t$e antenna out)ut end as re!erence )oint is as !ollo#s*
F N F1 J -F0 0 1.=G J -F% 0 1.=-GMG0. N 10Mlo-1.5582 J -% 0 1.=11.849 J -%.8184 0
1.=-11.849 C 0.1. N 2.%dB
At t$is time" t$e added to#er am)li!ier im)roves t$e noise coe!!icient" and
FDelta is 4.2dB" t$at is" t$e u)lin? is im)roved by 4.2 dB.
-%. Assum)tion %
'ereunder is a series o! assum)tions*
F1 N %.% dB -1.3193." noise coe!!icient o! t$e to#er am)li!ier
F% N%.2 dB -1.398%." noise coe!!icient o! t$e base station
G N 1% -11.849. dB" to#er am)li!ier ain
Goss o! t$e !eeder and ot$er )assive devices N 2 dB -%.
Gain o! t$e !eeder and ot$er )assive devices G0 N 02 dB -1=%.
Eoise coe!!icient o! t$e !eeder and ot$er )assive devices F0 N 1=G0
9$en t$e to#er am)li!ier is not added" t$e noise coe!!icient o! t$e base
station receivin system #it$ t$e antenna out)ut end as re!erence )oint is as
!ollo#s*
F N F0 J -F% 0 1.=G0 N 10Mlo -% J -1.398% 0 1.=0.1. N 1.2dB
9$en t$e to#er am)li!ier is added" t$e noise coe!!icient o! t$e base station
receivin system #it$ t$e antenna out)ut end as re!erence )oint is as !ollo#s*
F N F1 J -F0 0 1.=G J -F% 0 1.=-GMG0. N 10Mlo-1.3193J-% 0 1.=11.849 J -1.398% 0
1.=-11.849 C 0.1.. N %.3dB
At t$is time" t$e added to#er am)li!ier im)roves t$e noise coe!!icient" and
FDelta is %.5 dB" t$at is" t$e u)lin? is im)roved by %.5 dB.
Accordin to t$e )revious calculation" t$e !ollo#in conclusions can be
obtained*
+$e to#er am)li!ier im)roves t$e noise coe!!icient o! t$e base station receivin
system" t$us im)rovin t$e receivin sensitivity o! t$e base station.
+$e to#er am)li!ier im)roves u)lin? sinals e!!ectively" #$ic$ is also $el)!ul
!or im)rovin t$e receivin sensitivity o! t$e base station.
+$e ain o! t$e antenna am)li!ier reduces t$e e!!ect o! t$e com)onents
installed be$ind t$e to#er am)li!ier aainst noise coe!!icient.
9$en t$e !eeder is lon and t$e loss o! t$e !eeder is reat" i! t$e to#er
am)li!ier is added" t$e noise coe!!icient o! t$e base station receivin system
and t$e u)lin? sinals #ill be reatly im)roved.
+$e smaller t$e noise coe!!icient o! t$e to#er am)li!ier is" i! t$e to#er
am)li!ier is added" t$e reater t$e noise coe!!icient o! t$e base station
receivin system is im)roved. 'o#ever" i! t$e noise coe!!icient o! t$e to#er
am)li!ier is too reat" it may cause t$e noise coe!!icient o! t$e base station
receivin system to deteriorate.
9$en t$e receivin sensitivity o! t$e base station is reat and t$e !eeder is
s$ort" t$e to#er am)li!ier ma?es a little im)rovement on t$e noise coe!!icient
o! t$e base station.
>! t$e to#er am)li!ier im)roves t$e base station sensitivity" t$e base station is
more sensitive to outside inter!erence.
%.3.4 Cell Coverae ;stimation
>n actual )roFect )lannin" t$e e!!ective coverae area o! a base station larely
de)ends on t$e !ollo#in !actors*
;!!ective base station transmit )o#er
9or?in band -900M'( or 1800M'(. to be used
Antenna ty)e and location
@o#er budet
<adio )ro)aation environment
Carriers8 coverae requirements
Based on t$e inde:es o! QoS !or t$e mobile net#or? and t$e actual
a))lications" t$is section introduces t$e coverae area o! t$e base station in
di!!erent environments t$eoretically.
>! t$e !ollo#in assum)tions are )resent*
+$e antenna $ei$t o! GSM 900M'( and GSM 1800M'( base stations are 20
meters.
+$e sensitivities o! t$e GSM900 M'( %9 -22 dBm. mobile station and GSM
1800M'( 19 -20 dBm. mobile station are ,10% dBm and ,100 dBm res)ectively.
+$e mobile station $ei$t is 1.1 meters and t$e ain is 0 dB.
9$en t$e combiner and divider unit -CD/. is used" t$e sensitivities o! t$e
900M'( base station and 1800M'( base station are ,110dBm and ,108dBm
res)ectively.
+$e CD/ loss is 1.1dB" and t$e SC/ loss is 3.8dB.
+$e ain o! t$e 31,deree directional antenna is 12dBd !or t$e 900 M'( mobile
station and 13dBd !or t$e 1800M'( mobile station.
+$e !eeder is 10m in lent$. For 900M'( sinals" t$e !eeder loss is
4.02dBm=100m. For 1800M'( sinals" t$e !eeder loss is 1.85dB=100m.
>n eneral cities" select A?umura )ro)aation model.
Eo to#er am)li!ier and t$e do#nlin?s are restricted accordin to t$e
calculation o! t$e u)lin? and do#nlin? balance.
Accordin to t$e )revious assum)tions" t$e calculated results are as !ollo#s*
-1. Autdoor coverae radius o! t$e 900 M'( base station in urban areas
+$e minimum received level o! t$e mobile station dBm. +$e coverae radius is
calculated accordin to t$e ma:imum +<O transmit )o#er. +$e ma:imum +<O
transmit )o#er !or t$e 900 M'( base station 9 -43 dBm..
+$e ;><@ o! t$e base station antenna is*
-dBm.
'ere"
GCAM indicates t$e combiner loss
Gb! indicates t$e !eeder loss
Gab indicates t$e antenna ain o! t$e base station
And t$e allo#ed ma:imum )ro)aation loss is*
-dB.
Accordin to t$e A?umura )ro)aation model introduces earlier"
'ere"
indicates t$e antenna $ei$t o! t$e base station.
indicates t$e antenna $ei$t o! t$e mobile station.
6!7 N 900 M'(.
-dB.
Accordin to t$e )revious ?no#n number" t$e outdoor coverae radius o! t$e
900 M'( base station in urban areas can be obtained" t$at is" d N %.8?m.
-%. Coverae radius o! t$e 900 M'( base station in urban buildins
+$e minimum received level o! t$e mobile station -dBm..
-dB.
+$ere!ore" t$e coverae radius o! t$e 900 M'( base station in urban buildins
can be obtained" t$at is" d N 0.51?m.
>! t$e )revious assum)tions are )resent" t$is indicates t$at t$e 900 M'( base
station can cover t$e outdoor areas %.8 ?m a#ay" but !or t$e subscribers on t$e
!irst !loor o! t$e buildins 510 m a#ay" t$e quality o! t$e received sinals is not
satis!yin.
-2. Coverae radius o! t$e 900 M'( base station in suburban areas
+$e minimum received level o! t$e mobile station -dBm..
-dB.
+$e A?umura )ro)aation model in suburban areas must be modi!ied as
!ollo#s*
+$ere!ore" t$e coverae radius o! t$e 900 M'( base station in urban areas can
be obtained" t$at is" d N 1.4?m" so it is obvious t$at t$e coverae radius o! t$e
base station #it$ t$e same con!iuration is larer in suburban areas t$at in
urban areas.
-4. Autdoor coverae radius o! t$e 1800 M'( base station in urban areas
+$e minimum received level o! t$e mobile station -dBm.. Because t$e
ma:imum transmit )o#er o! t$e 1800 M'( +<O is 409 -43dBm." t$e coverae
radius is calculated based on t$is ma:imum transit )o#er.
-dBm.
-dB.
For t$e 1800 M'( base station" t$e A?umura )ro)aation model is*
>n addition" ! N 1800 M'( and -dB..
Accordin to t$e )revious ?no#n number" t$e outdoor coverae radius o! t$e
1800 M'( base station in urban areas can be obtained" t$at is" d N 1.5?m.
-1. Coverae radius o! t$e 1800 M'( base stations in urban buildins
+$e minimum received level o! t$e mobile station -dBm..
-dB.
>! t$e )revious assum)tions are )resent" t$is indicates t$at t$e 1800 M'( base
station can cover t$e outdoor areas 1.5?m a#ay" but !or t$e subscribers on t$e
!irst !loor o! t$e buildins 100m a#ay" t$e quality o! t$e received sinals is not
satis!yin.
%.3.1 Base Station Address @lannin
>. Avervie#
9$en )lannin base station addresses" !irst you must estimate t$e number o!
t$e base stations needed in various coverae areas accordin to t$e coverae
distance and t$e divisions o! t$e coverae areas. For t$e convenience o!
)rediction and emulation" you must )lan an initial layout t$e base station
addresses #it$ t$e $el) o! ma)s and t$e estimated results.
>>. @lannin met$ods
+$e base station address can be )lanned based on standard irds" or it can be
)lanned !rom a s)eci!ic area.
-1. @lan base station address based on standard rids
First you set t$e base stations in t$e coverae areas accordin to t$e distance
o! t$e standard rids" and t$en adFust t$e address layout and )roFect
)arameters accordin to t$e estimated coverae results to meet t$e coverae
requirement. A!ter t$at" continue t$e )lannin accordin to t$e !ollo#in
instructions*
>! a satis!yin address layout is obtained" you must analy(e t$e ca)acity o! t$e
base stations to be )lanned accordin to t$is layout" and determine t$e
reasonable number o! base stations. 9$en desinin t$e ca)acity" you must
calculate t$e number o! +<Os needs to be con!iured !or eac$ base station" and
t$en analy(e and adFust t$e con!iuration o! t$e base station accordin to t$e
number o! t$e con!iured +<Os.
+$e adFustment o! t$e con!iuration o! t$e base station is determined by
subscriber distribution. >! t$e number o! base stations in some areas does not
meet ca)acity requirement" anot$er base stations must be added.
-%. @lan base station address based on a s)eci!ic area
Accordin to t$is met$od" you are required to start t$e )lannin !rom t$e areas
#$ere t$e subscribers are most densely distributed or t$e )lannin #or? is
quite $ard to be )er!ormed. As a result" you must !ully survey t$e subscriber
distribution" land!orms" and round obFectives #it$in t$e coverae area to
)osition t$e ?ey coverae area #$ere t$e center base stations s$ould be
)lanned. And t$ese center base stations !unction as ensurin t$e coverae and
ca)acity in im)ortant areas.
A!ter t$e layout o! t$ese center base stations is determined" you can )lan
ot$er base station addresses accordin to coverae and ca)acity taret. And
t$is is $o# t$e !inal layout o! t$e base station addresses come !rom. A!ter t$e
overall solution is determined" t$e subsequent ste)s are )er!ormed accordin
to t$e !irst )lannin met$od.
^ Eote*
+$e di!!erence o! t$e tra!!ic intensity and t$e abnormality o! t$e land!orms
and round obFectives result in irreularity o! t$e radio coverae. +$ere!ore"
t$e distance bet#een base stations varies. Generally" t$is distance is smaller in
t$e areas #$ere tra!!ic intensity is reat. >n some $ot areas" you can ensure
t$e system ca)acity by usin micro cells and distributed antennas to )rovide
multi,layer coverae.
For restrictions !rom !requency resources are )resent" you must consider
avoidin inter!erence #$ile ensurin system ca)acity.
+$ere is no standard available !or t$e layout o! t$e base station addresses. A
ood )lannin solution is selected based on t$e interated )er!ormance o! t$e
net#or?.
%.3.3 Coverae @rediction
+$e coverae )rediction is to )redict t$e coverae o! t$e net#or? to be
constructed accordin to t$e selected base station addresses" desined base
station ty)es" suitable electronic ma)s" and net#or? )lannin tools to Fude
#$et$er t$e coverae meet t$e requirements o! t$e subscribers.
+$e coverae o! a base station is determined by t$e !ollo#in !actors*
>nde:es o! QoS
Aut)ut )o#er o! transmitters
Available sensitivity o! receivers
Direction and ain o! antennas
9or?in bands
@ro)aation environment -suc$ as land!orms" city constructions.
A))lication o! diversity rece)tion
>! t$e )redicted results o! t$e net#or? coverae !ail to meet t$e requirements"
you can ta?e t$e !ollo#in adFustin measures*
9$en t$ere are subscribers distributin beyond t$e cell coverae area" but it is
not economical !or you to install a base station" you can use a re)eater to
ensure t$e requirement o! t$ose subscriber.
9$en t$e sinals are #ea? or blind (ones are )resent #it$in t$e coverae area"
you can consider #$et$er to use micro cells accordin to actual conditions.
>! a lare blan? area is )resent bet#een nei$bor cells" you can increase t$e
antenna $ei$t and add base stations accordin to t$e )rinci)les o! cell
s)littin.
9$en t$e cell coverae area !ails to meet t$e co,c$annel inter!erence inde:"
you can adFust t$e !requency con!iuration o! t$e cell" adFust base station
addresses" or adFust desin o! t$e )arameters" suc$ as antenna s)eci!ication"
antenna $ei$t" a(imut$ anle" tilt anle" and transmit )o#er.
^ Eote*
9$en ta?in t$ese adFustin measures" you must consider t$e mutual e!!ect
bet#een base stations.
%.5 Desin o! Base Station Address
%.5.1 Address desin
Generally" in GSM radio net#or? )lannin" t$e base station address is desined
accordin to t$e !ollo#in requirements*
+$e address must serve to t$e reasonable cell structure.
Based on t$e com)re$ensive analysis o! t$e electronic ma)s and )a)er ma)s"
you can select several candidate addresses !rom t$e )ers)ective o! coverae"
anti,inter!erence" and tra!!ic balance.
>n actual conditions" carriers are required to discuss t$e selected addresses
#it$ o#ners. Generally" t$e addresses must be located #it$in t$e area 1=4
radius o! t$e cellular base station.
Durin t$e early construction stae #$en only a !e# base stations are installed"
t$e base stations must be located in t$e center o! t$e areas #$ere subscribers
are densely )o)ulated.
For t$e selection o! t$e base station addresses" t$e )riority must be iven to
t$e im)ortant areas" suc$ as overnment o!!ices" air)orts" train stations" ne#s
center" and reat $otels so t$at ood conversation quality can be assured.
Furt$ermore" overla))ed coverae must be avoided in t$ese areas.
For ot$er coverae areas" t$e base station addresses are desined accordin to
standard cellular structures. For t$e suburban areas" $i$roads" and
countryside areas" t$e desin o! base station addresses $as little relation #it$
cellular structures.
9it$out a!!ectin t$e layout o! base stations" you can select t$e
telecommunication buildins and )ost o!!ices as t$e base station addresses so
t$at t$e !acilities" suc$ as t$e equi)ment room" )o#er su))lier" and iron to#er
can be !ully utili(ed.
+$e direction o! antenna maFor lobe must be in accordance #it$ t$e area
#$ere t$e tra!!ic intensity is reat. >n t$is case" t$e sinal strent$ o! t$e area
can be en$anced" so does t$e conversation quality. Mean#$ile" t$e direction o!
t$e antenna maFor lobe must be deviated !rom intra,!requency cells so t$at t$e
inter!erence can be controlled e!!iciently.
>n urban areas" it is recommended t$at t$e overla))ed de)t$ o! t$e antennas
in adFacent sectors cannot e:cel 10L. >n suburban areas and small to#ns" t$e
overla))ed de)t$ bet#een coverae areas cannot be too reat" and t$e
included anle bet#een sectors must be equal to or $i$er t$an 90`.
>n addition" !or actual desin" you must consider t$e ma))in relations$i)
bet#een carrier number and cells. Generally" more carriers are con!iured !or
t$e cells #it$ $i$ intensity.
+$e a(imut$ anle must be desined accordin to not only t$e tra!!ic
distribution in t$e areas around t$e base stations" but also t$e )er!ormance o!
t$e overall net#or?.
Generally" it is recommended to ado)t t$e same a(imut$ anle !or t$e 2,sector
base stations in urban areas so t$at t$e com)licated net#or? )lannin can be
avoided a!ter cell s)littin in t$e !uture. Moreover" t$e antenna maFor lobe
cannot directly )oint to t$e strai$t streets in )o)ulated urban areas" because
it can cause cross,coverae.
>n t$e areas connectin urban and suburban areas" and alon trans)ort arteries"
you must adFust t$e a(imut$ anle accordin to coverae taret.
Generally" t$e base station address is not considered on t$e $i$ mountains in
urban and suburban areas. +o be more s)eci!ically" t$e $i$ mountains are
t$ose over %00 to 200 meters $i$er t$an above t$e sea,level.. At$er#ise" not
only stron inter!erence and #ea? sinals may be )resent #it$in t$e coverae
area" but also t$e base stations are $ard to be installed and maintained on $i$
mountains.
Ee# base stations must be installed at t$e s)ots #$ere t$e tra!!ic is
convenient" t$e )o#er su))ly is available" and t$e environment is secure. >n
contrast" ne# base stations must not be installed at t$e s)ots near t$e radio
transmit stations #it$ $i$ )o#er" radar stations" and ot$er equi)ments #$ic$
)roduces reat inter!erence" because t$e inter!erence,!ield intensity cannot be
reater t$an t$at de!ined by t$e base station.
+$e base station addresses must be !ar a#ay !rom !orests or #oods to ?ee) t$e
receivin sinals !rom !adin.
+$e transmission bet#een base station controllers must be considered in t$e
desin o! t$e base station address.
9$en selectin a base station address !rom $i$ buildins in urban areas" you
can divide t$e net#or? into several layers #it$ t$e $el) o! t$e buildin $ei$t.
+$e antenna $ei$t o! maFor base stations must be a little $i$er t$an t$e
averae $ei$t o! buildins. Generally" t$e antenna $ei$t o! t$e base stations
in )o)ulated urban areas ranes !rom %1 to 20 meters. >n suburban areas -or
t$e antenna )oints to suburban areas." t$e antenna $ei$t ranes !rom 40 to 10
meters.
Alon $i$roads or in mountain areas" t$e base station address is selected
based on !ull survey o! t$e land!orms. For e:am)le" t$e address can be
determined in an o)en area or at t$e turns o! t$e $i$roads.
9$en selectin a base station address !rom t$e cities c$aracteri(ed by
mountains and $ills and !rom t$e areas #$ere $i$ buildins are constructed
#it$ metals" you must consider t$e e!!ect o! time dis)ersion. >n t$is case" t$e
base station address must near re!lected obFectives. 9$en t$e base station is
!ar a#ay !rom re!lected obFectives" you must adFust t$e directional antenna to
t$e reverse direction o! t$e re!lected obFectives.
Caution*
+ime dis)ersion mainly re!ers to t$e intra,!requency inter!erence arisin !rom
t$e time di!!erence bet#een t$e master sinal and ot$er multi)at$ sinal
arrivin at t$e receiver in terms o! s)ace transmission. Accordin to t$e
requirements in GSM )rotocols" t$e equali(er o! t$e receiver must carry t$e
time #indo# #it$ 13Ks -equivalent to 4.8 ?m.. +$e multi)at$ sinal #it$ time
di!!erence reater t$an 13 Ks is rearded as intra,inter!erence sinal. >n t$is
case" you must consider #$et$er t$e level di!!erence bet#een t$e master
sinal and multi)at$ sinal meet t$e carrier,to,inter!erence ratio -C=>."
namely" t$e master sinal is 1% dB reater t$an t$e multi)at$ sinal at least.
%.5.% @roFect @arameter Decision
A!ter !inis$in desinin a base station address" you must decide t$e )roFect
)arameters needed !or t$e base station installation. +$ese )arameters include*
Gatitude and lonitude o! t$e location o! base station antenna
Antenna $ei$t
Directions o! t$e antenna
Antenna ain
A(imut$ anle
+ilt anle
Feeder s)eci!ications
+ransmit )o#er !or eac$ cell o! t$e base station
And t$e )revious )arameters are decided t$rou$ !ield survey.
Be!ore beinnin !ield survey" you must !amiliari(e yoursel! #it$ t$e overall
)roFect and collect t$e materials and tools relative to t$e )roFect. +$ey are*
All ty)es o! )roFect documents
Bac?round in!ormation
>n!ormation about t$e e:istin net#or?
Gocal ma)
Con!iuration lists required in contracts
<elative tools -includin diital camera" G@S" com)ass" ruler" and la)to)
com)uter.
^ Eote*
Ma?e sure t$at all t$e materials and tools are usable be!ore settin out.
+$e !ollo#in items must be em)$asi(ed be!ore !ield survey*
+$e G@S must be )laced in an o)en land to )osition t$e latitude and lonitude
o! a base station
Ma?e a detailed record o! t$e surroundins around t$e base station" suc$ as t$e
distribution o! t$e buildins" !acilities #it$ stron inter!erence" and t$e
equi)ments s$arin t$e same base station address.
>t is better to record t$e )revious in!ormation #it$ a camera.
@revent t$e com)ass !rom maneti(in" because t$e maneti(ation #ill cause
reat deviation durin t$e measurement.
Field survey determines t$e layout o! t$e base station addresses ultimately.
+$e !ield survey !or t$e base station includes o)tical measurement" s)ectrum
measurement" and base station address survey. +$ey are s)eci!ied as !ollo#s*
A)tical measurement
Measure i! a barrier t$at may re!lect electrical #aves around t$e base station"
suc$ as $i$ buildins.
S)ectrum measurement
C$ec? i! t$e electromanetic environments around t$e base stations are normal
at )resent or in recent days.
Base station address survey
C$ec? t$e installation conditions o! antenna and equi)ments" )o#er su))ly"
and natural environment.
+$e !ollo#in sections introduce t$e desin !or antenna installation.
>. ;nvironment !or antenna installation
+$e environment !or antenna installation can be divided into t$e environment
near t$e antenna and t$e base station. For t$e environment near t$e antenna"
you must consider t$e isolation bet#een antennas and t$e e!!ect o! iron to#er
and buildins aainst t$e antenna. For t$e environment near t$e base station"
you must consider t$e e!!ect t$e $i$ buildins #it$in 100 meters aainst t$e
base station. 'o#ever" i! t$e $ei$t o! t$e buildins is )ro)erly used" you can
obtain t$e intended coverae area.
>! a directional antenna is installed on t$e #all" t$e radiation direction o! t$e
antenna is )er!ectly )er)endicular to t$e #all. >! its a(imut$ anle must be
adFusted" t$e included anle bet#een t$e radiation direction and t$e #all is
required to be reater t$an 51`. >n t$is case" i! t$e !ront,to,bac? ratio o! t$e
antenna is reater t$an %0 dB" t$e e!!ect o! t$e sinals re!lected by t$e #all in
reverse direction aainst t$e sinals in t$e radiation direction is quite sli$t.
9$en installin an antenna" you must consider #$et$er lare s$ado#s #ill be
)resent #it$in t$e coverae area o! t$e antenna. +$e s$ado#s are )roduced
mainly because t$e base station is surrounded by some $ue barriers" suc$ as
$i$ buildins and reat mountains. +$ere!ore" t$e antenna must be installed
in t$e areas #it$ no suc$ barriers.
9$en a directional antenna is installed on buildin roo!s" you must )revent t$e
buildin edes !rom barrin t$e radiation o! antenna beams. +$ere!ore" to
reduce or ease t$e s$ado#" you can install t$e antenna near buildin edes.
Because t$e buildin roo!s are diversi!ied and com)licated" i! an antenna must
be installed !ar a#ay !rom buildin edes" t$e antenna must be installed $i$er
t$an t$e roo!. >n t$is case" t$e #ind load o! t$e antenna must be considered.
>>. Antenna isolation in GSM system
+o avoid inter,modulation inter!erence" you must leave certain isolation
bet#een t$e receiver and transmitter o! t$e GSM base station" namely" +: , <:*
20 dB and +: ,+:* 20 dB. +$ey are a))licable to t$e situation t$at a GSM
900M'( base station and a GSM 1800M'( base station s$are t$e same address.
+$e antenna isolation de)ends on t$e radiation diaram" s)ace distance" and
ain o! t$e antenna. Generally" t$e attenuation introduced by t$e voltae
standin #ave ratio -HS9<. is not considered. +$e antenna isolation is
calculated as !ollo#s*
For vertical arranement" Gv N %8 J 40l -?=a. -dB.
For $ori(ontal arranement" Gv N%% J %0l -d=a. 0 -G1JG%. 0 -S1 J S%. -dB.
'ere"
Gv indicates t$e required isolation.
a indicates t$e lent$ o! carrier #aves.
? indicates t$e vertical isolation distance.
d indicates t$e $ori(ontal isolation distance.
G1 indicates t$e ains o! t$e transmitter antenna in t$e ma:imum radiation
direction" in t$e unit o! dBi.
G% indicates t$e ains o! t$e receiver antenna in t$e ma:imum radiation
direction" in t$e unit o! dBi.
S1 indicates t$e levels o! t$e side lobes o! t$e transmitter antenna in t$e 90`
direction" in t$e unit o! dB)" and it is a neative value relative to t$e main
beam.
S% indicates t$e levels o! t$e side lobes o! t$e receiver antenna in t$e 90`
direction" in t$e unit o! dB)" and it is a neative value relative to t$e main
beam.
+$e !ollo#ins introduce t$e requirements on t$e antenna mount in GSM
900M'( and GSM 1800M'(.
-1. Directional antenna
>n one system" t$e !ollo#in requirements must be met in terms o! isolation*
+$e $ori(ontal distance bet#een t#o antennas in t$e same sector must be
equal to or reater t$an 0.4m.
+$e $ori(ontal distance bet#een t#o antennas in di!!erent sectors must be
equal to or reater t$an 0.1m.
>n di!!erent systems" t$e !ollo#in requirements must be met #$en t#o
antennas are in t$e same sector and direction*
+$e $ori(ontal distance bet#een t$e t#o antennas must be equal to or reater
t$an 1m.
+$e vertical distance bet#een t$e t#o antennas must be equal to or reater
t$an 0.1m.
+$e distance bet#een t$e bottom o! t$e antennas and t$e enclosin #all o!
buildin roo! must be equal to or reater t$an 0.1m.
+$e included anle bet#een t$e line connectin t$e bottom o! t$e antenna to
t$e antenna,!acin roo! and t$e $ori(ontal direction must be reater t$an 11`.
+$e bands o! t$e t#o systems are close to eac$ ot$er" t$e inter!erence aainst
eac$ ot$er #ill easily occur. Mostly" t$e transmission o! CDMA%000 1O base
station #ill inter!ere #it$ t$e rece)tion o! GSM 900M'( base station.
+$e disclosure sinals o! t$e CDMA band !allin into t$e c$annels o! t$e GSM
base station receivers #ill en$ance t$e noise level o! t$e GSM receivers. >n t$is
case" t$e GSM u)lin?s become #ea?" #$ic$ #ill reduce t$e coverae area o!
t$e base station and #orsen t$e quality o! t$e net#or?.
>! t$ere is not enou$ isolation bet#een base stations or t$e transmittin !ilter
inter!erin base stations does not )rovide enou$ out,o!,band attenuation" t$e
sinals !allin into t$e band o! t$e inter!ered base station receiver may stron"
#$ic$ #ill increase t$e noise level o! t$e receiver.
+$e deterioration o! t$e system )er!ormance is closely related to t$e strent$
o! inter!erence sinals" and t$e strent$ o! inter!erence sinals is determined
by t$e !actors" suc$ as t$e )er!ormance o! t$e transmittin elements o! t$e
inter!erin base stations" t$e )er!ormance o! t$e receivin elements o! t$e
inter!ered base stations" t$e distance bet#een bands" and t$e distance
bet#een antennas.
+$e sinal !rom t$e am)li!ier o! t$e inter!erin base station is !irst sent to t$e
transmittin !ilter" and t$en it attenuate due to t$e isolation bet#een t$e t#o
base stations. Finally" it is received by t$e receiver o! t$e inter!ered base
station. +$e )o#er o! t$e s)urious inter!erence arrivin at t$e antenna end o!
t$e inter!ered base station can be e:)ressed by t$e !ollo#in equation*
'ere"
>b indicates t$e inter!erence level received at t$e antenna receivin end o! t$e
inter!ered base station" in t$e unit o! dBm.
@+O,AM@ indicates t$e out)ut )o#er at t$e am)li!ier o! t$e inter!erin base
station" in t$e unit o! dBm.
@attenuation indicates t$e out,o!,band su))ression attenuation at t$e
transmittin !iler.
>isolation indicates t$e isolation bet#een t$e antennas o! t$e t#o base
stations" in t$e unit o! dB.
9Binter!ered indicates t$e band#idt$ o! t$e sinals at t$e inter!ered base
station.
9Binter!erin indicates t$e measurable band#idt$ o! t$e inter!erin sinals" or
it can be understood as t$e band#idt$ de!ined by s)urious radiation.
<eulate t$e )revious equation and t$e !ollo#in equation can be obtained*
Su))ose t$e transmit c$annel number o! CDMA%000 1O is t$e last one on its
#or?in band" t$at is" 858.49M'(" t$e s)urious sinal level on t$e band o! 890,
911M'( must be equal to or lo#er t$an ,12dBm=100?'(. >! you intend to )ut
t$is assum)tion into )ractice" you can !ilter and combine eac$ transmitted
c$annel number by usin band,limited !ilter #it$ a band#idt$ o! only 1.%2M'(.
+$e band,limited !ilter o! t$is ty)e $as reat out,o!,band attenuation" #$ic$
can reac$ 13 dB at 890 M'( and 80 dB at 909 M'(. 'ere you must consider t$e
#orst situation" t$at is" t$e !requencies at t$e $i$est end o! t$e CDMA system
inter!ere #it$ t$e !requencies at t$e lo#est end o! t$e GSM system.
>n t$is case" >isolation N -,12dBm=100?'(. , 13 , >b J 10l -%00?'(=100?'(.
'ere >b indicates t$e $i$est inter!erence level -dBm. allo#ed by t$e receivin
end o! t$e inter!ered base station. >! t$e receivin sensitivity o! t$e inter!ered
base station is ensured" t$e outside inter!erence level are required to be 10 dB
lo#er t$an t$e bac? noise o! t$e receiver. >n t$is case" t$e sensitivity a!!ected
only accounts to about 0.1 dB.
+$e bac? noise o! t$e GSM receiver is t$e sum o! t$e noise intensity"
band#idt$" and noise coe!!icient. >! t$e noise coe!!icient is 8 dB" t$e bac?
noise is ,154Jnoise coe!!icientJ10l -%00000. N ,154J8J12 N ,112 -dBm..
+$ere!ore" t$e ma:imum s)urious inter!erence allo#ed is ,112,10 N ,1%2
-dBm=%00?'(..
As a result" t$e s)urious inter!erences !rom ot$er systems !allin at t$e GSM
receivers are required to be smaller t$an ,1%2 -dBm=%00?'(.8 ot$er#ise" t$e
s)urious inter!erences #ill seriously a!!ect t$e GSM system.
+$ere!ore" >isolation N -,12dBm=100?'(. 0 13 , >b J 10l -%00?'(=100?'(. N ,12,
13, -,1%2dBm=%00?'(. J 10l -%00?'(=100?'(. N 15 dBm=%00?'(.
+$at is" accordin to t$e assum)tion" t$e isolation bet#een a CDMA antenna
and GSM 900M'( antenna must be at least 15dB reardless #$et$er t$ey s$are
t$e address or not.
Many #ays can be used to reduce t$e inter!erence. For e:am)le" you can ado)t
t$e !ollo#in #ays*
Desin enou$ distance bet#een antennas
Filter t$e out,o!,band inter!erence o! t$e transmitter
Add di!!erent equi)ments to t$e !ilter" suc$ as receiver" du)le:er" and divider.
Accordin to t$e requirements in +>A=;>A,95 )rotocols" t$e s)urious
inter!erence !rom t$e CDMA antenna inter!ace !allin #it$in t$e GSM 900M'(
receivin bands must be less t$an ,12 dBm=100?'(. +$ere!ore" t$e )roblems"
suc$ as mutual inter!erence and co,address construction must be considered in
t$e initial desin.
+o be s)eci!ic" you can !ilter and combine eac$ transmitted c$annel number
usin a limited,band !ilter #it$ t$e band#idt$ o! only 1.%2 M'(. +$e band,
limited !ilter o! t$is ty)e $as reat out,o!,band attenuation" t$us t$e s)ace
distance bet#een t$e antennas o! t$e CDMA system and GSM system must be
s$ortened.
>n addition" to minimi(e t$e inter!erence" you must ?ee) suitable isolation
bet#een t$e antennas o! t$e CDMA system and GSM system.
+$e antenna isolation is calculated accordin to t$e !ollo#in t#o !ormulas"
#$ic$ $as been introduced earlier*
For vertical arranement" Gv N %8 J 40l -?=a. -dB.
For $ori(ontal arranement" Gv N%% J %0l -d=a. 0 -G1JG%. 0 -S1 J S%. -dB.
Accordin to t$e t#o !ormulas" t$e requirements on t$e isolation bet#een t$e
antennas o! CDMA system and GSM 900 M'( system are s)eci!ied in t$e
!ollo#in t$ree circumstances.
+$e antennas o! t$e CDAM system and GSM 900M'( system do not s$are t$e
same address" #it$ t$e antennas $ori(ontally o))osite to eac$ ot$er" or t$e
antennas o! t$e t#o systems s$are t$e same address" #it$ t$e antenna ty)e o!
omni antenna.
Su))ose t$e e!!ective ains o! t$e antennas o! t$e t#o systems in t$e ma:imum
radiation direction are 10 dBi -#it$ t$e !eeder loss considered." and t$e
inter!erence sinals are 890M'(" accordin to )revious analysis" t$e isolation
bet#een t$e CDMA system and GSM system is required at least 15dB.
+$ere!ore" t$e !ollo#in equation can be obtained accordin to t$e )revious
!ormula*
15 N %% J %0l -D$=a. 0 -10 J 10.
+$e antennas o! t$e CDMA and GSM 900 M'( system s$are t$e same address
-t$e antennas are installed on t$e same )lat!orm and $ori(ontally se)arated."
#it$ t$e antenna ty)e o! directional antenna.
Su))ose t$at t$e t#o antennas are $ori(ontally )laced" and t$eir tilt anle is
31`" and t$at t$e e!!ective ains o! t$e t#o antennas in t$e radiation direction
are 11dBi.
And i! t$e side lobe o! t$e 31`antenna is ,18dB in t$e $ori(ontal )lane" t$e
e!!ective ain o! t$e antenna in t$is direction is -11 0 18. dBi N ,2 dBi.
+$ere!ore" 15N%%%J0l -D$=a. , U-11J11. J [-,18. J -,18.\V.
Accordin to t$e )revious equation" t$e $ori(ontal distance bet#een t$e t#o
antennas are d N 9.1m.
+$e antennas o! t$e CDMA and GSM 900 M'( antennas s$are t$e same address
-t$e antennas are not installed on t$e same )lat!orms o! t$e iron to#er and
vertically se)arated." #it$ t$e antenna ty)es o! directional antenna and omni
antenna.
>n t$is case" t$e equation 15N%8 J 40 l -?=a. is )resent.
Accordin to t$is equation" t$e vertical distance bet#een t$e t#o antennas is d
N 1.8m.
^ Eote*
+$e )revious descri)tions are Fust t$eoretical detections. >n actual net#or?in"
ot$er ty)es o! antennas may be installed at t$e same address. >n t$is case"
some equi)ment inde:es must be considered" amon #$ic$ t$e im)ortant ones
are s)urious radiation" t$e inter!erence )o#er o! t$e inter!erin sinals to
inter!ered sinals" and t$e antenna isolation.
>H. >nstallation distance bet#een antennas
Diversity tec$noloy is t$e most anti,!adin e!!ective. 9$en t#o sinals are
irrelevant to eac$ ot$er" t$e $ori(ontal distance bet#een t$e diversity
antennas must be 0.11 times t$at o! t$e valid antenna $ei$t. +$e $i$er )lace
t$e antenna is installed" t$e larer t$e $ori(ontal distance bet#een diversity
antennas is. 9$en t$e distance bet#een diversity antennas is equal to or
reater t$an 3m" $o#ever" t$e antenna is $ard to be installed on an iron
to#er.
>n addition" t$e distance required by vertical diversity antennas is 1 to 3 times
t$at o! t$e $ori(ontal diversity antennas #$en t$e same coverae is ensured.
+$ere!ore" t$e vertical diversity antenna is seldom used in actual )roFects" but
antennas are o!ten vertically installed to meet isolation requirement"
es)ecially omni antennas are vertically installed.
>n addition" !or $i$road coverae" t$e line connectin t#o receivin antennas
must be )er)endicular to t$e $i$road. >! s)ace diversity is used" t$e diversity
distance is t$e )er)endicular. >solation requirement* +:,+:" +: , <:* 20 dB
+$e installation !or GSM 900M'( and GSM 1800M'( antennas is !le:ible" but no
matter #$at s)eci!ications are used" t$ey must meet t$e requirements on
isolation and distance. >n addition" in actual )roFects" barriers are )resent
bet#een antennas. For e:am)le" a to#er is al#ays )resent bet#een t#o omni
antennas" so you must s$orten t$e $ori(ontal distance bet#een t$em.
H. Desin o! base station )arameters in residential areas
A lare number o! residential areas are distributed in urban areas" so t$is
section introduces t$e desin o! base station )arameters in t$ese areas.
-1. Features o! residential areas
Buildin intensity
Great,intensity residential areas* t$e distance bet#een buildins is #it$in 10
meters.
Middle,intensity residential areas* t$e distance bet#een buildins ranes !rom
10 to %0 meters.
Go#,intensity residential areas* t$e distance bet#een buildins is larer t$an
%0 meters.
Construction material
+$e #alls o! t$e residential areas are constructed #it$ concretes.
+$e #alls o! t$e residential areas are constructed #it$ bric?s and concretes.
+$e #alls o! t$e residential areas are constructed #it$ $ollo# bloc?s.
Eotes*
+$e t$ic?ness o! t$e buildins varies #it$ t$e reions and climates. +$ree
s)eci!ications are available" namely" %4m" 45m" and 49m. Generally" t$e #alls
are t$ic?er in sout$ern )arts and t$inner in nort$ern )arts.
-%. Antenna installation in residential areas
+$e address #$ere t$e antenna s$ould be installed in residential areas is $ard
to be determined. Generally" #$en ado)tin micro cells" you can install t$e
antenna #it$in a residential area near to t$e taret coverae area.
>n t$is case" t$e antenna can be installed in t$e !ollo#in s)ots*
An outer #alls -not roo!s. o! a buildin
An )illars
>nstall a micro cell in underround araes
>! t$e antenna is installed at a #all corner" t$e maFor lobe o! t$e antenna can
radiate t$e s)ace bet#een buildins. Generally" t$e maFor lobe o! t$e antenna
cannot !ace t$e #alls o! t$e buildins nearby directly.
>! !requencies are reusable amon t$ese micro cells" t$e directions o! antennas
must be consistent #it$ eac$ ot$er. >n addition" you can also use t$e cell
s)litter to enable a cell to coverae t$e areas in t#o directions. >n t$is case"
$o#ever" t$e !requency utili(ation ratio may decrease and e:tra )o#er s)litter
#ill introduce loss o! 2 dB.
For t$e residential areas #it$ reular arranement" t$e directional antennas
#$ose $ori(ontal beam #idt$ is 90` to 1%0` and vertical beam #idt$ is reater
t$an 20` are recommended.
/nder certain conditions" t$e micro cell antenna can be installed on t$e )illars
#it$in a residential area.
For t$e residential areas #it$ irreular arranement" t$e antenna can be
installed on t$e #alls o! a buildin" so t$e re!lected #aves can coverae t$e
#alls o! o))osite buildins. >n t$is case" t$e antennas #$ose $ori(ontal beam
#idt$ is reater t$an 1%0`and vertical beam #idt$ is reater t$an 20`are
recommended.
-2. Antenna selection
9$en t$e #alls o! a buildin is selected as an installed )osition" you can use
t$e build,in antenna o! t$e micro cell directly" or ot$er antennas #it$ small
si(e. Accordin to coverae !eatures o! residential areas" #$en selectin t$e
s)eci!ications !or t$e micro cell antennas to be used" you must consider t$e
!ollo#in !actors*
Antenna ain
'ori(ontal beam #idt$
Hertical beam #idt$
@olari(ation mode
Hisual e!!ect -antenna si(e" s$a)e" and #ei$t.
+$e antenna ain is recommended less t$an 9 dBi !or micro cell antennas.
Because t$e coverae area o! a micro cell antenna is small and t$e installed
)osition is near to t$e coverae area" t$e antenna ain can be adFusted to a
smaller value" es)ecially i! t$e ain o! an antenna is reater t$an 10dBi" its si(e
is lare" #$ic$ may cause o))osition !rom residents.
+$e selection o! t$e $ori(ontal and vertical beam #idt$ !or an antenna is
related to radio environment. >! a micro cell antenna is installed on a #all" t$e
antenna $ei$t is lo#er t$an t$e averae $ei$t o! surrounded buildins. >n t$is
case" i! bot$ t$e indoor coverae o! lo#er !loors and $i$er !loors can be
assured" you must select t$e antennas #it$ reater vertical beam #idt$.
Accordin to t$e $ei$t o! buildins" you can select t$e directional antennas
#$ose vertical beam #idt$ ranes !rom 21`to 80`.
+$e selection o! t$e $ori(ontal beam #idt$ o! t$e micro cell antenna and t$e
installed )osition o! t$e antenna are related to coverae taret. >n t$is case"
you can select t$e directional antennas #$ose beam #idt$ ranes !rom 30` to
110`" or you can c$oose omni antennas or bi,directional antennas -8,s$a)ed
antennas..
Bot$ vertical )olari(ation antennas and dual )olari(ation antennas can be
selected !or a micro cell. +$e coverae area o! a micro cell in urban areas is
small" so t$e diversity rece)tion is unnecessary. >n t$is case" a vertical
)olari(ation antenna can meet t$e coverae requirements in residential areas.
As !or t$e dual )olari(ation antenna" $o#ever" it is e:)ensive and lare in si(e"
so it is not recommended.
+$e visual e!!ect must be em)$asi(ed !or t$e micro cell antennas installed in
residential areas. +$ey must be small and moderate. >n addition" t$ey must be
li$t !or installation convenience. >! t$e contract bet#een t$e color o! t$e
antenna and t$at o! t$e surrounded buildins is reat" you must color t$e
antenna #it$ t$e same color o! t$e buildins.
>n some cases" you s$ould consider ado)tin dual,band antennas. 9$en
selectin a small,si(ed antenna" you s$ould consider #$et$er its ma:imum
out)ut )o#er can bear t$e micro cell out)ut )o#er. 9$en ado)tin s$ort
Fum)ers instead o! 5=8 !eeders" you s$ould consider #$et$er t$e antenna
connector -E,s$a)ed male=!emale" 5=13 D>E $eader. matc$es t$e Fum)er
connector.
%.8 Gocation Area Desin
%.8.1 De!inition o! Gocation Area
>n GSM )rotocols" a mobile communication net#or? is divided into multi)le
service areas accordin to t$e codes o! location areas. +$us t$e net#or? )aes
a mobile subscriber t$rou$ )ain its location area.
Gocation area is t$e basic unit o! )ain areas in a GSM system. +$at is" t$e
)ain messae o! a subscriber is sent in all cells o! a location area. A location
area contains one or more BSCs" but it belons to one MSC only.
Fiure 1,12 s$o#s t$e division o! service areas.
Fiure 1,1 Division o! service areas
%.8.% Division o! location areas
+$e coverae area o! eac$ GSM @GME is divided into multi)le location areas" in
#$ic$ an MS is )ositioned. +$e si(e o! a location area" namely" t$e area
covered by a location area code -GAC." )lays a ?ey role in a GSM system.
+$ere!ore" t$is section mainly introduces t$e )rinci)le !or )lannin location
areas.
>. Dividin t$e location area accordin to t$e distribution and be$aviour o!
mobile subscribers
+$e distribution o! location areas in cities and suburbs is di!!erent. Generally"
suburban areas or counties occu)y inde)endent location areas. >n cities" t$e
distribution o! location areas is similar to a concentric circle. -+$e areas in t$e
internal circle can be divided into several location areas due to t$e
requirements on ca)acity. +$e concentric circle can be divided into several
!raments..
>n addition" i! t#o or more location areas are )resent simultaneously in a bi
city o! reat tra!!ic" t$e land!orms" suc$ as mountains and rivers #it$in t$is
city can be used as edes o! t$e location areas. >n t$is case" t$e overla))ed
de)t$ bet#een t$e cells o! t$e t#o location areas can be reduced. >! no suc$
land!orms available #it$in t$is city" t$e areas -suc$ as streets and s$o))in
centers. #it$ reat tra!!ic cannot be used as edes o! t$e location areas.
Generally" t$e ede o! a location area is oblique instead o! )arallel or
)er)endicular to streets. >n t$e intersected areas o! urban areas and suburban
areas" to avoid !requent location u)date" you must desin t$e edes o! location
areas near t$e outer base stations instead o! t$e base stations Fust installed at
t$e intersections.
>>. Calculatin coverae area and ca)acity o! a location area
>! t$e coverae area o! a location area is too small" t$e mobile station #ill
)er!orm !requent location u)date. >n t$is case" t$e sinalin !lo# in t$e system
#ill increase. >! t$e coverae o! a location area is too larer" $o#ever" t$e
net#or? #ill send a )ain messae in multi)le cells until t$e mobile station is
)aed. >n t$is case" t$e @C' #ill be overloaded and t$e sinalin !lo# at t$e
Abis inter!ace #ill increase.
+$e calculation o! location areas varies #it$ t$e )ain strateies desined by
di!!erent carriers. Durin early net#or? construction stae" t$e tra!!ic is not
reat" so a location area can accommodate more +<Os. 'o#ever" it is still
necessary !or you to monitor t$e @C' load and tra!!ic ro#t$. 9$en t$e tra!!ic
ro#s reat" you can en$ance t$e @C' ca)acity by addin a BCC' to t$e
system" but t$e number o! voice c$annels can be added is reduced by one
accordinly.
Generally" t$e ca)acity o! a location area is calculated as !ollo#s*
+$e number o! )ain bloc?s sent in eac$ second C t$e number o! )ain
messaes sent in eac$ )ain bloc? N t$e ma:imum )ain times in eac$
second. As a result" t$e number o! )ain times in eac$ $our" t$e tra!!ic
allo#ed in eac$ location area" and t$e number o! carriers su))orted in eac$
location area can be deducted.
+$e !ollo#ins introduce t$e items )resent in t$e )revious )arara)$
res)ectively.
-1. +$e number o! )ain bloc?s sent in eac$ second
1 !rame N 4.31ms" 1 multi!rame N 11 !rames N 0.%214s8 su))ose t$e number o!
access rant bloc?s is AGB" t$e number o! bloc?s" t$e number o! )ain bloc?s
sent in eac$ second is calculated by t$e !ollo#in !ormulas*
For non,combined BCC'" t$e number o! )ain bloc?s sent in eac$ second N -9
0 AGB.=0.%241 -)ain bloc?=second..
For combined BCC'" t$e number o! )ain bloc?s sent in eac$ second N -2 0
AGB.=0.%241 -)ain bloc?=second..
For non,combined BCC'" t$e AGB is % accordin to 'ua#ei BSC. +$ere!ore" t$e
number o! )ain bloc?s sent in eac$ second is %9.5 -)ain bloc?=second.8
#$en AGB is 0" it is 28.% -)ain bloc?=second..
For combined,BCC'" t$e AGB is 1" so t$e number o! )ain bloc?s sent in eac$
second is 8.1 -)ain bloc?s=second.8 #$en t$e AGB is 0" it is 1%.5 -)ain
bloc?=second..
Accordin to t$e )revious analysis" t$e larer t$e number o! AGB" t$e smaller
t$e number o! t$e )ain bloc?s sent in eac$ second and t$e smaller t$e )ain
ca)acity is. Moreover" t$e )ain ca)acity o! t$e combined BCC' is !ar less
t$an t$at o! t$e non,combined BCC'.
^ Eote*
Generally" a combined,BCC' cell and a non,combined,BCC' cell are not
con!iured simultaneously #it$in a GAC" and t$e number o! AGB must be
consistent #it$ a location area8 ot$er#ise t$e )ain ca)acity o! t$e location
area #ill decrease -no# t$e )ain ca)acity o! t$e cell #it$ t$e least )ain
ca)acity is t$e )ain ca)acity o! t$e location area..
'o#ever" i! t$e ca)acity o! a location area is small and t$e GAC resource is
scarce" you can con!iure t$e combined,BCC' cell and non,combined,BCC' cell
#it$in a GAC to enlare t$e number o! tra!!ic c$annels !or A1 and S111 base
stations.
-%. +$e number o! )ain messaes sent in eac$ )ain bloc? -O.
Accordin to section 9.1.%% o! GSM0408 )rotocols" eac$ )ain bloc? $as %2
bytes" and can send % >MS> )aes" or % +MS> and 1 >MS> )aes" or 4 +MS> )aes.
Accordin to t$e )ain strateies o! 'ua#ei MSC" i! t$e >MS> )ain mec$anism
is ado)ted" t$e number o! )ain messaes sent in eac$ )ain bloc?s is %
-)ain times=)ain bloc?.8 i! t$e +MS> )ain mec$anism is ado)ted" it is 4
-)ain times=)ain bloc?.
-2. +$e ma:imum )ain times in eac$ second -@.
+$e ma:imum )ain times in eac$ second is calculated by t$e !ollo#in t#o
!ormulas*
For non,combined BCC'" @ N -9 0 AGB.=0.%241 -)ain bloc?=second. C -)ain
times=)ain bloc?..
For combined BCC'" @ N -2 0 AGB.=0.%241 -)ain bloc?=second. C -)ain
times=)ain bloc?s..
>! t$e >MS> )ain mec$anism is ado)ted" !or non,combined BCC'" #$en AGB N
%" @ N 19.45 -)ain times=second.8 #$en AGB N 0" @ N 53.45 -)ain
times=second.. For combined,BCC'" #$en AGB N 1" @ N 13.99 -)ain
times=second.8 #$en AGB N 0" @ N %1.49 -)ain times=second..
>! t$e +MS> )ain mec$anism is ado)ted" !or combined BCC'" #$en AGB N %" @
N 118.91 -)ain times=second.8 #$en AGB N 0" @ N 11%.92 -)ain
times=second.. For combined BCC'" #$en AGB N 1" @ N 22.98 -)ain
times=second.8 #$en AGB N 0" @ N 10.98 -)ain times=second..
Accordin to t$e )revious analysis" t$e )ain ca)acity under >MS> )ain
mec$anism is $al! o! t$at under +MS> )ain mec$anism.
-4. +$e tra!!ic allo#ed in eac$ location area -+.
9$en desinin t$e ca)acity !or a location area" you must be attention t$at
t$e )ain ca)acity o! a location area cannot brea? its limit. For net#or?
e:)ansion" you can collect t$e times o! t$e busy,$our )ain orders delivered
by BSC !rom AMC" and t$en convert t$e times into t$e number o! )ain orders
sent in eac$ second.
>! no tra!!ic measurement data is available" suc$ as in t$e case o! ne# net#or?
construction" you can calculate t$e tra!!ic allo#ed in eac$ location area by
assumin a tra!!ic model.
For e:am)le" i! t$e averae conversation duration is 30s and t$e ratio o! t$e
times !or t$e mobile station to be success!ully )aed to t$e times o! total
)aes is 20L" t$e 30s o! conversation duration matc$es 1=30 calls -in t$e unit
o! second. ;rl." and 20L o! calls is enerated by t$e called )arties. +$ere!ore"
t$e success!ul calls o! t$e 20L mobile stations are 0.01 times -t$at is" 1=30M20L
N 0.001." in t$e unit o! second. ;rl.
>! t$e 51L o! t$e mobile stations res)ond to t$e !irst )ae and %1L res)ond to
t$e second )ae" t$e mobile stations res)ondin to t$e t$ird )ae can be
nelected. ->t is Fust an assum)tion" #$ic$ may be di!!erent !rom actual
conditions... +$ere!ore" 1.%1 )aes are needed i! a mobile station is
success!ully called eac$ time -%1L o! t$e )aes must be resent.. >n t$is case"
t$e !ollo#in equation is )resent*
R N 0.001M-1J%1L. N 0.003%1 )ain times=-second. ;rl.
Su))ose t$e conestion on )ain c$annels #ill occur #$en t$e )ain ca)acity
is 10L reater t$an ma:imum t$eoretical )ain ca)acity" t$e oriinal )ain
messaes are still )resent even t$e )ain queue is !ull in t$e B+S. >n t$is case"
t$e )ain ca)acity in one second is @M10L.
+$ere!ore" t$e tra!!ic allo#ed in eac$ location area can be calculated
accordin to t$e !ormula + N @M10L=R.
-1. +$e number o! carriers su))orted by eac$ location area -E+<O.
;ac$ +<O $ad 5.% +C's in averae" so t$e ma:imum tra!!ic o! eac$ +<O in eac$
$our is 5.%.
+$ere!ore" t$e number o! carriers su))orted in eac$ location area can be
calculated accordin to E+<O N +=5.% and t$e s)eci!ic values are listed in
All t$e )revious assum)tions do not include t$e e!!ect o! t$e )oint,to,)oint
s$ort messaes aainst on )ain ca)acity. >! t$e conversation times o! a
subscriber are equal to t$e number o! t$e s$ort messaes to be sent" and i! t$e
sent ratio and received ratio are consistent #it$ eac$ ot$er" t$e )ain
times=second. ;rl #ill double in busy $our and t$e ca)acity o! t$e location area
#ill reduce by $al!. +$ere!ore" some common s$ort messaes must be sent on
CBC'.
%.8.2 At$ers
+$is section introduces some ot$er in!ormation about location area desin.
+$e ca)acity o! a location area is closely related to )ain mec$anism" and is
directly related to t$e combinations o! AGB and BCC'. 9$en t$e combinations
o! AGB and BCC' are inconsistent #it$ eac$ ot$er in a location area" t$e
ca)acity o! t$e location area is determined by t$e cell #it$ t$e smallest
ca)acity. +$ere!ore" t$e combinations o! AGB and BCC' must be desined to
be consistent in location area )lannin.
>! t$e number o! )oint,to,)oint messaes ro#s lare immediately" t$e number
o! )ain messaes #ill increase" but t$e number o! su))orted subscribers #ill
decrease. >n t$is case" you must control and )rotect t$e !lo#s in t$e system.
Because t$e tra!!ic density varies #it$ location areas" it is recommended t$at
t$e combined,BCC' cells" non,combined,BCC' cells" and multi,BCC' cells !orm
a location area res)ectively. 9$en a cell #it$ BCC'=SDCC' combination" t$e
location area can be as lare as )ossible #$en t$e )ain ca)acity o! t$e B+S
does not reac$ t$e limit. 'o#ever" because all )ain messaes #ill be
broadcasted in all cells #it$in a location area" t$e cell #it$ BCC'=SDCC'
combination is t$e bottlenec? o! t$e location area.
+$e GAC is a ?ind o! number resource. +$ere!ore" you must coo)erate #it$
carries to )lan location areas.
%.9 Dual,Band Eet#or? Desin
%.9.1 Eecessity !or Constructin Dual,Band Eet#or?
+$e earlier GSM mobile communication net#or? is constructed on t$e 900 M'(
band. 9it$ ra)id ro#t$ o! subscribers" t$e net#or? ca)acity also ro#s
ra)idly. +$ere!ore" t$e lac? o! !requency resources and radio c$annels is a
maFor concern !or mobile telecommunications.
Many met$ods can be used to e:)and t$e ca)acity o! a GSM system" includin*
Addin macro cell base stations to t$e system
<educin distance bet#een base stations
Ado)tin aressive !requency reuse tec$noloies -suc$ as M<@ and 1C2.
Addin micro cells to t$e system
A))lyin $al! rate to t$e system
'o#ever" all t$ese met$ods cannot t$orou$ly solve t$e )roblems concernin
net#or? ca)acity. As a result" t$e GSM 1800M'( net#or? is introduced -u)lin?*
180101880 M'(8 do#nlin?* 151001581 M'(.. And t$e net#or? interatin GSM
900M'( and GSM 1800M'( can meet t$e ro#t$ o! net#or? ca)acity.
+$e a))lication o! GSM 1800M'( can brin t$e !ollo#in advantaes*
>t does not occu)y t$e bands o! GSM 900M'( and $as a communication
band#idt$ o! 51M. +$ere!ore" it brea?s t$e bottlenec? o! GSM 900M'( in terms
o! !requency resources.
+$e system net#or?in" )roFect im)lementation" net#or? )lannin" and
net#or? maintenance o! a GSM 1800M'( net#or? are almost t$e same #it$
t$at o! a GSM 900M'( net#or?.
+$e GSM 1800M'( and GSM 900 M'( can s$are a base station" so a GSM
1800M'( net#or? can be !inis$ed in a s$ort time" #$ic$ is quite $el)!ul !or
net#or? e:)ansion.
Dual,band mobile )$ones no# accounts !or a maFor )art o! t$e total" so a GSM
1800M'( net#or? can )rovide services to t$e dual,band subscribers. >n t$is
case" t$e ca)acity )ressure on GSM 900M'( can be reatly eased.
%.9.% GSM 1800M'( Coverae Solutions
>. @ro)aaiton !eatures o! GSM 1800M'(
+$e )ro)aation !eatures o! t$e electromanetic #aves o! 900 M'( and 1800
M'( are di!!erent in t$e !ollo#in as)ects*
+$e )ro)aation loss in !ree s)ace
+$e )ro)aation loss o! t$e 1800 M'( sinals is 3 dB reater t$an t$at o! t$e
900 M'( sinals in !ree s)ace.
@enetration loss
+$e )enetration loss o! t$e 900 M'( sinals is reater t$an t$at o! t$e 1800
M'( sinals" but t$eir di!!erence is sli$t.
Di!!raction loss
+$e loner t$e #aves" t$e smaller t$e di!!raction loss is. +$e di!!raction ability
o! t$e 1800 M'( sinals is )oorer t$an t$at o! t$e 900 M'( sinals.
>>. Dual,Band Eet#or?in Mode
+$ere are t$ree dual,band net#or?in modes" namely" inde)endent MSC
net#or?in" co,MSC=inde)endent BSC net#or?in" and co,BSC net#or?in"
amon #$ic$ t$e !ormer t#o are called inde)endent net#or?in" and t$e later
is called $ybrid net#or?in.
>>>. Coverae requirements on GSM 1800 M'(
Autdoor coverae
+$e outdoor coverae can be easily reali(ed #$en t$e distance bet#een base
stations are not lare. >n necessary cases" you can add a GSM 1800M'( base
station at t$e address o! t$e oriinal GSM base station. And in some )laces" you
s$ould consider add a ne# base station.
>ndoor coverae
+o ensure t$at t$e indoor coverae o! GSM 1800M'( is ood" you must control
t$e distance bet#een t$e base stations installed in urban areas #it$in 1000
meters. >n C$ina" $o#ever" t$e buildins in most cities are constructed by
concretes and metals" so t$e )enetration loss is reat. As result" t$e distance
bet#een base stations in urban areas o! C$ina ranes !rom 100 to 800 meters.
>H. Coverae mode o! GSM 1800M'(
-1. Scattered coverae in $ots)ot areas
At t$e early net#or? construction stae" t$e GSM 1800M'( base stations are
scattered in $ots)ot areas. 9$en t$e ca)acity con!iured !or a GSM 1800 M'(
base station is small" you must solve t$e )roblems" suc$ as SDCC' conestion"
+C' conestion" and !requent u)date bet#een GSM 1800M'( and GSM 900M'(.
+$e cost in early construction stae is small.
Scattered coverae o! GSM 1800M'( in $ots)ot areas
+$e coverae o! t$e dual,band net#or? o! t$is mode is based on t$e oriinal
GSM 900M'( net#or?. +$e GSM 1800M'( base station is constructed in some
$ots)ot areas" so t$e seamless coverae o! GSM 1800M'( is not available in t$is
case.
>! a dual,band mobile )$one starts conversation in an area covered by GSM
1800M'(" a!ter leavin t$is coverae area" it $ands over to t$e GSM 900M'(
cell #$ere it oriinally #as. And t$e $andover o! t$is ty)e is called t$e inter,
band $andover caused by coverae.
>! a dual,band mobile )$one starts t$e conversation in an area covered by GSM
900M'(" but because t$e tra!!ic in t$is area is reat" t$e mobile )$one #ill
$and over to an area covered by GSM 1800M'(. And t$e $andover o! t$is ty)e is
called t$e inter,band $andover caused by ca)acity.
+$e scattered coverae in $ots)ot areas only relieves ca)acity )roblems in a
s$ort term. Moreover" !requent inter,band !requency $andover increase t$e
sinalin load" #$ic$ results in t$e loss o! system ca)acity.
-%. Seamless coverae in $ots)ot areas
>! t$e coverae o! t$is mode is available8 t$e GSM 1800 M'( net#or? can s$are
reater tra!!ic !or GSM 900M'( net#or? and e:)and t$e system ca)acity. >n
addition" it is cost,e!!ective.
-2. @er!ect seamless coverae
>! a GSM 1800M'( net#or? ado)ts t$e coverae o! t$is ty)e" t$e advantaes are
as !ollo#s*
+$e seamless coverae area #it$in a city can be reali(ed.
+$e GSM 1800M'( net#or? can s$are t$e tra!!ic load !or GSM 900M'( net#or?
as muc$ as )ossible.
+$e system ca)acity can be reatly e:)anded.
+$e ratio o! t$e $andover bet#een layers is small.
+$e quality o! t$e net#or? is quite satis!yin.
+$e !requencies can be )lanned by )atc$.
+$e carriers can be e:)anded ste) by ste).
'o#ever" t$ere are still disadvantaes. +$ey are as !ollo#s*
+$e number o! base stations is lare.
+$e #or? load o! net#or? )lannin and o)timi(ation is $ue.
+$e investment is lare.
+$e base station addresses cannot be decided once.
@er!ect coverae o! GSM 1800M'( in $ots)ot areas
>! a GSM 1800M'( net#or? ado)ts t$is coverae mode" it can be easily
e:)anded to meet !uture coverae.
Com)ared #it$ t$e scattered coverae in $ots)ot areas" t$e )er!ect seamless
coverae is c$aracteri(ed by reat intensity and lare area. +$ere!ore" t$e
ratio o! inter,band $andover under t$is coverae mode is !ar smaller t$an t$at
under scattered coverae mode. As a result" t$e sinalin load is reduced
reatly.
+$ere!ore" t$is coverae mode is an ideal coverae solution. >! a GSM 1800M'(
net#or? ado)ts t$is coverae mode" it does not necessarily attac$ to t$e GSM
900M'( net#or?" instead" it can !orm an inde)endent net#or?.
%.9.2 Gocation Area Division !or Dual,Band Eet#or?
+$e location area division !or dual,band net#or? is suested as !ollo#s*
>! 1800 M'( cells and 900 M'( cells are under t$e control o! t#o MSCs
res)ectively" t$eir location areas are di!!erent. +$ere!ore" you must set related
)arameters to maintain t$e mobile stations stay in t$e 1800 M'( cells #$ere
t$e tra!!ic is absorbed. >n t$is case" t$e times !or t$e mobile station to
$andover bet#een t$e t#o bands and reselect cells #ill decrease. Mean#$ile"
#$en desinin sinalin c$annels" you must !ully consider t$e load resulted
!rom location u)date.
>! 1800 M'( cells and 900 M'( cells s$are a MSC" at t$e early net#or?
construction stae" t$ey are suested to use t$e same location area #it$out
a!!ectin t$e net#or? ca)acity. >! t$e restriction on )ain ca)acity is )resent"
t#o location areas must be divided !or t$em eit$er in terms o! band or
eora)$ic location" as s$o#n in Fiure 1,15 and Fiure 1,18.
Gocation area division based on eora)$ic location
>! t$e location area is divided in terms o! band" because !requent location
u)dates are resulted !rom inter,band $andover and cell reselection" you must
set related )arameters to maintain t$e mobile stations stay in t$e 1800 M'(
cells #$ere t$e tra!!ic is absorbed. >n t$is case" t$e times !or t$e mobile
station to $andover bet#een t$e t#o bands and reselect cells #ill decrease.
Mean#$ile" #$en desinin sinalin c$annels" you must !ully consider t$e load
resulted !rom location u)date.
>! t$e location is divided in terms o! eora)$ic location" t$e !requent location
u)dates resulted !rom inter,band $andover and cell reselection can be avoided.
'o#ever" you need to modi!y t$e related data o! t$e oriinal 900 M'( net#or?.
>n addition" at t$e edes o! t$e location areas" because t$e location u)dates
caused by intra,band and inter,band $andover and cell reselection is )resent
simultaneously" t$e sinalin !lo# is $ue at t$ese edes. As a result" you must
care!ully desin t$e edes o! t$e location areas.
%.9.4 +ra!!ic Guidance and Control Strateies o! Dual,Band Eet#or?
>. +ra!!ic uide o! Dual,Band Eet#or?
At early construction stae o! a dual,band net#or?" tra!!ic control concerns
$o# to use t$e ne# GSM 1800M'( net#or? to s$are t$e tra!!ic !lo# !or t$e GSM
900M'( net#or?. Accordin to t$e oriinal intension o! t$e GSM 1800M'(
net#or?" t$e tra!!ic can be uided accordin to t$e !ollo#in )rinci)les*
1. At t$e early construction stae o! a dual,band net#or?" t$e GSM 1800M'(
net#or? is mainly a))lied to absorb t$e tra!!ic o! t$e dual,band subscribers so
t$at t$e load o! t$e GSM 900M'( net#or? can be eased.
%. 9$en t$e number o! dual,band subscriber ro#s lare" eac$ band must s$are
t$e tra!!ic so t$at t$e inter,band $andover times can be reduced.
@rocess o! tra!!ic uide and control strateies.
+$e various tra!!ic control strateies can be reali(ed t$rou$ adFustin
)arameter settins as !ollo#s*
1. >n idle mode" #$en t$e mobile station is selectin cells a!ter it is s#itc$ed
on and reselectin cells #$en it is in standby state" you can set $i$er
)riorities !or t$e 1800 M'( cells by desinin t$e system )arameters" includin
CBQ" CBA" C<A" +A" and @+. >n t$is case" subscribers are more li?ely to stay in
t$e 1800 M'( cells. As a result" t$eir calls are establis$ed on t$e 1800 M'(
cells.
%. >! tra!!ic conestion is )resent in t$e service cell #$en a mobile station is
settin u) a call" t$e system a))lies directed retry !unction to assin t$e
mobile station to a +C' in t$e nei$bor cells o! t$e service cell and adFust t$e
tra!!ic allocation.
2. >n conversation state" t$e tra!!ic must be uided to t$e 1800 M'( cells in
lo#er layers and levels accordin to t$e $ierarc$y cell structure. >n addition"
you can use 'ua#ei dual,band $andover alorit$ms so t$at t$e tra!!ic load can
be allocated more )ro)erly.
>>. 'ierarc$ical Cell Structure
Accordin to t$e $ierarc$y cell structure o! t$e dual,band net#or?" a GSM
system coverin an area can be divided into !our layers" as listed in +able 1,%1.
+o enable t$e net#or? to develo) smoot$ly and !le:ibly" you can divide eac$ o!
t$e !our layers into multi)le levels" and t$en you can set multi)le )riority
classes -!or e:am)le" 13 classes. !or t$e levels in eac$ layer. +$is met$od is not
only $el)!ul !or adFustin t$e tra!!ic load in )art o! t$e areas. +$ere!ore" t$e
$ierarc$ical cell structure en$ances t$e coo)eration o! t$e current net#or?
equi)ments and meets t$e devolvement o! t$e !uture net#or?.
>n terms o! tra!!ic )riority" t$e cells in lo#er layers and levels $as $i$er
)riorities" namely" t$e cells in lo#er layers $as t$e )riority to absorb t$e
tra!!ic.
%.9.1 Dual,Band Eet#or?in ;nineerin >m)lementation
Durin net#or? construction and o)timi(ation" a dual,band net#or? is
debued and commissioned ste) by ste)" #$ic$ !acilitates debuin t$e ne#
GSM 1800 M'( net#or?s and t$e oriinal GSM 900M'( net#or?s t$at $as been
e:)anded res)ectively. A!ter eac$ sinal net#or? is )er!ectly adFusted" you
must debu eac$ base station in t$e dual,band net#or?. And you cannot sto)
t$e debuin until t$e #$ole dual,band net#or? is !inis$ed.
+$e construction o! a #$ole dual,band net#or? can be divided into t$ree
staes" namely" de)loyment )re)aration" sinal 1800 M'( net#or? debuin"
and 900=1800 M'( dual,band net#or? debuin.
>. De)loyment )er)aration
+$e coordination o! dual,band tec$noloies and net#or? )lannin must be
!inis$ed in t$is stae. +$e coordination o! dual,band net#or? tec$noloies is a
)rerequisite !or t$e coo)eration o! di!!erent carriersD net#or?s. Eet#or?
)lannin is t$e !irst ste) in net#or? construction and involves many tas?s"
includin base station address survey" c$annel number )lannin"
electromanetic bac?round test" coverae test" and so on.
+$e !ollo#ins must be em)$asi(ed in dual,band coo)eration*
+$e customers" t$e t$ird )arty -t$e desinin institute or t$e oriinal
equi)ment su))lier." and t$e ne# equi)ment su))lier must be coo)erate #it$
eac$ ot$er #ell.
>! one )arty meets a tou$ )roblem durin t$e debuin o! t$e dual,band
net#or?" t$e enineers !rom a t$ird )arty must be )resent in site and $el)
)osition t$e )roblem.
+$e 900 M'( BSC and 1800 M'( BSC must sync$roni(e t$eir cloc?s #it$ t$e
same source cloc?. Mean#$ile" t$e cloc? o! eac$ base station in t$e e:istin
GSM 900 M'( net#or? can loc? t$e cloc? o! t$e BSC" and t$e cloc? o! t$e BSC
can loc? t$e cloc? o! t$e MSC.
9$en modi!yin t$e )arameters related to dual,band $andover -suc$ as
modi!yin t$e )arameters at t$e BSC side or MSC side." you must noti!y t$at to
ot$er t#o )arties.
>! t$e some )roblems concernin t$e coo)eration o! dual,band net#or? arise" a
meetin must be orani(ed" in #$ic$ eac$ )arty discuss #it$ eac$ ot$er on
$o# to solve t$e )roblems.
Bot$ t$e desinin institute or t$e oriinal equi)ment su))lier and t$e ne#
equi)ment su))lier must )rovide t$e )roFect im)lementation )lan" cutover
)lan" and )recise cell in!ormation.
>>. Sinal 1800 M'( net#or? debuin
At t$is stae" you need not modi!y any data o! t$e oriinal GSM 900 M'(
net#or?" but it is still t$e GSM 900M'( net#or? )rovides services to
subscribers. +$e GSM 1800M'( net#or? does not absorb tra!!ic.
9$en debuin t$e GSM 1800M'( net#or?" you must adFust t$e !ollo#in
)arameter so t$at t$e e:istin subscribers can be least a!!ected.
>n t$e system messae data list" set t$e )arameter 6CBA7 to 6EA7 to )revent
eneral subscribers !rom selectin and reselectin t$e 1800 M'( net#or?.
+$eoretically" eneral subscribers can $and over to t$e 1800 M'( net#or?" but
in !act" t$e $andover relations$i) is not con!iured #it$ t$e dual,band
net#or?" so t$e eneral subscribers cannot enter t$e 1800 M'( net#or?.
A!ter t$at" you use t$e testin mobile )$one #$ic$ can access t$e net#or? by
!orce to )er!orm dialin test in eac$ cell. >! all oes normal" you can test
coverae" $andover" )o#er control" inter!erence" do#nlin? and u)lin? balance"
)o#er adFustment" t$e coverae o! t$e GSM 900M'( net#or?" and t$e coverae
o! t$e GSM 1800M'( net#or?.
+$rou$ t$ese tests" you can not only discover t$e )roblems )resent in t$e
net#or?s" but also adFust t$e c$annel number" )o#er" tilt anle" and
)arameter settin and o)timi(e t$e )arameter con!iuration !or t$e GSM
1800M'( cell. >n t$is case" t$e coverae and o)eration o! t$e sinle GSM
1800M'( net#or? can be ensured.
>>>. 900=1800 M'( dual,band net#or? debuin
A!ter !inis$in t$e sinle GSM 1800M'( net#or? debuin" you must c$ane
bac? t$e )arameter 6CBA7 to 6R;S7 and con!iure t$e data !or dual,band
$andover. +$e tests involved into t$e dual,band net#or? debuin include*
Cell reselection and location u)date
+ra!!ic load control
Continuous conversation mode
Automatic dialin and scan
Dual,band net#or? $andover
Calls and $andovers initiated on maFor streets
Calls and $andovers initiated on ede areas
Dialin tests in )oor coverae areas and indoor environment
Dialin tests in outdoor and indoor environments in ?ey areas
+$e data includes nei$bor cell relations$i)" layer and level settin" $andover
ty)e" and $andover t$res$old. >n t$is case" #$en a mobile )$one is in idle
mode" it can reselect an 1800M'( cell" t$e GSM 1800M'( net#or? can absorb
t$e tra!!ic o! dual,band subscribers" and t$e subscribers can )er!orm $andover
bet#een 1800M'( cells and 900M'( cells.
At t$e beinnin" you can control t$e GSM 1800M'( net#or? to absorb only a
small )art o! t$e tra!!ic o! subscribers t$rou$ adFustin t$e settin o! C<A and
$andover t$res$old. 9$en ood cell reselection and dual,band $andover are
ensured" you can ta?e measures to enable t$e GSM 1800M'( net#or? to absorb
more tra!!ic" #it$ t$e )rerequisites t$at no conestion is )resent amon cells
and t$e net#or? quality is ensured.
At t$is stae" t$e !ollo#in )arameters must be con!iured*
+$e )arameters related to cell selection and reselection" includin CBA" CBQ"
ACCM>E" C<'" and C<A.
+$e )arameters related to nei$bor cell relations$i)" layer and level settin"
and $andover.
+$e con!iuration o! t$e )revious )arameters must be based on t$e
)rerequisite t$at t$e coo)eration o! t$e GSM 1800M'( cells and GSM 900M'(
cells is normal.
A!ter t$e GSM 900M'( and 1800M'( dual,band net#or? is enabled" you must do
t$e !ollo#ins*
1. Find out t$e )roblems )resent in t$e net#or? t$rou$ multi)le means" suc$
as drive test.
%. AdFust and o)timi(e t$e net#or? accordin to t$e )roblems so t$at t$e dual,
band net#or? can run stably.
2. C$ec? i! t$e dual,band net#or? runs stably" analy(e all t$e tra!!ic statistic
data" and c$ec? t$e net#or? o)eration inde:es.
4. Ma?e sure t$e )roblems and ta?e e!!ective measures accordin to t$e
analysis o! t$e drive test and tra!!ic statistics.
1. AdFust t$e related )arameters and retest t$e net#or? till t$e net#or?
inde:es meet t$e desin requirements.
+$us" a dual,band net#or? is constructed and o)timi(ed accordin to t$e t$ree
staes as introduced in t$is section.
%.10 Desin o! >ndoor Coverae System
%.10.1 C$aracteristics o! >ndoor coverae
9it$ t$e ra)id develo)ment o! economy" $otels" commercial centers" lare,
scale !lats" underround rail#ays" and underround )ar?in areas are arisin by
batc$. As a result" mobile stations are more !requently used in indoor
environment. +$us" t$ey require better indoor mobile communication services.
Generally" t$e !ollo#in )roblems are )resent in indoor mobile communication
systems*
From t$e )ers)ective o! coverae" t$e com)le: indoor structure and t$e
s$ieldin and absorbin e!!ect o! t$e buildins cause reat radio #ave
transmission loss. As a result" t$e sinals in some areas may be #ea?"
es)ecially t$e sinals in t$e !irst and second !loors in t$e underround are quite
#ea?" or even t$ere are dead (ones. >n t$is case" mobile stations cannot
necessarily access t$e net#or?" t$ere is no )ain res)onse" or subscribers are
not in service areas.
From t$e )ers)ective o! net#or? quality" t$e !actors inter!erin radio
!requencies are )robably )resent in u))er !loors o! $i$ buildins. >n t$is case"
t$e sinals in service areas are not stable" so 6)in )on e!!ect7 may occur and
conversation quality cannot be ensured.
From t$e )ers)ective o! net#or? ca)acity" i! mobile stations are !requently
used in buildins" suc$ as lare,scale s$o))in centers" con!erence $alls" some
areas in t$e net#or? cannot meet t$e requirements o! subscribers. >n t$is case"
conestion may occur on radio c$annels.
>! t$e indoor coverae is reali(ed by a re)eater" an outdoor $i$,)o#er base
station" or a reat,$ei$t outdoor antenna" t$e !ollo#in )roblems may arise*
+$e )enetration loss is reat" so t$e indoor coverae is not satis!yin. >n t$is
case" a lare number o! dead (ones are )resent" so subscribers cannot ?ee)
conversation.
>! a re)eater is ado)ted" t$e level o! oriinal sinals must be $i$. >n addition"
t$e cross,modulation and intra,!requency inter!erence is reat" so t$e
conversation quality is #ea? and call dro) ratio is $i$.
+$e net#or? ca)acity is limited and t$e call connected ratio is lo#.
+$e !requency )lannin is $ard to be )er!ormed !or t$e net#or? and t$e
net#or? ca)acity is $ard to be e:)anded.
+$e 6detac$ed island e!!ect7 is reat.
+$e value,added services are restricted !or rou) subscribers due to net#or?
quality and ca)acity.
+o en$ance t$e rade o! service" #e must im)rove indoor coverae
immediately. 9$en desinin an indoor coverae system" #e must ma?e t$e
!ollo#in considerations*
A ne# indoor coverae system cannot a!!ect t$e e:istin net#or?.
;nou$ ca)acity o! an indoor system must be ensured.
An indoor system must su))ort ne# services and !unctions.
+$e c$a)ter analy(es t$e desin o! indoor coverae system !rom t$e !ollo#in
as)ects*
>ndoor Antenna System Desin
Ca)acity Analysis and Desin
Frequency @lannin
+ra!!ic Control
%.10.% >ndoor Antenna System Desin
>. <F desin
-1. Gin? budet
>n an indoor coverae system" t$e lin? budet !ormula is as !ollo#s*
'ere"
@ant N antenna in)ut inter!ace )o#er
<Fmar N <alei$ !adin marin
>Fmar N access marin -de)ends on environment.
GEFmar N desin marin -enerally" it is 1 dB.
BG N body loss -900M'(* 1 dB8 1800=1900M'(* 2 dB.
MSsens N mobile station sensitivity
G)at$ N )at$ loss
'ere" G)at$ N %0lod -m. J 20lo! -M'(. , %8 dB J b. 9$en t$ere no barrier
loss" G) N %0lod -m. J 20lo! -M'(. , %8 dB. +$e 6b7 indicates t$e loss caused
by ot$er bariers.
Because t$e )enetration in cylindrical tunnels is reat" lea?y cables are a))lied
in cylindrical tunnels.
9$en )er!ormin lin? budet" you must consider t$e !ollo#ins*
>n an indoor multi,antenna system" t$e lin? budet !or test )oints must be in
accordance #it$ t$e lin? #it$ t$e minimum loss.
/nder t$e same convere area" t$e ;><@ at eac$ antenna inter!ace must be
consistent" and t$e error must be controlled #it$in 10 dB.
+$e u)lin? sinal must be desined to a $i$ value" so antenna diversity is
unnecessary.
+o reduce u)lin? inter!erence" you must )ro)erly set t$e ma:imum transit
)o#er o! t$e mobile station and enable t$e )o#er control !unction o! t$e
mobile station.
A certain marin must be leaved !or error correction and !uture system
e:)ansion.
+$e estimation and desin !or inter!erence marin vary #it$ t$e distance !rom
t$e outer #all. +$e smaller t$e distance" t$e larer t$e inter!erence marin is
desined.
-%. Service quality desin -inter!erence deree.
+$e actual inter!erence level c$anes #it$ net#or? layout and !requency re,
)lannin" and it can be tested accordin to actual situations.
-2. Service quality desin -inter!erence marin desin.
+$e reater t$e inter!erence in an area" t$e reater t$e inter!erence marin
->Fmar. is desined" and t$e $i$er t$e level t$e mobile station needs to
receive.
9$en a dual,band system is ado)ted in t$e indoor environment" t$e inde:es o!
mobile station receivin level are desined accordin to t$e 1800 M'( system
standard.
>>. Antenna system desin
9$en desinin an indoor distribution system" you must !irst survey t$e
buildin ty)e" structure" inter!erence environment" customers" and t$en
analy(e t$e )at$ loss. Finally" decide t$e antenna ty)e" number" and
installation location accordin to t$e requirements o! an area.
+$is section introduces t$e antenna desin uidelines in some ty)ical cases.
-1. Sinle cell
>! t$e indoor coverae is reali(ed by a sinal cell" eac$ antenna must be
desined to ensure t$at sinals are evenly distributed in t$e coverae area.
Generally" it is recommended to install t$e antenna in a (i(a #ay.
-%. Multi,cells
>! t$e indoor coverae is reali(ed by multi)le cells" a certain distance must be
leaved bet#een intra,!requency reuse cells. ;ac$ antenna must also be
desined to ensure t$at sinals are evenly distributed in t$e coverae area o!
eac$ cell. >! t$e !requencies are reused !requently" it is recommended to
install t$e antennas on di!!erent layers at t$e same )osition o! t$e layer.-2.
Closed buildin
A closed buildin $as t$e c$aracteristics" suc$ as t$ic? outer #all" reat sinal
attenuation" and little lea?ae. >n addition" it is little a!!ected by outdoor
intra,!requency cells. +$ere!ore" t$e !requency bet#een !loors is easily to be
)lanned. For t$e antenna desin uideline in a closed environment.
-4. 'al!,o)en environment
For a $al!,o)en buildin" t$e outer #all is made o! lasses" so t$e sinal
attenuation is small. 9it$in t$e buildin are t$e o)en con!erence $alls" #$ic$
are reatly a!!ected by outdoor intra,!requency cells" so you must )lan
dedicated !requencies or ado)t t$e multi,antenna system #it$ lo# out)ut
)o#er to limit t$e edes o! t$e indoor cells #it$in t$e buildin.
-1. Frame,structure buildin
For a !rame,structure buildin" t$e number o! internal #alls is lare and t$ey
are t$ic?. +$ere!ore" i! t$e antenna is installed at t$e corridors" t$e antenna
out)ut )o#er must be $i$ so t$at ood coverae can be ensured. >n t$is case"
sinals #ill lea? at t$e #indo#s near t$e corridor" so you must )lan dedicated
!requencies !or t$e buildin. +$e distance o! t$e intra,!requency cells bet#een
!loors is larer t$an t$at in ot$er environments. For t$e antenna desin
uideline in !rame,structure buildin.
-3. A!!ice buildin
+$e indoor environment o! o!!ice buildins requires $i$ rade o! services" so
its coverae is reali(ed by several directional and omni antennas. Rou can
control t$e coverae area easily t$rou$ )ro)erly desinin t$e e!!ective
radiation )o#er in t$e cell. For desin uideline" see -5. @ar?in area
@ar?in area $as no s)ecial requirement on ca)acity and mobile station
receivin level -,90 dBm.. For a )ar?in area" t$e elevator" escalator" entrance
and e:it are ?ey coverae areas.
-8. Su)ermar?et
Su)ermar?ets $ave certain requirements on coverae and ca)acity. +$e
antennas can be desined accordin to actual structure o! t$e buildins.
>>>. Survey
+$e antenna desin and installation is !inally decided accordin to t$e survey"
#$ic$ includes t$e !ollo#in as)ects*
Detailed coverae area and sinal quality and convere requirements
Distribution o! t$e sinals in coverae areas
Com)osition o! buildins in coverae areas
Sinal access location and mode
>nstallation )osition
Accordin to t$e survey" you must out)ut t$e !inal to)oloical structure
diaram" antenna cablin sc$eme" and list o! materials. Generally" t$e omni
antenna is installed at t$e ceilin center. +$e small directional antenna is $un
on t$e inner side o! t$e outer #all" #it$ t$e radiation directed to indoor )art.
>n t$is case" t$e e!!ect o! t$e antenna aainst t$e outdoor system can be
reduced to t$e minimum" so t$e C=> requirement o! t$e outdoor system can be
met.
>! )ossible" you can test t$e coverae and adFust t$e antenna desin accordin
to t$e test result" or re,)lan t$e !requency to ensure t$e voice quality.
Generally" i! t$e radiation )o#er at t$e antenna inter!ace is 10 dBm" t$e % dBi
small indoor omni antenna is used. >n t$is case" i! t$e #alls are densely
distributed in t$e areas #it$in 20 meters !rom t$e antenna" t$e coverae level
can reac$ ,50 dBm.
%.10.2 Ca)acity Analysis and Desin
Be!ore analy(in t$e ca)acity" you must de!ine t$e ty)e o! t$e indoor service
area.
De!inition o! indoor service area ty)e
>ndoor service area ty)e
C$aracteristic
;:am)le
@ublic service area
+$e tra!!ic is $ard to be )redicted.
+$e )o)ulation number varies #it$ day and ni$t.
+$e ca)acity c$aracteristics" suc$ as uneven distribution and burstin must be
considered.
+$e rade o! service and t$e tra!!ic o! eac$ subscriber are similar to t$at !or
outdoor cells.
Air)ort" s$o))in center" and )lay round.
Commercial service area
+$e e:isted !i:ed net#or?s are !requently used.
+$e tra!!ic is relatively !i:ed and easy to be calculated.
'i$ service quality is required.
Generally" t$e rade o! service -GoS. is 1L" t$e tra!!ic o! eac$ subscriber can
reac$ 0.1 ;rl.
A!!ice buildin and commercial $otels o! $i$ ran?s.
+$ere are t#o cell orani(ation modes o! distributed antenna system" namely"
sinle cell and multi)le vertical s)lit cells. +$e sinle cell is a))lied to t$e
indoor environment #$ic$ requires smell coverae area. +$e multi)le vertical
s)lit cells are a))lied to t$e indoor environment #it$ dense tra!!ic.
Gi?e#ise" a sinle cell #ill s)lit #$en t$e ca)acity does not meet t$e
requirement" #it$ vertical s)littin t$e s)littin mode. Generally" a cell #ill
vertically s)lit into at least t$ree cells so t$at !requency reuse can be ensured.
Four layers must be )resent bet#een t#o intra,!requency cells . +o avoid
inter!erence bet#een !requencies" you must ta?e measures to )revent a cell
!rom $ori(ontally s)littin.
%.10.4 Frequency @lannin
>! t$e dedicated !requency is ado)ted in indoors" t$e !requency )lannin is
relatively sim)le. Generally" t$e !requency reuse mode in business service
areas is almost t$e same as t$at in )ubic service areas. >! t$e !requency
resource is adequate" you must try best to use dedicated band !or indoor
coverae. >! not" you can searc$ t$e available c$annel numbers #it$ relatively
small inter!erence t$rou$ scannin t$e c$annel numbers. >! t$e !requency
resources o! t$e 900 M'( cannot meet requirements" you can introduce t$e
1800 M'( !requency8 namely" use a dual,band system.
>! you steal !requency resource !or indoor system due to no available dedicated
!requency" you must )ay attention to t$e !ollo#ins*
Do not select t$e !requencies o! t$e nei$bor cells.
;nsure t$at t$e BCC' !requencies are not inter!ered.
+$e inter!erence on t$e +C' !requencies can be reduced #it$ t$e $el) o! radio
!requency $o))in.
Searc$ t$e available u)lin? !requencies t$rou$ usin B+S equi)ments to
scannin t$e u)lin? c$annel numbers.
Searc$ t$e available do#nlin? !requencies t$rou$ usin drive test equi)ment
to scannin t$e do#nlin? c$annel numbers.
>! t$e $ierarc$ical cell structure is not used" t$e cell #it$ t$e stronest sinal
level is t$e service cell" and t$e inter!erence !rom nei$bor !requencies can be
nelected.
>! t$e $ierarc$ical cell structure is used" t$e cell #it$ t$e stronest sinal level
cannot necessarily be t$e service cell" so you must ta?e measures to reduce t$e
inter!erence !rom nei$bor !requencies.
Because t$e environment is urban areas is quite com)licated" es)ecially t$e
e!!ect o! t$e antenna bac? lobe is )resent" t$e service areas !or $i$ buildins
are reatly inter!ered" so you must care!ully )lan t$e !requencies !or t$e
indoor coverae o! $i$ buildins. Generally" !or t$e lo#er !loors" you can )lan
t$e !requencies accordin to eneral met$od. For t$e $i$er !loors #$ere t$e
inter!erence is stron" you can use dedicated c$annel numbers. 'o#ever" t$e
!inal !requency )lannin must be based on )ractical tests.
%.10.1 +ra!!ic Control
+$e indoor coverae system !or $i$ buildins can be ta?en as a system
inde)endent o! outdoor systems i! t$e coverae o! t$e indoor system is ood.
+$eoretically" you can only consider t$e cell selection and reselection"
$andover relations$i)" and t$e com)act on outdoor net#or?s at t$e entrances
and e:its o! t$e buildin.
'o#ever" t$e actual conditions are quite com)licated. For e:am)le" t$e sinals
outside o! t$e buildin may be stron. >n t$is case" i! a mobile station is
)o#ered o!!" it may cam) on an outside cell. +$ere!ore" #$en o)timi(in t$e
net#or?" you must set t$e one,#ay adFacent cell and t#o,#ay adFacent cell
accordin to actual conditions and set t$e )arameters" suc$ as C<A and +A to a
)ro)er value accordin to t$e reularity o! cell selection and reselection. >n
addition" you can set t$e indoor cells to a $i$ )riority so as to reserve more
tra!!ic. And t$e inter,layer $andover t$res$old and $ysteresis are de!ined and
adFusted accordin to actual conditions.
%.11 +unnel Coverae
%.11.1 C$aracteristic o! +unnel Coverae
At )resent" most o! t$e tunnels are dead (ones" so you must ma?e out s)ecial
solutions !or tunnel coverae. +$e tunnel ty)es include rail#ay tunnel"
$i$road tunnel" and underround rail#ay tunnel. ;ac$ tunnel $as its
c$aracteristics" and t$ey are s)eci!ied as !ollo#s.
For t$e $i$road tunnel" it is #ide. +$e coverae in t$e $i$road tunnels is
relatively stable. 9$en t$ere are ve$icles )assin by" you can select t$e
antennas #it$ a larer si(e to obtain a $i$er ain" so t$e coverae distance is
larer.
For t$e rail#ay tunnel" it is narro#" es)ecially #$en t$ere is a train )assin by8
only a little room is le!t in t$e tunnel" so t$e radio )ro)aation is reatly
a!!ected. Moreover" t$e train $as reat e!!ect on radio sinals. Since t$e
antenna installation room is quite limited" t$e antenna si(e and ain are
reatly restricted. >n addition" because eneral cars cannot be driven to suc$
tunnels" t$e tunnel coverae is $ard to be tested. +$ere!ore" t$e )lannin !or
$i$road coverae is di!!erent !rom t$at o! t$e rail#ay coverae.
+$e lent$ o! tunnels ranes !rom several $undred meters to several
?ilometers. For s$ort tunnels" you can ado)t !le:ible and economical means to
reali(e t$e coverae. For e:am)le" you can install a eneral antenna near one
end o! t$e tunnel" #it$ t$e radiation directed to t$e inside. For lon tunnels"
$o#ever" you must ado)t ot$er means. Actually" t$e coverae solution varies
#it$ tunnels" so it is desined accordin to actual conditions.
Cross section o! t$e sinle,trac? rail#ay tunnel and multi,trac? rail#ay tunnel*
+$e smaller t$e area o! t$e cross section" t$e reater t$e loss #$en a train
)asses t$rou$ t$e tunnel. +$e related calculation and analysis are based on
t$e multi,trac? rail#ay tunnels and $i$road tunnels. For t$e calculation and
analysis !or sinle,trac? tunnels" t$e )rotection marin can be 1 dB reater
t$an t$at o! multi,trac? rail#ay tunnels.
Be!ore )lannin tunnel coverae" you must )re)are !or t$e !ollo#in data*
Gent$ o! t$e tunnel
9idt$ o! t$e tunnel
Eumber o! tunnel $oles -1 or %.
Eeeded coverae )robability -10L" 90L" 98L or 99L.
Structure o! t$e tunnel -it is constructed #it$ metals or concretes.
Eumber o! needed carriers -1020.
Minimum receivin level in t$e tunnel -enerally" it ranes !rom ,81 dBm to
,10% dBm.
Distance bet#een tunnel $oles
9$et$er AC=DC is available
9$et$er t$e $ole can be )unc$ed in t$e tunnel #all
Sinal level at t$e tunnel entrance
;:isted sinal level in t$e tunnel
%.11.% +unnel Coverae Solution
>. Gin? budet
>ndoor radio lin? loss is mainly decided by )at$ loss medium value and s$ado#
!adin. A tunnel can be ta?en as a tube. +$e sinals are transmitted t$rou$
t$e re!lection o! #alls and strai$t transmission" #it$ strai$t transmission t$e
maFor !orm. >+/,< suests an indoor )ro)aation model on )ae 1%28" #$ic$ is
also e!!ective !or tunnel coverae. +$e !ormula is as !ollo#s*
G)at$ N %0 l ! J 20 l d J G! -n. , %8 dB
'ere"
6!7 indicates !requency -M'(.
6d7 indicates distance -m.
6G!7 indicates )enetration loss !actors bet#een !loors -dB.
6n7 indicates t$e number o! !loors lyin bet#een t$e mobile station and
antenna.
+$e G! -n. can be nelected in tunnel coverae" so t$e !ollo#in equation can
be a))lied in t$e calculation o! t$e radio )ro)aation in tunnels. +$at is*
G)at$ N %0 l! J 20 l d , %8 dB
>>. GSM sinal source selection
A GSM sinal source and a set o! distributed antenna system are a must !or
tunnel coverae. For tunnel coverae" t$e GSM sinal source is selected
accordin to t$e radio coverae" transmission" tra!!ic" and t$e e:istin net#or?
equi)ments near t$e tunnel. A macro cell base station" a micro cell base
station" or a re)eater can #or? as a GSM sinal source !or t$e tunnel coverae.
For t$e coverae o! rail#ay tunnels and $i$road tunnels" t$e indoor macro cell
base station is seldom used as sinal source" but it can be used !or an
underround rail#ay #$ic$ requires t$e coverae o! )lat!orms and entrances.
>n t$is case" t$e ca)acity o! t$e sinal source must be reat. >n most cases"
$o#ever" t$e tunnel coverae is reali(ed by micro cell sinals.
For t$e areas to be covered" i! t$e nearby net#or? ca)acity is adequate" t$e
ca)acity e:)ansion is unnecessary. And i! t$ere are ood GSM sinals available"
namely" t$e donor sinal level meets t$e requirements o! a re)eater -!or
e:am)le" ,50 dBm.8 a re)eater can #or? as t$e sinal source !or t$e tunnel
coverae. 9it$ t$e increase o! tra!!ic" $o#ever" you must use GSM base
stations to re)lace t$e re)eaters.
Adequate isolation must le!t bet#een donor antenna and retransmission
antenna" t$ou$ it #ill cause di!!iculty in antenna installation. Generally" t$e
lo,)eriodical antenna #it$ reat !ront,to,bac? ratio is used as t$e
retransmission antenna.
+$e eneral antenna -#ireless re)eater." coa:ial cable" and o)tical !iber
-o)tical re)eater. can connect a re)eater to a donor cell.
For tunnel coverae" t$e installation s)ace and au:iliary equi)ments are quite
limited" so micro cell base stations and re)eaters instead o! macro cell base
stations are o!ten a))lied in tunnel coverae.
>n mountain areas" re)eaters are more li?ely used because stron sinal level
o!ten e:ists at t$e mountain to)s near t$e tunnel. >n t$is case" t$e antenna
isolation requirement can be easily met. >! t$e sinal level o! t$e e:isted
net#or? near t$e tunnel is not stron enou$" you can use a micro cell !or t$e
tunnel coverae.
>>>. Antenna !eeder system selection
A!ter decidin t$e GSM sinal source" you must con!iure t$e antenna !eeder
system !or t$e tunnel coverae accordin to actual conditions. +$ree ty)es o!
con!iuration are available" namely" coa:ial !eeder )assive distributed
antenna" o)tical !iber !eeder active distributed antenna" and lea?y cable.
'ereunder introduces t$e tunnel coverae based on coa:ial !eeder )assive
distributed antenna and lea?y cable.
%.11.2 +unnel Coverae Based on Coa:ial distributed antenna system
>n a coa:ial distributed antenna system" t$e !ollo#in <F com)onents are used*
Feeder -2=8Y" 1=%Y" or 5=8Y. and Fum)er
@o#er s)litter
@o#er s)litter
Antenna
+$is section introduces t$ree tunnel coverae solutions based on t$e coa:ial
distributed antenna system.
>. Solution 1
+unnel coverae solution based on t$e bi,directional )assive distributed
antenna system.
+unnel coverae solution based on bi,directional )assive distributed antenna
system
Accordin to t$is solution" i! t$e needed minimum sinal level is ,81dBm -t$e
location )robability is 10L." you must add a marin o! 8 dB i! t$e #ant to
en$ance t$e location )robability to 90L.
>! t$e ain o! t$e bi,directional antenna is 1 dBi" t$e loss o! t$e equal
)robability )o#er s)litter and t$e Fum)er is % dB" and t$e !eeder #it$ t$e
s)eci!ication o! 5=8Y is used" t$e )at$ loss in 100 meters is 4 dB and t$e out)ut
)o#er o! t$e equi)ment is 29 dBm.
Su))ose t$at t$e level o! t$e sinals transmitted by t$e !irst bi,directional
antenna is ,81 dBm at t$e tunnel entrance" you can calculate t$e distance
bet#een t$e antenna and t$e tunnel entrance usin t$e !ollo#in equation*
@out, G)at$ -d. 0 Gcable -d. 0 GFum)er J Gant N ,81dBm J 8dB90LPloc.@rob
'ere"
@out indicates t$e out)ut )o#er -29dBm..
G)at$ -d. indicates t$e )at$ loss !rom t$e !irst bi,directional antenna to t$e
tunnel entrance.
Gcable -d. indicates t$e cable loss.
GFum)er indicates t$e Fum)er loss -% C % dB..
Gant indicates t$e antenna ain -1 dBi..
>! introducin t$e )revious data to t$e equation" you can obtain t$e sum o! t$e
G)at$ -d. and Gcable -d." t$at is" 115 dB.
For t$e relations$i) bet#een distance 6d7 and G)at$ -d. and Gcable -d." see
Fiure 1,24" in #$ic$ t$e curve indicates G)at$ -d. and t$e slant line indicates
Gcable -d..
1 Rou can obtain t$at d N 201m t$rou$ estimation.
>! a )o#er s)litter is ado)ted !or t$e !irst antenna" a loss o! 2dB must be
added. >n t$is case" t$e sum o! G)at$ -d. and Gcable -d. is 114 dB.
% Rou can also obtain t$at d N %31m t$rou$ estimation.
For rail#ay tunnels" train !illin #ill a!!ect sinal )ro)aation" so a )rotection
marin o! 1dB must be considered #$en t$e antenna is installed in t$e tunnel.
>n t$is case" d N %40m. +$at is" i! a bi,directional antenna is installed in t$e
tunnel" it can coverae a distance o! 480m.
>! a )o#er s)litter is ado)ted !or t$e second antenna" t$e coverae distance
bet#een t$e !irst antenna and t$e second antenna #ill be s$ortened unless an
am)li!ier is used.
+$e !ollo#ins analy(e t$e coverae #$en no am)li!ier is ado)ted !or t$e
second antenna.
+$e total )o#er out)ut by t$e !irst )o#er s)litter -it is installed at t$e !irst
antenna. @out1 is e:)ressed as !ollo#s*
@out1 N @out 0 Gcable -d. , GFum)er , Gs)litter N 29dBm 0Gcable -%31m. , %dB ,
2dBN %2.13 dBm. -+$e cable loss in %31m is about 10.44 dB" Fum)er loss is % dB"
and t$e )o#er s)litter intersection loss is 2dB..
Su))ose t$e overla))in level bet#een t$e t#o antennas is ,81 dBm" t$e
distance bet#een t$e second antenna and t$e !irst antenna is* d% N d J :.
'ere" 6d7 indicates t$e coverae distance o! t$e !irst antenna -%31m." and 6:7
indicates t$e coverae distance o! t$e second antenna in t$e sinle direction.
Accordin to t$e )revious analysis" t$e !ollo#in t#o equations can be
obtained*
@out1 0 Gcable -%31m. 0 Gcable -:. 0 GFum)er J Gant 0 G)at$ -:. N , 81dBm J
8dB90LPloc.@rob
G)at$ -:. J Gcable -:. N 108.13dB
@lus t$e t#o equations" you can obtain t$e value o! :" t$at is" 100m. +$is
means t$at #$en no am)li!ier is ado)ted" t#o antennas can coverae a tunnel
distance o! 5%%m" namely" %M-%31 J 100. m N 5%%m.
>! you ado)t cascaded antennas" t$e transmit )o#er is relative lo# due to t$e
coa:ial cable loss. >n t$is case" you can use t$e am)li!ier to am)li!y t$e )o#er.
>>. Solution %
>! a tunnel is not lon" you can ado)t a sim)ler coverae mode.
+unnel coverae solution based on a sinle antenna
Accordin to t$is solution" a directional antenna is installed at t$e tunnel
entrance" #it$ t$e radiation directed to t$e inside. +$e !ollo#in analy(e t$is
coverae solution.
>n t$is solution" @out N 29 dBm -su))ose t$at t$e out)ut )o#er o! t$e GSM
sinal source is 89..
>! t$e G)at$ -d. indicates )ro)aation loss" t$e sum o! Gcable -d. and GFum)er
is 1dB" t$e antenna ain Gant is 8 dBi" and t$e needed received level is
,55dBm" t$e G)at$ -d. is e:)ressed as !ollo#s*
G)at$ -d. N 29dBm , 1dB J 8dBi 0 -,55dBm. N 119 dB
Accordin to t$e equation G)at$ -d. N %0 l10! J 20 l10d , %8 dB" t$e value o!
6d7 can be obtained" t$at is" 818m.
+$e )revious analysis is a))licable to $i$road tunnels. For rail#ay tunnels" you
can consider a marin o! 10 dB due to t$e e!!ect o! train !illin" but t$e
coverae distance o! t$e antenna in rail#ay tunnels is calculated t$e same as
t$at in $i$road tunnels. Accordin to t$e calculation" d N 298m.
%.11.4 +unnel Coverae Based on Gea?y Cable System
>! ado)tin lea?y cables to reali(e t$e tunnel coverae" you must !ind t$e
s)eci!ications o! t$e lea?y cables and com)lete t$e lea?y cable desin
accordin to t$e !ollo#in ste)s*
1. Decide coverae !actor
%. Calculate t$e ain o! t$e bi,directional am)li!ier
2. ;stimate t$e lent$ o! t$e lea?y cable bet#een t$e !eeder source and t$e
!irst am)li!ier
4. ;stimate t$e lent$ o! t$e lea?y cable bet#een t$e am)li!iers
1. Decide t$e number o! needed am)li!iers
+$e !ollo#ins describe t$ese ste)s in details.
>. Decide coverae !actor
+$e !ollo#in in!ormation is needed !or decidin t$e coverae !actor*
Cou)ler loss
Eumber o! carriers
Coverae )robability
Coverae !actor indicates t$e loss in t$e areas % meters beyond t$e lea?y cable
-alon t$e vertical direction.. +$is loss includes t$e cou)ler loss o! t$e lea?y
cable and )rotection marin required by t$e coverae )robability. >! 90L o!
coverae )robability is required" you must add 8dB to t$e medium level. Some
lea?y cables s)eci!y t$e relations$i) bet#een t$e coverae )robability and
cou)ler loss.
+$e coverae !actor is determined by t$e )arameters" suc$ as cou)ler loss" <F
carrier number" coverae )robability" and tunnel ty)e. For t$e decision o!
coverae !actor in concreter tunnels. For t$e decision o! coverae !actor in
metal tunnels. 9$en decidin t$e coverae !actor" you can !i: a )oint in t$e
ra)$ and mar? a $ori(ontal line t$rou$ t$is )oint" and t$is line intersects
required coverae )robability. +$is intersection )oint is t$e coverae !actor.
Coverae !actor in metal tunnels*
For e:am)le" i! t$e lea?y cable #it$ a cou)ler loss o! 51 -900 M'(. is used" t$e
<F carrier number is 18" and t$e coverae )robability is 90" t$e coverae !actor
in a concrete tunnel is ,55
>>. Decide cable lent$ bet#een GSM sinal source and t$e !irst am)li!ier
Be!ore decidin cable lent$ bet#een GSM source and t$e !irst am)li!ier" you
must obtain t$e !ollo#in in!ormation*
+ransmit )o#er o! t$e sinal source -dBm.
Xum)er loss* 1 dB
Connector loss* 1 dB
Gea?y cable loss* % dB
+ransmit )o#er at t$e !eeder source -dBm.
9$en calculatin t$e )o#er at a )oint o! t$e !eeder" you must subtract t$e
!eeder )ro)aation loss !rom t$e GSM sinal source. >! a #ireless re)eater #it$
an out)ut )o#er o! 18 dBm -18 carriers. is used as t$e GSM sinal source" and
t$e attenuation !rom t$e Fum)er to !eeder" and !rom t$e !eeder to t$e lea?y
cable is 5 dB -+$at is" t$e )o#er !rom t$e re)eater is transmitted !rom a
Fum)er to a !eeder" and t$en !rom t$e Fum)er to a lea?y cable" so !our
connectors are needed. Generally" t$e attenuation is % dB !or eac$ Fum)er" 1
dB !or eac$ !eeder" and 0.1 dB !or eac$ connector" so t$e total attenuation is 5
dB.." t$e transmit )o#er at t$is )oint is 11 dB. For t$e connection o! lea?y
cable.
Connection sc$eme o! lea?y cable*
Su))ose t$e needed sinal level in a tunnel is ,81 dBm" t$e sinal level at t$e
!irst am)li!ier must be equal to or reater t$an ,81 dBm. +$e cou)ler loss and
lonitudinal )ro)aation loss o! t$e lea?y cable are )resent bet#een t$e sinal
!eeder )oint and t$e !irst am)li!ier. +$ey are calculated accordin to t$e
!ollo#in equation*
GossGon N 11dBm 0 -,81dBm. J Gosscou). 'ere" Gosscou) indicates t$e
coverae !actor" and it is ,55dB #$en 90L coverae is ensured. +$ere!ore" t$e
GossGon is 19 dB -t$at is" 11dBm J 81dBm ,55dB N 19dB..
+$e cable lent$ bet#een t$e sinal !eeder source and t$e !irst am)li!ier can
be obtained accordin to Fiure 1,29 and Fiure 1,40. For e:am)le" su))ose
t$at t$e attenuation is 4.2dB=100 !or t$e lea?y cable" you can mar? a )lumb
line at t$e )oint indicatin 4.2dB. +$is )lumb line #ill intersect t$e curve
indicatin 19 dB at a )oint" and t$en you mar? a $ori(ontal line startin !rom
t$is )oint. +$e $ori(ontal line #ill intersect t$e ri$t vertical a:is at a )oint.
And t$is )oint s$o#s t$e cable lent$. Accordin to t$is e:am)le" t$e distance
bet#een t$e sinal source and t$e !irst am)li!ier is 440m -t$at is" 19=4.2 N
440m..
Cable lent$ bet#een am)li!iers in concrete cables*
Accordin to t$e )revious !iures" t$e le!t vertical a:is indicates 6<equired
<AD>AM@ Gain7" #$ic$ can be re)laced by t$e radial loss o! t$e lea?y cable"
but it ma?es no di!!erence.
>>>. Eeeded am)li!ier ain
Be!ore calculatin t$e ma:imum am)li!ier ain" you must collect t$e !ollo#in
in!ormation*
+$e minimum acce)table sinal level -dBm.
Coverae !actor -dB.
+$e ma:imum out)ut loss allo#ed by a sinle carrier -dBm.
>! t$e am)li!ier is not added" t$e sinal level out)ut by t$e lea?y cable !or t$e
lonest transmission distance is equal to t$e di!!erence o! t$e minimum
acce)table sinal level and t$e coverae !actor.
+$e sinal level at t$e lea?y cable beyond t$e lonest transmission distance
may be lo#er t$e minimum acce)table level" so an am)li!ier must be added to
am)li!y t$e sinals to t$e ma:imum out)ut )o#er allo#ed by a sinle carrier.
+$e am)li!ication o! t$is )o#er is related to t$e s)eci!ications o! t$e am)li!ier
and t$e number o! carriers. >! t$e ma:imum out)ut )o#er allo#ed by a sinle
carrier is ?no#n" t$e am)li!ier ain can be calculated as !ollo#s*
Eeeded am)li!ier ain N t$e ma:imum out)ut )o#er allo#ed by a sinle carrier
-it de)ends on t$e number o! carriers. 0 -t$e minimum acce)table sinal level 0
coverae !actor.
Alon t$e lea?y cable" t$e ma:imum out)ut )o#er allo#ed by eac$ carrier o! a
bi,directional am)li!ier is related to t$e number o! carriers t$at $ave been
am)li!ied. +$is is considered mainly !or t$e intermodulation inter!erence is
)resent" because t$e intermodulation inter!erence #ill increase #it$ t$e total
number o! carriers t$at $ave been am)li!ied.
<elations$i) bet#een t$e ma:imum out)ut )o#er allo#ed by a sinle carrier
and t$e number o! carriers t$at $ave been am)li!ied*
Eeeded am)li!ier ain N t$e minimum acce)table sinal level 0 coverae !actor
J t$e ma:imum out)ut )o#er allo#ed by a sinle carrier.
Accordin to t$e )revious equation" i! t$e minimum acce)table sinal level is
,81 dBm" t$e coverae !actor is ,55" and t$e ma:imum out)ut )o#er allo#ed
by a sinle carrier is 1 dBm" t$e needed am)li!ier ain is 12 dB.
>H. Decide cable lent$ bet#een am)li!iers
Be!ore decidin t$e cable lent$ bet#een am)li!iers" you must ?no# t$e
needed am)li!ier ain and t$e cable loss -dB=100m.. Fiure 1,29 and Fiure 1,
40 $el) you decide t$e cable lent$ bet#een am)li!iers. For e:am)le" in a
concrete tunnel" i! t$e am)li!ier ain is 12 dB and t$e cable attenuation is
4.2dB=100m" t$e cable lent$ bet#een t#o am)li!iers is 200m.
H. Decide t$e number o! needed am)li!iers
Be!ore decidin t$e number o! needed am)li!iers" you must ?no# t$e !ollo#in
in!ormation*
+$e cable lent$ bet#een t$e !eeder source and t$e !irst am)li!ier
+$e cable lent$ bet#een am)li!iers
+$e tunnel lent$
>! t$e )revious in!ormation is ?no#n" t$e !ollo#in !ormula can be used to
calculate t$e number o! needed am)li!iers. +$at is*
+$e number o! am)li!iers c -t$e tunnel lent$ 0 t$e cable lent$ bet#een t$e
!eeder source and t$e !irst am)li!ier.=-t$e cable lent$ bet#een am)li!iers."
roundin u) to t$e nearest inteer.
Accordin to t$e !ormula" i! t$e tunnel lent$ is 1000m" t$e cable lent$
bet#een am)li!iers is 200m" and t$e cable lent$ bet#een t$e !eeder source
and t$e !irst am)li!ier is 4%0m" % am)li!ies are needed. +$at is" -1000 0
4%0.=200 N 1.92" so t$e nearest inteer is %.
A!ter decidin t$e number o! needed am)li!iers" you can o)timi(e t$e distance
bet#een am)li!iers. +$at is" you can obtain t$e distance bet#een t$e t#o
am)li!iers by dividin t$e remainin distance by t$e number o! needed
am)li!ier. Accordin to t$e )revious e:am)le" it is 180=% N %90m" namely" t$e
distance bet#een t$e t#o am)li!iers is %90m.
H>. <emar?s on lea?y cable installation
+$e lea?y cable must not touc$ any metal. Generally" a lea?y cable must be
installed at a s)ot 1m a#ay !rom concrete #alls and at least 10m a#ay !rom
metal #alls. >n addition" a lea?y cable must be installed near to t$e coverae
area. Rou cannot necessarily consider t$e line,o!,si$t )ro)aation" because
t$e sinals lea?in !rom t$e cable #ill !ill t$e s)ace nearby.
+$is section introduces t$e coverae solutions to tunnels in di!!erent lent$. >n
actual net#or?in" t$e !ollo#in coverae solutions may be used*
Micro base station -or re)eater. J a sinle antenna
Micro base station -or re)eater. J distributed antenna system
Micro base station -or re)eater. J lea?y cable
Be!ore decidin #$ic$ coverae solution s$ould be ado)ted" you must consider
t$e !ollo#ins*
>s t$e GSM sinal near t$e tunnel entrance stron enou$d
>s t$ere any available transmission lin? near t$e tunneld
Generally" i! t$e e:isted sinal level near t$e tunnel entrance -includin nearby
mountains. is lo#er t$an ,80 dBm" t$e micro base station is recommended. >! it
is reater t$an ,80 dBm" t$e micro base station or t$e re)eater is
recommended. >! )roblems concernin transmission are )resent" t$e re)eater is
recommended. 9$en usin t$e re)eater" you must consider t$at certain
isolation is required bet#een re)eaters.
>. Coverae solution to s$ort tunnels
Generally" t$e tunnels s$orter t$an 100m are de!ined as s$ort tunnels. 9$en
)lannin t$e coverae !or t$ese tunnels" you must consider t$e coverae areas
near t$e tunnels. >! several tunnels are close to eac$ ot$er" you can install a
base station or a re)eater bet#een t$e tunnels. >! ado)tin a micro base
station" you must ado)t t$e bi,directional antenna. >! t$e antenna ain is 1 dBi"
you s$ould install t$e antenna at t$e tunnel entrance so as to ensure coverae.
9$en desinin tunnel coverae solutions" you must !ully consider t$at !act
t$at cars and trains move at a $i$ s)eed" so $o# to ensure normal $andover
a!ter t$e cars or trains steerin into t$e tunnels is o! vital im)ortance.
>! t$e re)eater is used as t$e GSM sinal source and t$e sinals outside t$e
tunnel and t$e sinals #it$in t$e tunnel belon to t$e same cell" no $andover
)roblem #ill occur. >! t$e micro cell is used as t$e GSM sinal source and t$e
sinals outside t$e tunnels and t$e sinals #it$in t$e tunnel belon to di!!erent
cells" t$e sinals in t$e outside cell #ill dro) dramatically #$en t$e train steers
into t$e tunnel. >n t$is case" $andover !ailure may occur and call dro) #ill be
resulted in.
+o solve t$is )roblem" you can consider ado)tin t$e !ollo#in met$ods*
Ado)t t$e bi,directional antenna !or t$e tunnel coverae" because it can
)rovide enou$ overla))in area !or $andover.
;nable s)ecial $andover alorit$ms" suc$ as !ast level !all $andover alorit$m.
>n t$is case" a mobile station can $and over to anot$er cell #$en t$e sinal
level !alls !ast.
Select t$e directional antenna #it$ small !ront,to,bac? ratio.
>>. Coverae solution to middle,lent$ tunnels
+$is section introduces several ty)ical coverae solutions to rail#ay tunnels.
+$e !ollo#ins are a series o! assum)tions*
+$e 'ua#ei B+S2001C -t$e ma:imum out)ut )o#er is 89. is used as t$e GSM
sinal source.
+$e re)eater #it$ 1 am)li!ied carrier and a ma:imum out)ut )o#er o! %9 is
considered.
+$e lo#est receivin level is desined to ,81 dBm" and t$e coverae )robability
is 90L -#it$ a )rotection marin o! 8 dB..
For rail#ay tunnel coverae" because t$e train #ill a!!ect sinal transmission"
i! t$e antenna is installed at t$e tunnel entrance" t$e )rotection marin must
be increased by 10 dB. >! t$e antenna is installed in t$e tunnel" t$e )rotection
marin must be increased by 1dB.
+$e dedicated directional antenna #it$ t$e s)eci!ication o! DB551S10ESR" t$e
$ori(ontal $al! )o#er anle o! 30`" and t$e antenna ain o! 8 dBi is used at t$e
tunnel entrance.
+$e bi,directional antenna #it$ t$e s)eci!ication o! &528443 and antenna ain
o! 1 dBi is used #it$in t$e tunnel.
Accordin to t$ese assum)tions" i! a micro base station -29 dBm. is used as t$e
GSM sinal source" t$e coverae distance is 400m #$en t$e antenna #it$ a ain
o! 8 dBi is installed at t$e tunnel entrance" and t$e coverae distance is 480m
#$en t$e bi,directional antenna #it$ a ain o! 1 dBi is installed in t$e tunnel.
>! a re)eater -22 dBm. is used as t$e GSM sinal source" t$e coverae distance
is %10m #$en t$e antenna #it$ a ain o! 8 dBi is installed at t$e tunnel
entrance" and t$e coverae distance is 230m #$en t$e bi,directional antenna
#it$ a ain o! 1 dBi is installed in t$e tunnel.
+$ere!ore" !or t$e tunnels s$orter t$an 100m" you can use t$e combination o! a
micro base station and a sinle antenna -or a re)eater. !or t$e tunnel
coverae. For curve tunnels" you can install a bi,directional antaean in t$e
tunnel.
Accordin to on,site survey on t$e cross,section" t$e available antenna si(e"
and t$e tunnel lent$" you can use t$e antenna #it$ a $i$er ain to coverae
t$e tunnels a little loner t$an 100m.
>>>. Coverae solution to lon tunnels
For t$e tunnels loner t$an 100m" you need to use t$e distributed antenna
system or t$e lea?y cable !or t$e coverae. +$e !ollo#ins introduce t$e
coverae reali(ed by t$e combination o! a micro base station and a lea?y cable
-or a re)eater..
'ereunder is a series o! assum)tions*
+$e 'ua#ei B+S2001C -t$e ma:imum out)ut )o#er is 89. is used as t$e GSM
sinal source.
+$e re)eater #it$ 1 am)li!ied carrier and a ma:imum out)ut )o#er o! %9 is
considered.
+$e lo#est receivin level is desined to ,81 dBm" and t$e coverae )robability
is 90L -#it$ a )rotection marin o! 8 dB..
+$e lea?y cable #it$ t$e s)eci!ication o! SG9R,10,%% and t$e radial loss o!
1dB=100 m is used.
+$e cou)ler loss may be 55 dB #$en t$e 90L o! sinals are received.
Accordin to t$ese assum)tions" i! a micro base station -29 dBm. is used as t$e
GSM sinal source" t$e coverae distance is 800m #$en only t$e lea?y cable
but no am)li!ier is used. >! a re)eater -22 dBm. is used as t$e GSM sinal
source" t$e coverae distance is 380m #$en only t$e lea?y cable but no
am)li!ier is used. +$e coverae distance #ill be larer i! lea?y cables #it$
smaller loss are used.
For t$e coverae o! still loner tunnels" you must use am)li!iers to am)li!y
sinals. +$at is" you can use eit$er t$e distributed antenna system or t$e lea?y
cable !or t$e coverae solution. >n terms o! tec$nical inde:es and installation
s)ace" coverae solution based on lea?y cable is recommended. >n terms o!
cost" you must select a suitable coverae solution base on actual conditions.
%.1% <e)eater @lannin
%.1%.1 A))lication Bac?round
9it$ ra)id develo)ment o! mobile communication net#or?s" )eo)le $ave
$i$er requirements on service quality. +$ey $o)e to enFoy mobile services
any#$ere and anytime. As !or telecommunication carriers" t$ey cannot enable
a base station in some dead (ones due to t$e reasons suc$ as cost and
transmission conditions. >n t$is case" a re)eater can )rovide an au:iliary and
economical means to coverae t$e dead (ones.
>. <e)eater ty)es
A #ireless re)eater ado)ts a set o! donor antenna to receive t$e sinals !rom
t$e base station. A!ter am)li!yin t$e sinals" it ado)ts a set o! retransmission
antenna to !or#ard t$e sinals in anot$er direction. Generally" a #ireless
re)eater $as only one receivin )at$" so t$e diversity antenna is unnecessary.
A)tical re)eater
An o)tical re)eater transmits sinals usin o)tical !ibers" so t$e re)eater side
and base station side must $ave t$e o)tical transmission ca)ability.
C$annel band#idt$
Band#idt$ selection re)eater
A band#idt$ selection re)eater is also called #ideband re)eater" and it can
select a !requency -!or e:am)le" t$e !requency #it$ a band#idt$ o! 3M" 19M"
or %1M. and am)li!y it.
C$annel selection re)eater
A c$annel selection re)eater is also called narro# band re)eater or !requency
selection re)eater. >t am)li!ies t$e selected c$annel numbers only. >t is a
narro# band re)eater and am)li!ies a limited c$annel numbers.
Ee# style
Solar enery re)eater
A solar enery re)eater is o! t$e #ideband ty)e. >t is similar to a eneral
#ideband re)eater e:ce)t t$at its )o#er is solar enery.
@roduct ty)e
9ireless !requency selection re)eater
Currently" t$e ty)es o! t$e re)eaters listed in t$e le!t column are in
commercial use.
A)tical !requency selection re)eater
9ireless #ideband re)eater
A)tical #ideband re)eater
>>. Com)arison bet#een re)eater and micro cell
Many equi)ments and a lon )eriod are needed !or constructin a micro cell.
A re)eater is installed in a !le:ible #ay and t$e base station equi)ments and
transmission equi)ments are unnecessary.
A micro cell can e:)and t$e system ca)acity. 9$en t$e cells near a base
station are busy" a micro cell can be used to ease t$e conestion.
A re)eater can absorb tra!!ic. 9$en a cell is idle" it brins t$e tra!!ic to t$is
cell" t$us en$ancin t$e utili(ation ratio o! t$e equi)ments. A re)eater does
not e:)and t$e ca)acity !or a system.
+$e system needs to allocate c$annel numbers to a micro cell" but t$is is $ard
to be reali(ed in t$e areas #$ere t$e !requency resource is scarce.
+$e system does not need to allocate c$annel numbers to a re)eater" but it
must )revent t$e re)eater !rom inter!erin #it$ ot$er cells.
Eote*
+$e !ilter o! an intra,!requency re)eater #ill )roduce a delay o! about 1Ks.
+$eoretically" t$e ma:imum e!!ective coverae distance o! a GSM cell #ill be
smaller t$an 21?m in t$is case.
A GSM system must enable t$e dynamic )o#er control !unction" #$ic$ is
trans)arent to a re)eater. Generally" you must ado)t t$e automatic level
control tec$noloies -AGC. !or a re)eater.
^ Eote*
9$en t$e AGC tec$noloy is a))lied to a re)eater" i! a mobile station is too
near to t$e re)eater" t$e re)eater #ill reduce t$e ains !or all t$e mobile
stations #it$in its service area. >n t$is case" t$e conversation quality o! some
mobile stations #ill become )oor" or even call dro) may occur8 es)ecially t$e
mobile stations !ar a#ay !rom t$e re)eater are reatly a!!ected.
>>>. A))lication c$aracteristics
<e)eaters are mainly used to cover t$e dead (ones in vast o)en land" and t$ey
are t$e e:tension o! t$e base stations. A re)eater im)roves t$e coverae but
does add u) to t$e tra!!ic ca)acity o! a net#or?. 'o#ever" because it enlares
t$e coverae o! t$e base station" t$e total tra!!ic volume increases.
A #ireless re)eater a))lies t$e radio transmission mode" #it$ s$ort
construction )eriod and e!!ective cost. An o)tical re)eater ado)ts o)tical !iber
as transmission medium" so t$e transmission loss is small and transmission
distance is lare" but construction cost is reater t$an t$at o! t$e #ireless
re)eater.
+$e a))lication advantae o! t$e #ireless re)eater lies in lo# transmission
requirement. >! you )lant t$e o)tical !iber" t$ere is no )rice advantae aainst
t$e construction o! a micro cell base station. >n t$is case" considerin t$e
net#or? quality" you are recommended to select t$e micro cell base station.
Com)ared #it$ #ideband re)eater" a narro# band re)eater $as better
)er!ormance and )rovides better sinal quality. 'o#ever" t$e !ollo#in
)roblems are still )resent in a))lication*
+$e carriers o! a narro# band re)eater must outnumber t$e carriers con!iured
!or t$e source base station8 ot$er#ise t$e re)eater cannot ca)ture a c$annel.
+$e number o! )at$s o! many re)eaters is set to 4" so t$e base stations
outnumber 4 carriers cannot #or? as t$e sinal source.
For t$e base stations #it$ radio !requency and !requency $o))in" i! t$e
!requencies in t$e !requency $o))in set outnumber t$e )at$s selected by t$e
re)eater" t$e conversation cannot be maintained.
9$en t$e c$annel number o! t$e donor cell o! t$e re)eater c$anes" you must
adFust t$e c$annel number" ot$er#ise t$e )roblems suc$ as c$annel assinment
!ailure" call dro)" and inter!erence #ill occur.
+$e #ideband re)eater allo#s t$e base station to ado)t !requency $o))in"
and you do not $ave to adFust t$e c$annel number o! t$e re)eater a!ter t$e
c$annel number o! t$e donor cell c$anes i! t$e c$annel number is #it$in t$e
band#idt$ o! t$e re)eater. 'o#ever" t$e #ideband re)eater #ill am)li!y all
t$e sinals #it$in t$e band" so it causes reat inter!erence aainst ot$er cells.
Eo matter #$et$er t$e o)tical !iber or #ireless re)eater is a))lied" t$e sum o!
t$e radius o! t$e service area o! t$e re)eater and t$e distance bet#een t$e
re)eater and base station cannot brea? t$e +A limitation. For eneral base
stations" t$e distance bet#een a re)eater and t$e base station must be s$orter
t$an 21 ?ilometers.
+$e o)tical re)eater can be used in t$e areas #$ere t$e GSM radio sinals
cannot reac$ and no s)ace is le!t !or a re)eater. Because t$e transmission loss
o! o)tical !iber is small and its band#idt$ is #ide" t$e o)tical re)eater is quite
$el)!ul !or transmittin <F sinals.
;it$er an omni antenna or a directional antenna can be selected !or an o)tical
re)eater accordin to t$e actual land!orms. For an o)tical re)eater" its
transmission does not $ave to be isolated !rom t$e rece)tion. >n addition" t$e
address o! an o)tical re)eater is easy to be decided. Generally" an o)tical
re)eater is a))lied in t$e dead (ones #it$in countryside" $i$roads" tourin
areas" !actories" and urban areas.
>n remote mountain areas and alon $i$roads" you can also consider usin a
solar enery re)eater.
>n conclusion" t$e re)eater is used !or t$e !ollo#in )ur)oses*
;nlare coverae area and eliminate dead (ones.
Strent$ t$e !ield strent$ and enlare convere o! t$e base stations in urban
areas.
;nsure t$e coverae alon t$e $i$roads and tunnels.
<eali(e indoor coverae.
%.1%.% 9or?in @rinci)les o! <e)eater
>. 9ireless !requency selection re)eater
Fiure 1,41 s$o#s t$e #or?in )rinci)les o! a #ireless !requency selection
re)eater. +$e re)eater receives t$e <F sinals !rom t$e selected base station
-donor antenna. and am)li!ies and !or#ards t$e sinals. +$e antenna receivin
t$e sinals !rom t$e base station is called donor antenna" t$e ot$er antenna is
called retransmission antenna.
9or?in )rinci)les o! a #ireless !requency selection re)eater are as !ollo#s*
1. +$e lo#,noise )o#er am)li!ier )rocesses t$e sinals -received by t$e donor
antenna. !rom do#nlin? carriers.
%. +$e sinals -900 M'( <F sinals. are do#n converted into 51 M'(
intermediate !requency ->F. sinals.
2. +$e >F !ilter -#it$ a band#idt$ o! %00 &'(. am)li!ies t$e 51 M'( >F sinals
and u) converts t$e sinals into t$e 900 M'( <F sinals.
4. +$e retransmission antenna -service antenna. transmits t$e sinals to t$e
coverae areas.
+$e u)lin? sinals are also )rocessed accordin to t$e )revious )rocedures.
>>. 9ireless #ideband re)eater
+$e #ireless #ideband re)eater #or?s as t$e same #ay as t$e #ireless
!requency selection re)eater e:ce)t t$e !ilter )art. +$e band#idt$ o! t$e !ilter
o! t$e #ireless #ideband re)eater is !i:ed. Generally" it is 3M" 19M" or %1M.
>>>. A)tical re)eater
+$e di!!erence bet#een t$e o)tical !requency selection re)eater and t$e
o)tical #ideband re)eater lies in t$e coverae end. +$e !ormer ado)ts t$e
!requency selection com)onents" but t$e later ado)ts t$e variable band#idt$
o)tions.
Com)ared #it$ t$e #ireless re)eater" t$e o)tical re)eater does require
isolation bet#een donor antenna and retransmission antenna.
%.1%.2 <e)eater Eet#or? @lannin
>. <e)eater address selection
+$ere is no s)ecial requirement on t$e re)eater address selection e:ce)t t$e
!ollo#in items*
A re)eater address must lie bet#een t$e donor base station and t$e dead (one"
and t$e a(imut$ anle bet#een t$e donor antenna and t$e retransmission
antenna cannot be smaller t$an 90`" as s$o#n in t$e !ollo#in !iure.
>! t$e service antenna is a directional antenna" t$e re)eater must be installed
about %00 to 100 meters beyond t$e dead (one. >! t$e re)eater is installed
#it$in t$e dead (one" t$e coverae quality cannot reac$ t$e best" as s$o#n in
t$e !ollo#in !iure.
9$en t$e re)eater is used to coverae t$e dense residential areas at t$e edes
o! t$e urban area" it cannot !ace t$e buildins" because reat )enetration loss
#ill be caused. >n t$is case" t$e re)eater must be installed at t$e one side o!
t$e buildin" as s$o#n in t$e !ollo#in !iure.
+$e areas to be covered must meet t$e requirement o! line,o!,si$t
transmission.
+$e re)eater address must ensure t$e received sinal level required by t$e
re)eater. Generally" t$e received sinal level ranes !rom ,10 dBm to ,80 dBm.
Eo stron carrier #$ose c$annel number is t$e same as t$at o! t$e donor base
station is )resent at near t$e re)eater address.
+$e land!orms" buildins" or to#ers #$ere t$e donor antenna and
retransmission antenna can be installed. -+$e donor antenna must be directed
to t$e base station and t$e retransmission antenna must be directed to t$e
service area o! t$e re)eater. >n addition" t$e isolation bet#een t$e t#o
antennas must be reater t$an 150 dBc..
>>. Antenna selection
9$en selectin t$e antenna !or a re)eater" you must consider t$e !ollo#ins*
Select t$e )ro)er antenna ain accordin to t$e sinals and coverae condition
Do not ado)t t$e omni antenna because t$e #ireless re)eater is a!!iliated to
t$e intra,!requency relay system" ot$er#ise t$e system #ill )er!orm sel!,
e:citation.
+$e communication bet#een t$e donor antenna and t$e donor base station
antenna is )oint,to,)oint communication" so you must select t$e antenna #it$
$i$ ain or narro# $ori(ontal beam #idt$. For e:am)le" to reduce
inter!erence" you can select t$e re!lector antenna or t$e lo,)eriodical
antenna.
Select retransmission antenna accordin to t$e c$aracteristics o! a coverae
area. For a lare coverae area" you can select t$e eneral directional antenna
#it$ $i$ ain. For tunnel coverae" you can select t$e Rai antenna or t$e
s)iral antenna. For indoor coverae" you must select t$e antenna s)ecially
desined !or indoor use. Eo matter in #$at occasions" you must control t$e
transmit direction o! t$e retransmission antenna to )revent t$e retransmitted
sinals !rom !eedin in t$e donor antenna.
+$e !ront,to,bac? ratio o! t$e antenna must be as reat as )ossible -it is better
to be reater t$an 20 dB. so t$at a better isolation bet#een t$e donor antenna
and retransmission antenna can be ensured.
>>>. <equirements on antenna isolation
+$e isolation bet#een re)eater antennas de)ends on t$e $ost ain" but t$e
$ost ain cannot e:cel t$e isolation coe!!icient !or sel!,e:citation. Accordin to
t$e requirements in GSM )rotocols 02.20" t$e isolation must be at least 11 dB
reater t$an t$e $ost ain. >n actual )roFect desin" you can Fude #$et$er t$e
installation )osition meets t$e requirements on antenna isolation accordin to
on,site measurement.
Accordin to t$e !ormulas calculatin t$e antenna $ori(ontal isolation" t$e
!ollo#in !ormula can be deducted*
A' N 21.3 J %0 ld 0 -Gt J Gr. dB -900 M'(.
A' N 25.3 J %0 ld 0 -Gt J Gr. dB -1800 M'(.
'ere" 6d7 indicates t$e distance bet#een t$e donor antenna and
retransmission antenna" in t$e unit o! meter. Gt and Gr indicate t$e antenna
ain relative to t$e maFor lobe in t$e direction o! t$e t#o antennas. >! t$e t#o
antennas are bac?,to,bac? installed" Gt and Gr indicate t$e !ront,to,bac? ratio
o! t$e antenna.
'ori(ontal isolation o! re)eater antennas*
+$e !ormula calculatin t$e vertical isolation o! re)eater antennas is as
!ollo#s*
Av N 45.2 J 40 lod dB -900 M'(.
Av N 19.2 J 40 lod dB -1800 M'(.
Hertical isolation o! re)eater antennas*
>! t$e $ori(ontal isolation and vertical isolation are )resent simultaneously" t$e
total isolation can be calculated by t$e !ollo#in !ormula*
AS N -AH , A'. a=90 J A'" $ere AH indicates t$e vertical isolation8 A' indicates
t$e $ori(ontal isolation8 and 6a7 indicates t$e antenna included anle.
Donor antenna and retransmission antenna are installed on t$e to) o! t$e
buildin. Su))ose t$e $ost ain is 100 dB" t$e isolation bet#een t$e t#o
antennas can be 1%0 dB. >! t$e !ront,to,bac? ratio o! t$e donor antenna and t$e
retransmission antenna is 20 dB" #$en no barriers are )resent bet#een t$e t#o
antennas" t$e requirement on t$e isolation can be met.
>! t$e s)ace loss o! t$e sinals bet#een t$e t#o antennas is 30 dB" t$e
$ori(ontal isolation distance can be obtained" t$at is" d N %3m.
Durin )roFect im)lementation" you must select t$e antenna installation
)osition accordin to on,site measurement. Rou can use a sinal source and a
receiver !or t$e re)eater. >! t$e sinal attenuation bet#een t$e sinal source
and t$e receiver reac$es 30 dB" it means t$at t$e antenna installation )osition
meets t$e requirement on antenna isolation.
9$en installin t$e antenna !or a re)eater" you must )ay attention to t$e
!ollo#in items*
>! t$e antennas are $ori(ontally installed" t$e $ost o! t$e re)eater must be
installed bet#een t$e donor antenna and t$e retransmission antenna -it must
be nearer to t$e donor antenna..
A ood isolation must be ensured reardless t$at t$e antennas are $ori(ontally
or vertically installed. 9$en t$ey are $ori(ontally installed" it is better t$at
t$ere are some barriers lyin bet#een t$e donor antenna and t$e
retransmission antenna" because you do not $ave to )articularly desin a lare
installation s)ace to ensure antenna isolation in t$is case.
>H. /)lin? and do#nlin? balancce calculation
For a GSM re)eater" t$e lin? balance is reali(ed by !our lin?s" namely" t$e
u)lin? and do#nlin? bet#een t$e donor base station and re)eater" and t$e
u)lin? and do#nlin? bet#een t$e re)eater and mobile station.
+$is section em)loys t$e #ireless re)eater a))lied in outdoors as an e:am)le
to calculate t$e lin? balance. +o sim)li!y t$e calculation" #e introduce t$e
e!!ective donor )at$ loss -;Do@G." #$ic$ includes all t$e loss and ain !rom t$e
out)ut end o! t$e base station combiner or t$e in)ut end o! t$e multi,)at$
cou)ler to t$e in)ut end o! t$e re)eater.
+$e lin? balance is calculated accordin to t$e !ollo#in t#o !ormulas*
For do#nlin?s" @bout , ;Do@G J G<D , G<F J G<A , G)ass , @mn N @min.
For u)lin?s" @mout , G)ass J G<A , G<F J G</ , ;Do@G , @bn N @bin.
'ere"
@bout indicates t$e out)ut )o#er o! t$e base station.
@mout indicates t$e out)ut )o#er o! t$e mobile station.
G<D indicates t$e do#nlin? ain o! t$e re)eater.
G</ indicates t$e u)lin? ain o! t$e re)eater.
G<F indicates t$e !eeder loss o! t$e retransmission antenna.
G<A indicates t$e ain o! t$e retransmission antenna.
G)ass indicates t$e )at$ loss t$e mobile stations !rom t$e re)eater to t$e
service area.
@bn indicates t$e attenuation marin o! t$e mobile station.
@bin indicates t$e receivin level o! t$e base station.
@min indicates t$e receivin level o! t$e mobile station.
B+Ssens indicates t$e base station sensitivity.
MSsens indicates t$e mobile station sensitivity.
>! t$e u)lin? ;Do@G and do#nlin? ;Do@G are equal to t$e u)lin? )at$ loss and
t$e do#nlin? )at$ loss !rom t$e re)eater and mobile station" t$e attenuation
marin o! t$e base station is equal to t$at o! t$e mobile station. +$ere!ore" i!
you subtract t$e !ormula calculatin u)lin? balance !rom t$e !ormula
calculatin do#nlin? balance" you can et @bout , @mout J G<D , G</ N @min ,
@bin.
>! t$e lin?s are balance" t$e equation @min , @bin N Dsens N MSsens, B+Ssens is
)resent. >n t$is case" t$e !ormula calculatin lin? balance is @bout , @mout J
G<D , G</ N Dsens.
+$ere!ore" t$e Dsens is !i:ed a!ter t$e base station equi)ments are selected.
Moreover" t$e out)ut )o#er o! t$e base station and mobile station may be
decided in GSM system )lannin. As a result" to ac$ieve t$e balance o! t$e
#$ole lin?s" you need to adFust t$e u)lin? ain and do#nlin? ain o! t$e
re)eater only.
+$e !ollo#ins em)loy t$e re)eater system installed in outdoors as an e:am)le
to calculate t$e #$ole lin? balance.
For do#nlin? budet o! t$e outdoor re)eater " out)ut )o#er o! t$e transmitter
-J42dBm. 0 loss o! t$e combiner -4dB. 0 ;do@G -90dB. N in)ut )o#er o! t$e
re)eater -,11dBm. J do#nlin? ain o! t$e re)eater -80dB. N do#nlin? out)ut
)o#er o! t$e re)eater -J%9dBm. 0 !eeder loss o! t$e retransmission antenna
-2dB. J ain o! t$e retransmission antenna -18dBi. 0 )at$ loss o! t$e re)eater in
t$e coverae area -1%5dB. N in)ut level o! t$e mobile station -,82dBm. 0
attenuation marin -%0dBm. N t$e mobile station sensitivity -,102dBm..
^ Eote*
+o obtain t$e value o! ;Do@G" you can measure t$e in)ut level o! t$e donor
re)eater and out)ut level o! t$e base station combiner !irst" and t$en obtain
t$e di!!erence bet#een t$e t#o" and t$e di!!erence is t$e value o! ;Do@G. >n
addition" t$e ain o! t$e mobile antenna must be converted to 0 dBi.
For u)lin? budet o! t$e outdoor re)eater" out)ut )o#er o! t$e mobile station
transmitter -J22dBm. 0 )at$ loss o! t$e re)eater in t$e coverae area -1%5dB. J
ain o! t$e retransmission antenna -18dBi. 0 !eeder loss o! t$e retransmission
antenna -2dB. N in)ut )o#er o! t$e re)eater -,59dBm. J u)lin? ain o! t$e
re)eater -80dB. N out)ut )o#er o! t$e re)eater -J1dBm. 0;do@G -90dB. N in)ut
level o! t$e base station -,89dBm. 0 attenuation marin -%0dBm. N base station
sensitivity -,109dBm..
^ Eote*
Because you do not $ave to consider t$e diversity !unction" t$e attenuation
marin on u)lin?s is t$e same as t$at on do#nlin?s. Accordin to t$e )revious
lin? budet" t$e do#nlin?s are restricted by t$e out)ut )o#er o! t$e re)eater"
t$e u)lin?s are restricted by t$e out)ut )o#er o! t$e mobile station" and t$e
noise restricts t$e ma:imum ain -;Do@G,10 dB." so t$e lin? balance is )resent.
'o#ever" t$is is t$e most common situation. Actually" you must calculate t$e
marin !or all lin?s #$en installin or o)timi(in t$e re)eater system. +$e
latest re)eater su))orts t$e u)lin? ain and do#nlin? ain to be set
res)ectively.
'ereunder is an e:am)le.
+$ere is a base station coverin )arts o! a $i$road. >ts coverae radius is about
%0 ?m.
+$e measured sinal strent$ at t$e edes o! t$e base station cells is ,92dBm.
+$e micro#ave lin? to#er on t$e to) o! t$e $ill near t$e base station is
selected as t$e address o! t$e re)eater.
>n t$e areas -includin mountains. 210m belo# t$e to) o! t$e to#er" t$e
received level o! t$e mobile station is ,51 dBm.
+$e lo,)eriodical antenna #it$ a ain o! 18dBi and an a(imut$ anle o! 21`is
used as t$e donor antenna.
+$e antenna is installed at 11 meters under t$e to#er to) and !aces t$e base
station.
>! t$e )revious conditions are )resent" t$e sinals out)ut by t$e re)eater are
,14 dBm. >! a )lane antenna #it$ a ain o! 15 dBi and a $ori(ontal a(imut$
anle o! 30 derees is installed at t$e to) o! t$e to#er and t$e antenna
radiates to t$e reverse direction o! t$e donor antenna" t$e requirements on
antenna isolation can be met even i! t$e ain o! t$e re)eater reac$es 81 dB. >n
t$is case" t$e out)ut )o#er o! t$e re)eater is 20 dBm. And t$e level o! t$e
sinals in t$e areas alon t$e $i$road #$ic$ are %0 ?m beyond t$e to#er can
reac$ ,90 dBm. +$ere!ore" t$e radius o! t$e cell alon t$e $i$road is enlared
by 10L.
^ Eote*
>! a retransmission antenna is installed at t$e to) o! t$e to#er" you must
ensure t$at t$e received sinal level in t$e (ero )oint !illin areas near t$e
to#er.
H. <e)eater out)ut )o#er control
9$en ado)tin a re)eater" you must )ay s)ecial attention to t$e e!!ect o! t$e
intermodulation )roducts aainst t$e system. +$e intermodulation )roducts o!
t$e re)eater de)end on t$e number o! t$e am)li!ied carriers" t$e out)ut )o#er
o! eac$ carrier" and t$e linearity o! t$e am)li!ier.
Ginearity o! t$e am)li!ier*
+$ird order intermodulation #ill increase #it$ out)ut )o#er due to t$e
nonlinearity o! t$e am)li!ier. +$ere!ore" you must control t$e out)ut to a
certain deree to ensure t$at t$at t$e inde:es on t$ird order intermodulation
meet t$e requirements. +$e !ollo#in !ormula s$o#s t$e relations$i) bet#een
t$e out)ut )o#er o! eac$ carrier o! t$e re)eater and t$e requirements on t$ird
order intermodulation.
@o N >@2 J -@>M@=%. J10 l -E=%.
'ere"
@o indicates t$e out)ut )o#er o! eac$ carrier -dBm.
>@2 indicates t$e t$ird order section o! t$e am)li!ier -dBm.
@>M@ indicates t$e level o! t$e t$ird order intermodulation -dBc.
E indicates t$e number o! carriers
>! t$e t$ird order section o! t$e am)li!ier o! a ty)ical re)eater is 10 dBm" and
t$e intermodulation level must be lo#er t$an ,41 dBc accordin to t$e
requirement o! t$e #ireless communication institutes in Britain.
H>. <e)eater ain settin
+$e ain o! t$e early re)eaters must be set manually" but t$e latest ain o! t$e
latest re)eaters can be automatically set. For t$e re)eaters #$ose ain is set
manually" t$e sum o! t$e re)eater ain and t$e )rotection marin must be
equal to or smaller t$an t$e re)eater isolation8 ot$er#ise" t$e sel!,e:citation
o! t$e re)eater #ill be caused. 'ere t$e re)eater isolation indicates t$e
isolation bet#een t$e donor antenna and t$e retransmission antenna o! t$e
re)eater. Generally" t$e )rotection marin ranes !rom 10 dB to 11 dB.
H>>. <e)eater adFacent cell )lannin
+$e coverae areas o! a re)eater may overla) ot$er donor cells" so you must
con!iure t$e corres)ondin adFacent cell relations$i) !or t$e re)eater to
ensure normal $andover. >n addition" you must )ay attention t$at t$e
!requencies in t$e coverae areas o! t$e re)eater and t$at in t$e donor cells
cannot be t$e same !requency and nei$bor !requency.
H>>>. ;!!ect o! delay )rocessin aainst re)eater )lannin
>! only one re)eater cannot !ully cover an area -suc$ as a narro# and lon
tunnel." you can use several cascaded re)eaters to )rovide t$e coverae. +$e
selection o! t$e address and antenna !or t$e re)eater o! eac$ level is t$e same
as t$at !or a sinle re)eater.
'o#ever" t$e re)eater #ill am)li!y t$e same !requency and it ta?es some time
!or t$e re)eater to )rocess t$e sinal" so t$ere is a delay !or eac$ sinal
sement. >! t$e delay is reater t$an t$e time !or t$e GSM system to identi!y
t$e time #indo#" t$e intra,!requency inter!erence #ill occur. +$ere!ore" you
must consider t$e e!!ect o! t$e delay #$en ado)tin cascaded re)eaters"
because t$e delay #ill also accelerate t$e time dis)ersion and s$orten t$e
coverae distance.
>! ado)tin t$e o)tical re)eater" you must consider t$at t$e transmission s)eed
o! t$e sinals in o)tical !ibers is %=2 t$at o! in !ree s)ace" namely" i! t$e
e:tension cell tec$noloy is not used" t$e ma:imum transmission distance o!
t$e sinals in o)tical !iber is 21 ?m multi)lies %=2 -about %2.2 ?m. due to t$e
restriction on transmission delay.
>n addition" i! one o! t$ree sync$ronous cells ado)tin t$e o)tical re)eater" t$e
+A o! t#o cells #ill be di!!erent due to t$e di!!erence o! transmission mode
and rate. >n t$is case" t$e sync$ronous $andover !ailure #ill occur. +$ere!ore"
you must ado)t t$e async$ronous $andover to obtain t$e +A o! a ne# cell"
#$ic$ #or?s as t$e $andover taret cell.
+$e delay )rocessin varies #it$ re)eater ty)es. Some ta?e % to 2 Ks and some
ta?es 1 to 3Ks. >n a GSM system" t$e delay o! t#o sinals cannot be reater
t$an 13Ks. For t$e e!!ect o! re)eater delay )rocessin aainst time dis)ersion.
Distance bet#een )oint A and t$e re)eater 6d7 is %.1?m. +$e delay !or t$e
mobile station at )oint A to receive t$e sinals !rom t$e re)eater and t$e cell
is as !ollo#s*
-%.1?m J %.1?m.=c -li$t s)eed. J 2Ks N 14Ks J 2Ks N 15Ks S 13Ks.
>n t$is case" t$e intra,!requency inter!erence may be )resent. >! t$e di!!erence
o! t$e levels o! t$e t#o sinals is equal to or lo#er t$an 1% dB" t$e
conversation quality #ill be a!!ected.
+$e time dis)ersion #ill cause intra,!requency inter!erence" and t$e time
dis)ersion is caused by t$e overla) o! t$e sinal source cell and t$e re)eater
coverae area. +$ere!ore" you must select t$e sinal o! t$e secondary cells in
t$e coverae areas o! t$e re)eater instead o! t$e sinals o! t$e maFor service
cell as t$e source sinal o! t$e re)eater. >n t$is case" t$e time dis)ersion
caused by overla) can be avoided.
>O. ;!!ect o! bac?round noise aainst re)eater )lannin
Su))ose t$at t$e ma:imum received noise level allo#ed by t$e base station is
DE" i! t$e u)lin? bac?round noise level o! t$e re)eater $ost is too reat" t$e
base station c$annels #ill be conested #$en t$e noise level at t$e base
station is reater t$an DE. 'o#ever" $o# to set t$e re)eater #it$out a!!ectin
t$e base stationd +$ey are introduces as !ollo#s.
>! t$e !ollo#in assum)tions are )resent*
+$e transmitted sinal strent$ o! t$e base station is +b.
+$e received sinal strent$ o! t$e base station is <b.
+$e received do#nlin? sinal strent$ o! t$e base station $ost is Dr.
+$e transmitted u)lin? sinal strent$ o! t$e base station $ost is /t.
>n t$is case" t$e )at$ loss bet#een t$e base station and t$e re)eater is +b,Dr"
so <b N /t 0 -+b,Dr.. As a result" i! t$e re)eater does not a!!ect t$e base
station" <b T DE" so t$e !ollo#in t#o inequities are )resent*
/t 0 -+b , Dr. T DE
/t T +b,Dr J DE
Accordin to t$e )revious analysis" t$e re)eater does not a!!ect t$e base
station i! t$e u)lin? bac?round noise level out)ut by t$e re)eater $ost is
lo#er t$an -+b,DrJDE.. From t$is )ers)ective o! revie#" t$e bac?round noise
must be )articularly em)$asi(ed in re)eater )lannin because it is easier to
brin inter!erence t$an ot$er ty)es o! base stations.
O. S)eci!ications o! #ireless re)eaters
%.12 Conclusion
Eet#or? )lannin is t$e !oundation o! a mobile communication net#or?"
es)ecially t$e #ireless )arts in a mobile communication net#or? costs reat
and is o! vital im)ortance to net#or? quality" so you must ma?e a ood
)lannin at earlier stae" #$ic$ is $el)!ul !or net#or? e:)ansion and service
u)date in t$e !uture.
Eet#or? )lannin requires enineers to analy(e coverae" decide net#or?
layers" and analy(e tra!!ic based on relative tec$noloies and )arameters" and
!inally out)ut t$e results o! <F )lannin" includin base station layout and
scale.
<F )lannin" as #ell as t$e a))lication o! cell )arameters" determines t$e cell
coverae. +$e cell coverae must be )ro)erly desined so t$at t$e mobile
station can al#ays enFoy t$e best service at t$e best cells. >n addition" t$e cell
coverae must be desined in a #ay conducive to net#or? ca)acity e:)ansion.
+$is c$a)ter also introduces t$e solutions to dual,band net#or?" indoor
coverae" tunnel coverae" and so on. Gast" t$is c$a)ter introduces t$e
re)eater a))lication
2 GSM Frequency @lannin
2.1 Avervie#
Frequency resource is scarce !or t$e mobile communication" so $o# to
ma:imi(e t$e s)ectrum utili(ation ratio is a reat concern !or many carriers"
equi)ment )roviders" and sc$olars. And t$eir researc$ into t$is )roblem $as
accelerated t$e develo)ment o! t$e communication tec$noloies. By no#" t$e
mobile communication $as e:)erienced t$ree )$ases* analo +ACS=AM@S"
GSM=CDMA >S91" and 9CDMA=CDMA%000.
+$e )ur)ose to en$ance t$e s)ectrum utili(ation ratio is to e:)and t$e net#or?
ca)acity based on t$e limited s)ectrum resource #$ile ensurin t$e net#or?
quality. >! not considerin addin !requencies to t$e net#or?" you can e:)and
t$e ca)acity o! a GSM net#or? usin t$e t#o met$ods. Ane is to increase t$e
number o! base stations in t$e net#or?8 t$e ot$er is to use t$e !requency reuse
tec$noloies. +$is c$a)ter mainly describes t$e GSM !requency reuse
tec$noloies" namely" !requency )lannin tec$noloies.
+o e:)and t$e net#or? ca)acity" you must reuse t$e limited !requency
resources. +$ou$ !requency reuse is bene!icial !or net#or? e:)ansion" it
brins into anot$er )roblem. +$at is" it deteriorates t$e conversation quality.
+$e more aressive t$e !requencies are reused" t$e reater t$e inter!erence
#ill arise in t$e net#or?. +$ere!ore" $o# to see? a balance bet#een net#or?
ca)acity and conversation quality is a demandin tas? in !requency )lannin.
Currently" t$e 4 : 2" 2 : 2" % : 3" 1 : 2" 1 : 1" M<@" and concentric circles are
t$e GSM !requency tec$noloies in common use. For t$e 4 : 2 !requency reuse
)attern" t$e !requency utili(ation ratio is relatively lo#" but t$e $i$er carrier,
to,inter!erence ratio -C=>. can be obtained" so you can enFoy better
conversation quality. Com)ared #it$ t$e 4 : 2 !requency reuse )attern" t$e 1 :
2 !requency reuse )attern ensures a relatively $i$ !requency utili(ation ratio"
but t$e reuse distance is s$orter" so inter!erence is reater and t$e
conversation quality is )oorer. >n t$is case" you s$ould ta?e some measures"
suc$ as t$e !requency $o))in and D+O" aainst t$e inter!erence.
+$e !requency )lannin is a ?ey tec$noloy !or GSM net#or?" so t$e quality o!
t$e !requency )lannin #ill determine t$e net#or? quality.
+$is c$a)ter introduces t$e rules o! !requency reuse based on t$e !requency
reuse )atterns and t$e net#or? requirement. Mean#$ile" it also )rovides
e:am)les to detail t$e !requency division" C=>" !requency reuse deree under
eac$ reuse )attern.
2.% Frequency Division and C=> <equirement
2.%.1 Frequency Division
+$e GSM cellular system can be divided into GSM 900M'( system and DCS
1800M'( system in terms o! t$e band to be used. +$e carrier s)acin is %00
&'(.
>. GSM 900M'(
>t $as 1%4 c$annel numbers. +$e absolute radio !requency c$annel number
-A<FCE. is 101%4" and a )rotection band #it$ %00 &'( in #idt$ is reserved at
t$e t#o ends. Accordin to t$e documents )rescribed by t$e relative
overnment de)artment o! C$ina" C$ina Mobile uses t$e 8900909=9230914M'(
band" and t$e corres)ondin A<FCE is 1091 -enerally" t$e c$annel number 91
is !or reservation only.. For C$ina /nicom" it uses t$e 9090911=9140930M'(
band" and t$e corres)ondin A<FCE is 9301%4. For t$e bands de!ined !or t$e
carriers !rom ot$er countries" t$ey can be calculated by t$e !ollo#in
!ormulas*
Base station rece)tion* !1 -n. N [890.% J -n 0 1. : 0.%\ M'(
Base station transmit* !% -n. N [!1 -n. J 41\ M'(
>>. DSC 1800M'(
>t $as 254 c$annel numbers. +$e A<FCE is 11%0881. +$e relations$i) bet#een
t$e !requency and t$e c$annel number -n. are listed in t$e !ollo#in*
Base station rece)tion* !1 -n. N [1510.% J -n 0 11%. : 0.%\ M'(
Base station transmit* !% -n. N [!1-n. J 91\ M'(
C$ina Mobile uses t$e 1510015%0 M'( band" and t$e corres)ondin A<FCE is
11%0131. C$ina /nicom uses t$e 154101511 M'(" and t$e corres)ondin A<FCE
is 3850523.
2.%.% C=>
C=> stands !or carrier,to,inter!erence ratio. >n t$e GSM system" !requency reuse
#ill cause intra,!requency inter!erence. +$e intra,!requency is related to bot$
t$e reuse distance and t$e cell radius.
>! t$e intra,!requency cell and t$e service cell #or? at t$e same time" t$e MS
locatin in t$e center o! t$e service cell #ill receive bot$ t$e use!ul sinals
!rom t$is service cell and t$e inter!erin sinals !rom t$e intra,!requency cells.
For t$e omni,directional base station #it$ reular !requency reuse" t$ere are 3
intra,!requency inter!erence sources at t$e !irst layer" namely" t$e 3 intra,
!requency reuse cells in orane. +$ere are 1% intra,!requency inter!erence
sources at t$e second layer" namely" t$e 1% intra,!requency reuse cells in
yello#. 'o#ever" t$e 1% intra,!requency inter!erence sources $as only a little
e!!ect on t$e 3 inter!erence sources at t$e !irst layer" so it can be nelected.
>! t$e radio )ro)aation environment bet#een t$e 3 intra,!requency reuse cells
and t$e service cell is t$e ?ee)s stable.
9$en t$e MS locates at t$e ede o! t$e service cell" it #ill receive t$e )oorest
sinals !orm t$e service cell but t$e stronest inter!erin sinals. >n t$is case"
t$e needed C=> can be e:)ressed by t$e !ollo#in equation*
>! t$e cellular layout is im)ro)erly desined" t$e inter!erin sources #ill
increase and t$e C=> #ill decrease. Accordin to t$e )revious equations" t$e
more t$e cells in eac$ cluster" t$e reater t$e C=> and t$e better t$e net#or?
quality are" but t$e !requency utili(ation ratio #ill be lo#er. >n addition" t$e
GSM inter!erence is related to t$e tra!!ic load. +$e intra,!requency
inter!erence reac$es t$e reatest #$en t$e tra!!ic load reac$es t$e )ea?.
Generally" t$e 4 : 2 !requency reuse )attern is used in GSM !requency
)lannin. For t$e areas #$ere t$e tra!!ic is reat" you can use ot$er !requency
reuse )atterns" suc$ as 2 : 2 and 1 : 2. Eo matter #$ic$ !requency reuse
)attern you ta?e" you must meet t$e requirement on inter!erence,to,
)rotection ratio.
A)art !rom t$e intra,!requency inter!erence caused by normal !requency reuse"
t$ere are ot$er abnormal inter!erences. +$ey are listed in t$e !ollo#in*
Multi)at$ sinal inter!erence ->t occurs #$en use!ul sinals !all outside t$e
delay equali(er o! t$e system..
Autside sinal inter!erence ->t re!ers to t$e sinals !rom t$e radar" illeal
#ireless equi)ments" and environment noises..
>n t$e GSM system" t$e requirements on t$e C=> are listed in t$e !ollo#in*
For intra,!requency C=>" it must be equal to reater t$an 9 dB. >n actual
)roFectin" a marin o! 2 dB is needed" namely" it is equal to or reater t$an 1%
dB.
For adFacent,!requency C=>" it must be equal to or reater t$an ,9 dB. >n actual
)roFectin" a marin o! 2 dB is needed" namely" it is equal to or reater t$an ,3
dB.
9$en t$e carrier o!!set reac$es 400 &'(" t$e C=> must be equal to or reater
t$an ,41 dB.
2.2 Frequency @lannin @rinci)le
Generally" #$en )lannin t$e !requency !or t$e net#or?" you #ill divide t$e
eora)$ic area into smaller slices" but you must reserve a certain amount o!
c$annel number at t$e intersection area bet#een slices i! t$e !requency
resource is adequate.
+$e intersection area must be !ar a#ay !rom t$e areas #$ere t$e tra!!ic is
reat and t$e areas #$ere t$e net#or?in is com)le:. Generally" you s$ould
bein t$e )lannin #it$ t$e area #$ere base stations are intensively
distributed. >! t$ere are rivers or bi la?es in t$e )lannin area" you must
consider t$e re!ection e!!ect o! t$e sur!ace.
Generally" base stations irreularly distributed" so you cannot )er!orm t$e
!requency )lannin com)letely accordin to 4 : 2 !requency reuse )attern or 2
: 2 !requency reuse )attern. >nstead" you must ma?e !le:ible adFustment
accordin to actual conditions.
Eo matter #$ic$ reuse )attern you ta?e" you must obey t$e !ollo#in
)rinci)les*
, Generally" t$e intra,!requencies and adFacent cannel numbers are allo#ed to
a))ear #it$in a base station.
, +$e !requency s)acin bet#een t$e BCC' and +C' must be reater t$an 400
&'( #it$in a cell.
, +$e !requency s)acin bet#een t$e +C's must be reater t$an 400 &'(
#it$in a cell. -9$en !requency $o))in is used" you can meet t$is by )ro)erly
settin t$e mobile allocation inde: o!!set..
, +$e adFacent base stations cannot use t$e same !requency.
, Considerin t$e com)le:ity o! t$e antenna $ei$t and radio )ro)aation
environment" t$e base stations near eac$ ot$er cannot use t$e same
!requency.
, Generally" i! usin t$e 1 : 2 !requency reuse )attern" you must ensure t$at
t$e number o! !requency $o))in c$annel numbers is at least t#ice t$at o! t$e
!requency $o)in carriers.
, @ay s)ecial attention to t$e intra,!requency reuse. +$e adFacent areas are
not allo#ed to s$are t$e BCC' and t$e BS>C.
3. !or"al Fre#uency $euse Technology
3..1 C%& under ' 3 Fre#uency $euse Pattern
The spectrum utilization ratio can be expressed by frequency reuse degree,
which reveals the aggressiveness of the frequency reuse. The frequency reuse
degree can be expressed by the following equation: f
reuse
=N
!"#N
$

N
T!%
&ere N
!"#N
is the total number of the available channel numbers, and N
T!%
is
the number of T!%s configured for the cell.
"or the n x m frequency reuse pattern, 'n' indicates the number of the base
stations in the reuse clusters, and 'm' indicates the number of the cells under
each base station. (n this case, the frequency reuse degree can be expressed
by the following equation:
f
reuse
= n x m
(n actual planning, however, the allocated number of channel numbers will be
greater than n x m, so the actual f
reuse
is usually greater than n x m.
Therefore, the smaller the f
reuse
, the more aggressive the frequency is reused
and the higher the frequency utilization ratio is. s the aggressiveness of the
frequency reuse grows, however, it will bring greater interference to the
networ). (n this case, you must enable the technologies, including *T% and
power control, to solve this problem. The more aggressive the frequency is
reused, the lower the spectrum utilization ratio is, but the conversation quality
is better at this time.

+$e )ur)ose t$e !requency )lannin is to reac$ a balance bet#een t$e
!requency utili(ation ratio and t$e net#or? ca)acity. Based on t$e assurance o!
t$e net#or? quality" you must ta?e measures to ma:imi(e t$e net#or?
ca)acity.
(n the +,- system, the . x / frequency reuse pattern is in basic use. &ere '.'
indicates . base stations 0each base station consists of / cells1, and '/'
indicates the / cells under the control of each base station. Therefore, there
are 23 sectors are available. nd the 23 sectors ma)es up of a frequency reuse
cluster, but the frequency in the same cluster cannot be reused.
"or the . x / frequency reuse pattern, the intra4frequency spacing is great, so
it can meet +,- system5s requirement on the intra4frequency interference
protection ratio and ad6acent frequency interference protection ratio. s a
result, this frequency reuse pattern is good for the networ) quality and
security. 7nder the . x / frequency reuse pattern, the frequency reuse
aggressiveness is 23.
"or the aggressive reuse introduced hereunder, because the 8##& plays an
important role in the networ) and you cannot use the apply the anti4
interference measures, such as downlin) power control and *T%, to the
8##&, you must apply the . x / frequency reuse pattern or looser reuse
patterns to the 8##& carriers.
Eormal 4 : 2 !requency reuse )attern.
3..2 1(M)* +andwidth ' 3 Fre#uency $euse
&ereunder are several assumptions:
+$e available band#idt$ is 10M'(.
+$e c$annel number is 41094.
>! t$e c$annel numbers ranin !rom 81094 -14 c$annel
numbers in total. are allocated to t$e BCC'" and t$e ot$er
c$annel numbers are allocated to +C'.
(f the previous assumptions are present, the frequency planning under . x /
frequency reuse pattern is provided in
Frequency )lannin under 4 : 2 !requency reuse )attern -a.
"requency
group
number
A1 B1 C1 D1 A2 B2 C2 D2 A3 B3 C3 D3
#hannel
number of
each
frequency
group
94 93 92 91 90 89 88 87 86 85 84 83
80 79 78 77 76 75 74 73 72 71 70 69
68 67 66 65 64 63 62 61 60 59 58 57
56 55 54 53 52 51 50 49 48 47 46 45

ccording to this table, the channel numbers in the first line are 8##&
numbers, in which the channel numbers 92 and 93 are standby channel
numbers. The channel number of 8##& of the cell 2 is :.. (t is 9;, <9 and
=< for other carriers, and so on.
(n a cluster which contains 23 cells, the frequency group for base station is
>2, 3, and /?@ the frequency group for base station 8 is >82, 83, and
8/?@ the frequency group for base station # is >#2, #3, and #/?@ and the
frequency group for base station * is >*2, *3, and */?.
Therefore, as listed in this table, no channel number is reused within a cluster.
(n addition, the intra4frequency and ad6acent frequency are not available for
the ad6acent cells and the same cell.
&owever, the drawbac)s of this frequency reuse pattern are that the frequency
reuse ratio is low and the capacity expansion needs a great amount of the
frequency resources. Therefore, this reuse pattern is not used in the areas
where the networ) capacity needs to be constantly expanded.
(f the bandwidth is 2;-&z, the maximum base station configuration is ,.$.$.
under the normal . x / frequency reuse pattern, and the frequency reuse
degree is 23.= 0=;$. = 23.=1.
& !ote,
The maximum base station type mentioned in the chapter refers to the
configuration type that most continuous base stations can reach. (t does not
include standalone base station.

3..3 1-M)* +andwidth ' 3 Fre#uency $euse
"or the 2:-&z frequency 02 to :.1 used by #hina -obile, the . x /
frequency reuse pattern are used for the frequency planning. The channel
numbers ranging from A: to :. 02< channel numbers in total1 are allocated to
the 8##&, and other channel numbers are allocated to T#&. No channel
number is reserved for micro cells. (n this case, the frequency planning
solution is provided in "requency planning under . x / frequency reuse
pattern 0b1
Frequency )lannin under 4 : 2 !requency reuse )attern -b.
"requency
group
number
A1 B1 C1 D1 A2 B2 C2 D2 A3 B3 C3 D3
#hannel
number of
each
frequency
group
94 93 92 91 90 89 88 87 86 85 84 83
78 77 76 75 74 73 72 71 70 69 68 67
66 65 64 63 62 61 60 59 58 57 56 55
54 53 52 51 50 49 48 47 46 45 44 43
42 41 40 39 38 37 36 35 34 33 32 31
30 29 28 27 26 25 24 23 22 21 20 19
18 17 16 15 14 13 12 11 10 9 8 7
6 5 4 3 2 1

s listed in this table, the channel numbers ranging from A: to 93 are standby
channel numbers. "or the 2:-&z bandwidth, the maximum base station type
can be ,9$A$A under . x / frequency reuse pattern. The frequency reuse
degrees are 22.A=, 2/../, and 2/../, so the average value is 23.9A.
3.. .M)* +andwidth ' 3 Fre#uency $euse
"or the <-&z frequency 0:< to 23.1 used by #hina 7nicom, the . x /
frequency reuse pattern is used for the frequency planning. The channel
numbers ranging from 222 to 23. 02. channel numbers in total1 are allocated
to the 8##&, and other channel numbers are allocated to T#&. No channel
number is reserved for micro cells. (n this case, the frequency planning
solution is provided in:
Frequency )lannin under 4 : 2 !requency reuse )attern -c.
"requency
group
number
A1 B1 C1 D1 A2 B2 C2 D2 A3 B3 C3 D3
#hannel
number of
each
frequency
group
124 123 122 121 120 119 118 117 116 115 114 113
110 109 108 107 106 105 104 103 102 101 100 99
98 97 96

s listed in this table, the channel numbers ranging from 222 to 223 are
standby channel numbers. "or the <-&z bandwidth, the maximum base
station type can be ,/$3$3 under . x / frequency reuse pattern. The frequency
reuse degrees are :.<A, 2/.=, and 2/.=, so the average value is 23.33.
3../ ' 3 Fre#uency $euse Conclusion
The . x / frequency reuse pattern is a basic technology applied in frequency
planning. (t is applicable to other frequency aggressive reuse technologies
that are used for the 8##&.
Theoretical analysis shows that when the base stations are regularly
distributed and azimuths of the cells are consistent with each other, the
interference can be reduced to the minimum. Therefore, if you intend to
expand the networ) capacity, you can )eep the base stations to be distributed
as regular as possible and plan the azimuths of the cells along the same
direction. (n addition, you can also maintain the antennas at a similar height.
&owever, sometimes you need to ad6ust the azimuth of the antenna to
improve the coverage, which seems contradicts to the capacity expansion.
Therefore, sometimes you must ma)e find a balance between the coverage
and capacity.
(f the networ) capacity needs to be further expanded, you can ta)e the
following measures:
S)lit a cell into smaller cells. At )resent" $o#ever" t$e
averae coverae radius o! t$e macro cell base stations in urban
areas is already s$orter t$an 100m" so !urt$er cell s)littin #ill
meet di!!iculty in cost and tec$noloy.
/tili(e ne# !requency resources. For e:am)le" you can
em)loy t$e 1800M'( band to establis$ a DSC 1800M'( net#or?.
/nder t$e current 900M'( net#or?" use more aressive
!requency reuse tec$noloy to e:)and t$e net#or? ca)acity.
t present, the aggressive frequency reuse technology wor)s as the most
economical and convenient way to expand the networ) capacity, so it is also
the most popular with carriers.
The typical frequency reuse technology includes / x /, 3 x <, 3 x /, 2 x /, and
2 x 2.
3./ Aggressi0e Fre#uency $euse Technology
3./.1 3 ' 3 Fre#uency $euse Pattern
The / x / frequency reuse pattern can be used in the areas with high traffic.
That is, three base stations form a group, and each base station has three cells,
so there are : cells, which form a frequency reuse cluster. &owever, the :
cells use different frequencies. #ompared with the . x / frequency reuse
pattern, the intra4frequency reuse distance under the / x / frequency reuse
pattern is small, so on4line interference is greater.
>! t$e available band#idt$ is 10M'( and t$e c$annel numbers are !rom 41 to
94" you can use normal 4 : 2 !requency reuse )attern on BCC'. >n t$is case" t$e
!requency ranes !rom 81 to 94" so 14 c$annel numbers are available. For +C'"
you can use 2 : 2 !requency reuse )attern. >n t$is case" t$e !requency ranes
!rom 41 to 80" so 23 c$annel numbers are available.
For t$e !requency )lannin under 2 : 2 !requency reuse )attern" see
Frequency )lannin under 2 : 2 !requency reuse )attern
Frequency )lannin under 2 : 2 !requency reuse )attern Frequency )lannin
under 2 : 2 !requency reuse )attern
"requency
group
number
A1 B1 C1 A2 B2 C2 A3 B3 C3
#hannel
number of
each
80 79 78 77 76 75 74 73 72
71 70 69 68 67 66 65 64 63
62 61 60 59 58 57 56 55 54
frequency
53 52 51 50 49 48 47 46 45

(f / x / reusing the 2;-&z band, you can configure the maximum base
station type as ,=$=$=, and the frequency reuse degree is 2;.
Accordin to )revious equations" because t$e number o! base stations is 2 -E N
2." t$e intra,!requency inter!erence attenuation !actor is 2 -q N 2.. >n t$is case"
t$e number o! t$e intra,!requency inter!erence sources is % at t$e !irst layer.
9$en t$e band#idt$ is 10M'(" t$e base station ty)e can be con!iured as
S1=1=1 under 2 : 2 !requency reuse )attern. For 4 : 2 !requency reuse )attern"
t$e ma:imum base station con!iuration ty)e can only be con!iured as
S4=4=4=. +$ere!ore" net#or? ca)acity under 2 : 2 !requency reuse )attern is
reater t$an t$at under 4 : 2 !requency reuse )attern #$en t$e band#idt$ is
t$e same.
Bhen the number of subscribers in a networ) is not great, you can use the / x
/ frequency reuse pattern to ease the pressure of networ) capacity. (n actual
conditions, however, because base stations are irregularly distributed, the
antenna height is different, and the coverage area of each base station varies,
the interference in the networ) will increase. (n this case, if you intend to
obtain better voice quality, you must ta)e some anti4interference measures,
such as using frequency hopping and *T%.
The characteristic of the / x / frequency reuse pattern are as follows:
+$e adFustment !or net#or? structure is unnecessary.
+$e !requencies can be easily rou)ed and t$e system
ca)acity is reat.
Com)ared #it$ 4 : 2 !requency reuse )attern" 2 : 2
!requency reuse )attern brins reater inter!erence" but t$e
overall inter!erence can be controlled to a lo#er level.
>! !requency $o))in is used" adequate band#idt$ is
needed.
3./.2 2 ' . $euse Pattern
The 3 x < frequency reuse pattern is developed from the . x / frequency reuse
pattern. 7nder the . x / frequency reuse pattern, you can add anther 3 cells to
each base station, so 3 base stations 0each base station has < <;C4sectorized
cells1 has 23 cells, which form a frequency reuse cluster. (n this case, a
frequency reuse cluster contains 23 <;C4sectorized cells, and this is defined as
3 x < frequency reuse pattern.
3 x < frequency reuse pattern3 x < frequency reuse pattern.
% : 3 !requency reuse )attern
7nder the 3 x < frequency reuse pattern, .
8ecause each cell is <;C4directional cell under 3 x < frequency reuse pattern,
the interference source of each cell is reduced to 2 at the first layer. (n this
case, the theoretical #$( can be expressed by the following equation:
(n actual conditions, because base stations are irregularly distributed, the
antenna height is different, and the effect from radio environment, the value
of #$( cannot be as high as 2=.< d8.
(f the available bandwidth is 2;-&z, the channel numbers range from .= to
:., you can also use 3 x < frequency reused pattern. #onsidering the
characteristics of the 3 x < cellular structures, you can also use the 3 x <
frequency reuse for 8##&. The frequencies are from 92 to :., 2. channel
numbers in total, and the others are T#& numbers.
For t$e !requency )lannin under % : 3 !requency reuse )attern" see Frequency
)lannin under % : 3 !requency reuse )attern
Frequency )lannin under % : 3 !requency reuse )atternFrequency )lannin
under % : 3 !requency reuse )attern
"requency
group
number
A1 B1 A2 B2 A3 B3 A4 B4 A5 B5 A6 B6
#hannel
number of
each
frequency
group
94 93 92 91 90 89 88 87 86 85 84 83
80 79 78 77 76 75 74 73 72 71 70 69
68 67 66 65 64 63 62 61 60 59 58 57
56 55 54 53 52 51 50 49 48 47 46 45

s listed in this table, when allocating frequency to the base station, you can
select the frequency according to the regularity of >2, 3, /, ., ., <?
and >82, 83, 8/, 8., 8=, 8<?. Note that intra4frequency and neighbor
frequency cannot be present within the same cell and ad6acent cells.
7nder the 3 x < frequency reuse pattern, you can enhance the system capacity
by adding new cells to the base station. #ompared with . x / frequency reuse
pattern, the maximum base station type can be configured as ,.$.$.$.$.$.
under 3 x < frequency reuse pattern, so the capacity of a single base station is
twice that of the base station under the . x / frequency reuse pattern.
7nder this frequency reuse pattern, however, the intra4frequency reuse
distance is further shortened, which increases networ) interference greatly. (n
addition, as the number of cells increases, the requirements on the half4power
angle and other antenna indexes are higher. -oreover, you must add antenna
feeders to the system if using the 3 x < frequency reuse pattern, which brings
great difficulty to pro6ect implementation. Therefore, the 3 x < frequency
reuse pattern is seldom used.
"or the 3 x < frequency reuse pattern, the frequency reuse degree is 23.=. nd
its characteristics are listed in the following:
+$rou$ add more cells to eac$ base station" you can
en$ance t$e ca)acity o! t$e base station reatly.
+$e antennas #it$ smaller $al!,)o#er anle and ood
)er!ormance are needed and t$e requirement on antenna and
base station address is strict.
+$e sinals radiated by antennas are more concentrated"
#$ic$ is ood !or indoor coverae.
+$e BSS system must su))ort 3 sectors.
More antennas are needed under t$e % : 3 !requency reuse
)attern t$an t$at under 4 : 2 !requency reuse )attern" so you
must adFust and o)timi(e t$e )lannin !or antenna system and
!requencies.
+$e times o! $andovers under t$e % : 3 !requency reuse
)attern are more t$an t$at under t$e 4 : 2 !requency reuse
)attern.
+$e intra,!requency reuse distance is small" so t$e
inter!erence #it$in t$e net#or? is reat. +$ere!ore" you must
ta?e anti,!requency measures" suc$ as usin D+O and !requency
$o))in.
3./.3 2 ' 3 Fre#uency $euse Pattern
7nder 3 x / frequency reuse pattern, there are 3 base stations. Dach one has /
cells, so < cells form a frequency reuse cluster. The cells in the same cluster
use the different frequencies, and the cells in different clusters use the same
frequency group. This is defined as the 3 x / frequency reuse pattern.
Drror: !eference source not found3 x / frequency reuse pattern:
7nder 3 x / frequency reuse pattern, each intra4frequency cell is interfered by
/ cells. 8ecause the number of base stations in each frequency cluster is 3 0N
= 31, the intra4frequency interference attenuation factor 0q1 can be expressed
by the following equation:
"or regularly4arranged cells, the theoretical carrier4to4interference ratio 0#$(1
can be expressed by the following equation:
Dven if the cells are regularly arranged, however, the value of #$( cannot
meet the requirement of the networ). Therefore, you must ta)e anti4frequency
measures, such as frequency hopping, power control, and *T%.
"or 2;-&z bandwidth, the available channel numbers are from .= to :.. (f
the 2. channel numbers 0924:.1 are 8##& numbers, and the others are T#&
numbers, the frequencies are planned according to 3 x / frequency reuse
pattern.
Frequency )lannin under % : 2 !requency reuse )attern
"requency group
number
A1 B1 A2 B2 A3 B3
#hannel number of
each frequency
group
80 79 78 77 76 75
74 73 72 71 70 69
68 67 66 65 64 63
62 61 60 59 58 57
56 55 54 53 52 51
50 49 48 47 46 45

Eou can use looser . x / frequency reuse pattern and allocate 2. channel
numbers for 8##&. (f the bandwidth is 2;-&z, you can configure the
maximum base station type as ,A$A$A under the 3 x / frequency reuse pattern.
(n this case, the frequency reuse degree is A.2..
The networ) capacity is great under the 3 x / frequency reuse pattern, but
small intra4frequency reuse distance will cause great interference. (n addition,
the cell traffic cannot 2;;F reach the designated value. (n actual conditions,
therefore, you can use the looser . x / frequency reuse pattern for 8##& and
the 3 x / frequency reuse pattern for T#&.
The characteristics of the 3 x / frequency reuse pattern are listed below:
+$e net#or? ca)acity is relatively reat.
+$e adFustment !or t$e net#or? structure is unnecessary.
+$e net#or? ca)acity can be e:)anded #it$out #ide
!requency band.
Small intra,!requency reuse distance #ill cause reat
inter!erence" so you must ta?e anti,inter!erence measures to
ensure net#or? quality.
<adio !requency -<F. $o))in tec$noloy must be used to
su))ort t$e equi)ments.
+$e antennas must be directed to t$e same direction as
muc$ as )ossible.
3./. 1 ' 3 Fre#uency $euse Pattern
2 x / frequency reuse pattern is also called fractional reuse. "or 2 x / or 2 x 2
frequency reuse pattern, the reuse distance is quite small, so the interference
in the networ) is quite great. Therefore, to avoid frequency collision, you
must use !" hopping technology and set the parameters, including -
0mobile allocation1, &,N 0hopping sequence number1, and -(G 0mobile
allocation index offset1. The ratio of number of the T!%s to that of the
frequency hopping is "! HG* 0generally, it is smaller than =;F1.
7nder the 2 x / frequency reuse pattern, the interference in the networ) can
also indicates the probability of the collision of intra4frequencies and
neighbor frequencies. Dmulation shows that probability of the collision is
related to "! only.
ccording to 2 x / frequency reuse pattern, the / cells of a base station form
a frequency reuse cluster. The same4directional cells of each base station use
the same frequency group.
2 x / frequency reuse pattern.
For t$e 1 : 2 !requency reuse )attern" t$e number o! base station is
1 -E N 1." so

" and

.
8ecause the value of #$( here is far lower than the protection value required
by the system, you must ta)e anti4interference measures, such as frequency
hopping, power control, and *T%, to enhance the value of #$(.
(f the available bandwidth is 2;-&z, the available channel numbers are from
.= to :.. 8ecause !" hopping must be used under 2 x / frequency reuse
pattern, considering the importance of 8##&, you can use . x / frequency
reuse pattern for 8##& and 2 x / frequency reuse pattern for T#&.
"or 8##&, 2. channel numbers 0924:.1 are available@ for T#&, /< channel
numbers 0.=49;1 are available.
The channel numbers used for T#& are divided according to two ways. They
are space grouping and sequence grouping. "or the 2 x / frequency reuse
spacing grouping, see 2 % / frequency reuse space grouping
1 O 2 !requency reuse s)ace rou)in -a.
"requency
group number
#hannel number -(G
9;, AA, A., A2, <9, <=, <3, =:, =<, =/, =;, .A ;, 3, .,<, 9, 2;
8 A:, A<, A/, A;, <A, <., <2, =9, ==, =3,.:, .< 2, /, =, A, :, 22
# A9, A=, A3, <:, <<, </, <;, =A, =., =2, .9, .= ;, 3, ., <, 9, 2;

"or the 2 x / frequency reuse sequence grouping, see 2 x / frequency reuse
sequence grouping 0a1
1 : 2 !requency reuse sequence rou)in -a.
"requency
group number
#hannel number -(G
9;, A:, A9, AA, A<, A=, A., A/,A3, A2, A;, <: ;, 3, ., <, 9, 2;
8 <9, <A, <<, <=, <., </, <3, <2, <;, =:, =9, =A ;, 3, ., <, 9, 2;
# =<, ==, =., =/, =3, =2, =;, .:, .9, .A, .<, .= ;, 3, ., <, 9, 2;

8ecause the ratio of the number of carriers to that of frequency hopping is
required to be 2 to 3, if the bandwidth is 2;-&z, you can configure the
maximum base station type as ,A$A$A. (n this case, the frequency reuse degree
is A.2..
The / cells of the same base station use the same &,N, and the cells of
different base stations use different &,Ns. To avoid the interference from
neighbor frequencies, you can configure a proper -(G for the cells of the
same base station.
(f the available bandwidth is <-&z, the available channel numbers are from
:< to 23.. (n this case, you can use . x / frequency reuse pattern for 8##&
0the available channel numbers are from 222 to 23., namely, 2. in total1. "or
T#&, you can use 2 x / frequency reuse pattern 0the available channel
numbers are from :< to 22;, namely, 2= in total.
"or the 2 x / frequency reuse space grouping when the bandwidth is <-&z,
see 2 x / frequency reuse space grouping 0b1
1 : 2 !requency reuse s)ace rou)in -b.
"requency
group number
#hannel number -(G
:<, ::, 2;3, 2;=, 2;9 ;, 3, .
8 :A, 2;;, 2;/, 2;<, 2;: 2, /
# :9, 2;2, 2;., 2;A, 22; ;, 3

Bhen the bandwidth is <-&z, you can configure the maximum base station
type as ,.$/$/ under 2 x / frequency reuse space grouping. (n this case, the
frequency reuse degree is A.3=$:.<A$:.<A, with 9.9< in average.
"or the 2 x / frequency reuse sequence grouping, see 2 x / frequency reuse
space grouping 0b1
1 : 2 !requency sequence rou)in -b.
"requency
group number
#hannel number -(G
:<, :A, :9, ::, 2;; ;, 3
8 2;2, 2;3, 2;/, 2;., 2;= ;, 3
# 2;<, 2;A, 2;9, 2;:, 22; ;, 3

8ecause the ratio of the number of carriers to that of frequency hopping is
required to be 2 to 3, if the bandwidth is <-&z, you can configure the
maximum base station type as ,/$/$/. (n this case, the frequency reuse degree
is :.<A.
"or T#&, both the space grouping and sequence grouping have drawbac)s.
+enerally, for the urban areas where base stations are regularly and densely
distributed, you should better use sequence grouping. "or the areas where
base stations are fragmentary and irregularly distributed, you should better
use space grouping.
The characteristics of 2 x / frequency reuse pattern are listed below:
+$e !requencies are more aressively reused" so t$e
net#or? ca)acity is reat.
+$e net#or? ca)acity under s)ace rou)in is a little
reater t$an t$at under sequence rou)in.
9$en )lannin a net#or?" you need to )lan c$annel
numbers !or BCC' only.
<e,)lannin !or !requencies is unnecessary durin net#or?
o)timi(ation.
+$e e!!iciency !or net#or? )lannin is $i$.
9ideband combiner must be used" but t$e cavity combiner
#it$ !requency selectivity is ina))licable.
+$is !requency reuse )attern requires #ideband re)eater.
+$e inter!erence amon intra,!requencies and nei$bor
!requencies increases as t$e !requency reuse distance decreases.
<F $o))in must be used" and t$e c$annel numbers
)artici)atin !requency $o))in is t#ice t$at o! t$e number o!
carriers at least.
>n actual conditions" you cannot ta?e anti,inter!erence
measures" suc$ as <F $o))in" D+O" and )o#er control" !or
BCC'. +$ere!ore" to ensure net#or? quality" you can use t$e
looser 4 : 2 !requency reuse )attern !or BCC' only.
3././ 1 ' 1 Fre#uency $euse Pattern
Gne cell of one base station forms a frequency reuse cluster, and this is
defined 2 x 3 frequency reuse pattern. Gther cells and this cell use the same
frequency group.
(f the available bandwidth is <-&z, the available channel numbers are from
:< to 23.. 8ecause !" hopping must be used under 2 x 2 frequency reuse
pattern, considering the importance of 8##&, you can use . x / frequency
reuse pattern for 8##& and 2 x 2 frequency reuse pattern for T#&.
(f . x / frequency reuse pattern is used for 8##&, the available channel
numbers are from 222 to 23., 2. in total. The channel numbers from :< to
22; are used for T#&, 2= in total.
"or the frequency planning under 2 x 2 frequency reuse pattern, see
"requency planning under 2 x 2 frequency reuse pattern.
Frequency )lannin under 1 : 1 !requency reuse )attern.
"requency
group number
#hannel number -(G
:<,:A,:9,::,2;;,2;2,2;3,2;/,2;.,2;=,2;<,2;A,2;9,2;:,22
;
;,3,.
8 :<,:A,:9,::,2;;,2;2,2;3,2;/,2;.,2;=,2;<,2;A,2;9,2;:,22
;
<,9
# :<,:A,:9,::,2;;,2;2,2;3,2;/,2;.,2;=,2;<,2;A,2;9,2;:,22
;
2;,23

(f the bandwidth is <-&z, you can configure the maximum base station type
as ,.$/$/$ under 2 x 2 frequency reuse pattern. (n this case, the frequency
reuse degree is A.3=$:.<A$:.<A, so the average value is 9.9<.
Therefore, the maximum base station configuration under 2 x 2 frequency
reuse pattern is the same as that under 2 x / frequency reuse space grouping
pattern, so is the networ) capacity.
3./.. A 1 + Fre#uency $euse Pattern
The I 8 frequency reuse pattern is developed from 2 x / frequency reuse
pattern. Bhen the bandwidth is narrow but the capacity is great, you can use
this frequency reuse pattern. (n this case, you must use !" hopping. 7nder
the I 8 frequency reuse pattern, the frequencies can be divided into three
groups. They are >f2?, >f3?, and >f/?. "or frequency planning, see I 8
frequency reuse pattern
A J B !requency reuse )attern
ccording to I 8 frequency reuse pattern, you can increase frequency
diversity gain by increasing the number of channel numbers participating
frequency hopping within the cell, because the increase of the frequency
diversity gain can improve the carrier4to4interference ratio. To avoid
interference among intra4frequencies and neighbor frequencies, you can
configure a proper -(G for the cells within the same base station. The
probability of the collision of the intra4frequencies and neighbor frequencies
will decrease as the number of channel numbers participating frequency
hopping increases among cells of different base stations.
(f the available bandwidth is <-&z, the available channel numbers are :< to
23.. "or I 8 frequency reuse pattern, you must use !" hopping, but the
8##& does not participate in !" hopping. Therefore, in actual planning, to
ensure good networ) quality, you can use looser . x / frequency reuse pattern
for 8##& and I 8 frequency reuse pattern for T#&.
(f you use . x / frequency reuse for 8##&, the available channel numbers
are 222 to 23., 2. in total, in which two channel numbers are standby ones.
"or T#&, the available channel numbers are :< to 22;, 2= in total.
For t$e !requency )lannin under A J B !requency reuse )attern" see Frequency
)lannin under A J B !requency reuse )attern
Frequency )lannin under A J B !requency reuse )attern
"requency
group number
#hannel number -(G
:<, :A, :9, ::, 2;;, 2;2, 2;3, 2;/, 2;., 2;= ;, 3, .
8 2;2, 2;3, 2;/, 2;., 2;=, 2;<, 2;A, 2;9, 2;:, 22; 2, /
# :<, :A, :9, ::, 2;;, 2;<, 2;A, 2;9, 2;:, 22; =, A

Bhen the bandwidth is 2;-&z, you can configure the maximum base station
type as ,.$/$/ under I 8 frequency reuse pattern. (n this case, the
frequency reuse degree is A.3=$:.<A$:.<A, so the average value is 9.9<.
(n actual conditions, the irregular distribution of base stations and antenna
height may deteriorate the performance of parts of the networ). Therefore,
the I 8 frequency reuse pattern are not recommended in large networ)s.
3.. Concentric Cell Technology
3...1 Concept
(n the +,- networ), concentric cell technology is used to divide the service
area into two parts: overlay and underlay. (n essence, the concentric cell
technology concerns channel allocation and handover. Bhen combining this
technology with various frequency planning technologies, you can both
expand networ) capacity and improve networ) quality.
The underlay covers the traditional cells, and the overlay covers the areas
near the base station. +enerally, . x / frequency reuse pattern is used for the
underlay. "or overlay, the frequency reuse patterns, such as / x /, 3 x /, or 2
x /, are used. Therefore, all carriers can be divided into two groups, one for
underlay, and the other one for overlay. The overlay and underlay share the
same base station address, one set of antenna feeder system, and one 8##&,
so you must set the 8##& on the underlay.
>! t$e ca)acity o! t$e overlay is reat" you can rou) t$e c$annel numbers
accordin to C$annel number rou)in !or 3M'( band#idt$ concentric cell -a..
>n t$is case" t$e overlay $as more c$annel numbers" #$ic$ is bene!icial !or t$e
base station to absorb nearby tra!!ic volume.
C$annel number rou)in !or 3M'( band#idt$ concentric cell -a.



Hogical
channel
#hannel number
7nderla
y
0231
<< <A <9 <: A; A2 A3 A/ A. A= A< AA
Gverlay
0291
A9 A: 9; 92 93 9/ 9. 9= 9< 9A 99 :9 :; :2 :3 :/ :. :=




>! tra!!ic volume is evenly distributed" you can en$ance t$e underlay ca)acity
t$rou$ rou)in t$e c$annel numbers accordin to C$annel number rou)in
!or 3M'( band#idt$ concentric cell -b..
>n t$is case" t$e underlay can absorb more tra!!ic volume.
C$annel number rou)in !or 3M'( band#idt$ concentric cell -b.




3...2 General 2nderlay 30erlay
+eneral underlay overlay 0+7G1 aims to restrict the intra4frequency
interference. To realize this purpose, you can reduce the overlay coverage
area. That is, if the transmit power of the overlay carriers is lower than that of
the underlay carriers, the coverage area of the overlay is smaller than that of
the underlay.
The handover between the overlay and underlay is related to the receiving
level of the -, and the T 0timing advance1 from the -, to the base station.
Eou should allocate the channel numbers 0such as 8##& number1 with
looser frequency reuse aggressiveness to the -,s in the underlay. "or the
-,s in the overlay, you should allocate the channel numbers with aggressive
frequency reuse to them. (n this case, you can expand the networ) capacity by
using aggressive frequency reuse pattern in overlay.
"or general underlay overlay, the coverage area of the underlay is
inconsistent with that of the overlay, so problems concerning traffic and
handover control are often caused. The general underlay overlay is applicable
to the areas near the base station where the traffic is concentrated. The more
concentrated the traffic near the base station, the more apparent the effect of
capacity expansion is. &owever, the transmit power of the carriers in the
overlay is low, so it is hard for the base station to absorb indoor traffic
Hogical
channel
#hannel number
7nderlay
03.1
<
<
<
A
<
9
<
:
A
;
A
2
A
3
A
/
A
.
A
=
A
<
A
A
A
9
A
:
9
;
9
2
9
3
9
/
9
.
9
=
9
<
9
A
9
9
9
:

Gverlay
0< 1

:
;
:
2
:
3
:
/
:
.
:
=
volume. (n this case, when the traffic volume is evenly distributed, the
general underlay overlay has little effect on capacity expansion.
3...3 &ntelligent 2nderlay 30erlay
>ntellient underlay overlay ->/A. tec$noloy can ensure t$at t$e coverae
areas o! call carriers are t$e same. For an >/A" t$e transmit )o#er o! t$e
carriers in t$e underlay and overlay is t$e same.
>n an >/A" t$e !requencies o! a base station are divided into t#o layers* one is
reular layer" and t$e ot$er one is su))er layer. At t$e reular layer" t$e
!requency reuse distance is lare" so you can use looser !requency reuse
)attern" suc$ as 4 : 2 !requency reuse )attern. At t$e su))er layer" t$e
!requency reuse distance is relatively small" so you can use aressive
!requency reuse )atterns" suc$ as % : 2 and 1 : 2 !requency reuse )attern.
(n an (7G, the interference at the supper layer is great, so designated
equipments and handover algorithms on #$( must be provided.
(n an (7G, the conversation is first established at the supper layer, and then
the 8,# monitors the #$( of the channels at the supper layer without any stop.
(f the #$( is greater than the +ood #$( Threshold, the conversation seizes a
channel at the supper layer. (f the #$( is smaller than the 8ad #$( Threshold,
the conversation seizes a channel at the regular layer. (n addition, you can
control the traffic volume at the supper layer and the regular layer by
ad6usting the handover threshold.
"or an (7G, the transmit power of the carriers at the regular layer is the same
as that at the supper layer, so the networ) can absorb the traffic flexibly,
which is beneficial for the expansion for actual networ) capacity.
(f the (7G technology is used, you must add the functions, including the
estimation of intra4frequency protection #$( for downlin) channels and the
handover algorithms related to (7G, to the system.
3... Characteristics o4 Concentric Cell Technology
The characteristics of concentric cell technology are listed below:
Any c$ane o! t$e net#or? structure is unnecessary.
S)ecial so!t#are and desinated alorit$ms on c$annel
allocation and $andover are needed.
+$e system $as no s)ecial requirement on $ard#are.
G/A is a))licable to t$e areas near t$e base station #$ere
t$e tra!!ic is concentrated.
+$e overlay coverae o! t$e G/A is small" so t$e intra,
!requency reuse attenuation !actor -q. is reat" #$ic$ increases
inter!erence in t$e net#or?.
+$e transmit )o#er o! t$e overlay carriers in t$e G/A is
lo#" so it is $ard !or t$e carriers to absorb indoor tra!!ic.
+$e transmit )o#er o! t$e underlay carriers in t$e G/A is
t$e same" so t$e carriers can absorb indoor tra!!ic" #$ic$
contributes to net#or? ca)acity e:)ansion and ood conversation
quality.
For t$e com)arison bet#een t$e G/A and >/A" see A com)arison bet#een G/A
and >/A.
A com)arison bet#een G/A and >/A. A com)arison bet#een G/A and >/A.

#overage
area
"requency reuse
pattern
Transmit
power
Hogical channel
allocation
&andover
algorithm
+7G
7nderlay . x / &igh 8##&$T#&
JowerK
*istance
Gverlay / x /$3 x /$2 x / How T#&
(7G
7nderlay . x / ,ame 8##&$T#& #$(
Gverlay / x /$3 x /$2 x / ,ame T#&

3.7 Multiple $euse Pattern Technology
3.7.1 +asic Principle
ccording to multiple reuse pattern 0-!J1, the carriers are divided into
several groups. The carries in each group wor) as an independent layer, and
each layer uses a different frequency reuse pattern. *uring frequency
planning, you can configure the carriers layer by layer, with reuse
aggressiveness increases layer by layer.
M<@ $as no s)ecial requirement on $ard#are. >t is develo)ed !rom t$e conce)t
o! carrier layerin. +$at is" t$e available c$annel numbers are divided into
multi)le rou)s" and eac$ rou) #or?s as a carrier layer. Accordin to t$e rules
o! t$e aressive !requency reuse )attern" t$e c$annel numbers allocated !or
eac$ layer are listed in C$annel number allocation !or eac$ layer
C$annel number allocation !or eac$ layer
Hayer #hannel number
8##& n2
T#& 2 n3
T#& 3 n/
L L
T#&m42 nm
Note:
n1 n2 n3 n4 nm.
"or -!J, first you must divide an available band into several sub4bands.
+enerally, the sub4bands wor) as the bands for 8##&. The reasons are listed
below:
BS>C decodin #ill not be a!!ected by tra!!ic. +C' numbers
cannot a!!ect se)arated BCC' numbers" #$ic$ is $el)!ul !or t$e
MS to decode t$e BS>C.
+$e )lannin !or adFacent cell list can be sim)li!ied. +$e
se)arated BCC' numbers contributes t$e sim)li!ication o!
adFacent cell list" so t$e MS can ca)ture t$e use!ul BCC' quic?ly.
Ma:imum ain can be obtained !rom )o#er control and D+O.
Do#nlin? )o#er control and D+O can be a))lied to +C' carriers
only" so t$e se)arated BCC' numbers can ma:imi(e t$e !unction
o! do#nlin? )o#er control and D+O.
+$e re,)lannin !or +C' numbers #ill not a!!ect BCC'. 9$en a
+<O is added to t$e system" i! not considerin t$e isolation o!
combiner and adFacent !requency inter!erence" you do not $ave
to c$ane t$e BCC' numbers.
fter that, you must divide the remaining channel numbers into multiple T#&
bands. "or -!J, different frequency reuse patterns must be used for different
T#& bands.
ccording to the carrier allocation in the networ), you can decide the average
frequency reuse degree. ccording to the maximum number of carriers
configured in each cell and the number of cells configured in the networ),
you can ad6ust the average frequency reuse degree to a proper value. (n this
way, you can effectively control networ) quality.
The increase of the carries has little effect on the frequency allocation plan.
The increased channel numbers affect other cells that have more carriers than
the service cell has. "or example, if a cell has four carriers, the cells that have
been configured with more than four cells will be affected.
-!J technology enables carriers to be configured flexibly. ccording to
-!J, the frequencies of a cell can never be completely identical with that of
the ad6acent cells. Therefore, the -!J improves both the intra4frequency
interference protection ratio and frequency hopping effect.
ccording to the requirements defined in +,- protocols, all the downlin)
timeslots of the 8##& carriers must transit with full power and the
interference features of the 8##& are different from that of the T#&.
Therefore, to ensure networ) quality and security, you are recommended to
use . x / frequency reuse pattern for 8##&. (n this case, the channel
numbers used for 8##& are equal to or more than 23. (n actual conditions,
they are from 23 to 2=.
(f the available bandwidth is A.3-&z, the available channel numbers are from
<; to :=, /< in total, and they can be divided into . groups
+o ensure net#or? security" you must !inis$ BCC' number allocation
!irst. +o be s)eci!ic" )lan t$e 1% c$annel numbers accordin to 4 : 2
!requency reuse )attern and allocate 1 BCC' number to eac$ o! t$e
1% cells. A!ter t$at" you s$ould allocate 1 carrier at t$e +C'2 layer
to eac$ cell" and t$en you s$ould allocate t$e +C'% and +C'1
numbers to t$e cells. >n t$is case" you can con!iure !our c$annel
numbers !or eac$ cell o! a base station -S4=4=4.. +$e remainin 2
c$annel numbers can be con!iured !or micro cells or mini,micro
cells.
3.7.2 M$P Se#uence Grouping
8ecause 8##& numbers and T#& numbers are selected in different ways,
the -!J can be divided into two types. They are -!J sequence grouping
and -!J space grouping, the first of which is introduced hereunder.
(f the available bandwidth is 2;-&z, the channel numbers are from .< to :..
(n this case, you can plan the frequencies at the 8##& and T#& carrier
layers according to the sequence of the channel numbers. (f using the
sequence planning, you should add 2 to 3 extra channel numbers to the
8##& numbers. "or the -!J sequence grouping, see:
M<@ sequence rou)in
#arrier type
!"#N of the available
channel number
vailable channel numbers
8##& 9/M:. 23
T#&2 A.M93 :
T#&3 <<MA/ 9
T#&/ =9M<= 9
T#&. =3M=A <
T#&= .<M=2 <
Note:
ARFCN stands for absolute radio frequency channel number.
ccording to this table, the channel numbers can be divided into < groups.
"or 8##&, 23 channel numbers can be reused at the carrier layer. Traffic
channels can be divided into = groups, from T#&2 to T#&=. "or T#&2, :
channel numbers can be reused@ for T#&3 and T#&/, 9 channel numbers can
be reused@ and for T#&. and T#&=, < channel numbers can be reused.
Therefore, when the bandwidth is 2;-&z, the base station type can be
configured as ,<$<$<. (f the traditional .$23 frequency reuse pattern is used,
the maximum base station type can be configured as ,.$.$. only.
"or -!J sequence grouping, intra4frequency and neighbor frequency
interference may exist within the frequency layer, and the interference
between frequency layers exist at the critical points of the frequencies.
3.7.3 M$P Space Grouping
"or -!J space grouping, neighbor frequency interference does not exist
within the frequency layer, but exist between frequency layers. Bhen the
traffic is not busy, this frequency reuse pattern can reduce networ)
interference.
>! t$e available band#idt$ is 10M'(" t$e available c$annel numbers
are !rom 43 to 94. >n t$is case" t$e !requencies can be allocated
accordin to
#arrier
type
!"#N of the available channel number
vailable channel
numbers
8##& .<, .9, =;, =3, =., =<, =9, <;, <3, <., <<, <9 23
T#&2 A;, A3, A., A<, A9, 9;, 93, 9., 9< :
T#&3 99, :;, :3, :., .A, .:, =2, =/ 9
T#&/ ==, =A, =:, <2, </, <=, <A, <: 9
T#&. A2, A/, A=, AA, A:, 92 <
T#&= 9/, 9=, 9A, 9:, :2, :/ <
Note:
ARFCN stands for absolute radio frequency channel number.
t the very beginning, not each cell needs the T!% of the last layer, so the
T!% of the last layer can reuse the frequencies more aggressively. (n
addition, though interference increases after the -!J is enabled, the T!%s in
the cells also increase. (n this case, more the channel numbers will participate
in frequency, which enhances frequency hopping gain.
(f both the channel numbers with a little interference and the channel numbers
with great interference exist simultaneously within a cell, the frequency
hopping technology will average the interference through mixing these
channel numbers. (n this case, the system can still decode the signals
normally.
Bhen allocating the frequencies according to -!J, you must notice that the
minimum frequency reuse degree at the T#& layer must be equal to or
greater than <. (n actual conditions, however, the minimum average
frequency reuse degree at the T#& layer ranges from A to 9. Therefore, when
the frequency resource is adequate, you can reserve some channel numbers to
for future use during frequency planning.
"ixed -!J means that the channel numbers allocated to each T#& are fixed.
They are independent of each other, as shown in Drror: !eference source not
found. "or -!J, you should plan the channel numbers layer by layer so that
the T#& numbers can be easily ad6usted. (n this case, if interference is
present at a T#& layer, you need to ad6ust the channel numbers allocated to
that layer only.
3.7.4 Characteristics o4 M$P Technology
-!J technology can enables you to plan the frequencies flexibly according
to traffic distribution. #ompared with / x / frequency reuse pattern, -!J
contributes to greater networ) capacity. #ompared with 3 x / and 2 x /
frequency reuse pattern, -!J has little effect against networ) quality. (n
addition, -!J technology is compatible with the technologies, such as
frequency hopping, power control, *T%. -oreover, it has no special
requirement on hardware and software.
+enerally, the advantages of the -!J are listed below:
+$e net#or? ca)acity is reat and !requency utili(ation rate is
$i$.
+$e c$annel con!iuration is !le:ible. +$e !requency reuse
)attern is selected accordin to net#or? ca)acity and tra!!ic
distribution. >n t$e areas #$ere t$e tra!!ic is $i$" you can add
carriers to t$ese areas.
Eo t#o cells $ave t$e same c$annel numbers" so no intra,
!requency cell e:ists in t$e system i! t$e M<@ is used.
Baseband $o))in and <F $o))in can be used.
+$e base station ty)e can be con!iures !le:ibly" #$ic$ is ood
!or net#or? quality.
+$e c$annels to be allocated are #ei$ted" #$ic$ en$ances
t$e net#or? quality.
3.7.5 Co"parison 5etween M$P and 1 6 3 Fre#uency $euse Pattern
(n fact, 2 x / frequency reuse pattern is a special )ind of -!J. The
configuration for the equivalent -!J is 23$/$/$/$/$/. The following is a
comparison between -!J and 2 x / frequency reuse pattern.
+$e net#or? ca)acity under 1 : 2 !requency reuse )attern is
reater t$an t$at under M<@.
For 1 : 2 !requency reuse )attern" you need to )lan a rou) o!
!requencies !or +C' only. >! you $ave to add ne# carriers to t$e
system #it$out addin ne# base stations" you do not $ave to re,
)lan t$e !requencies. +$ere!ore" t$e !requency )lannin is
sim)ler under 1 : 2 !requency reuse )attern t$an t$at under
M<@.
>! t$e net#or? is irreular in land!orms and tra!!ic
distribution" you s$ould better not use 1 : 2 !requency reuse
)attern. >n most cases" a base station is inter!ered by many base
stations nearby. >! t$e 1 : 2 !requency reuse )attern is used" you
#ill !ind it $ard to )osition t$e inter!erence source. +$ere!ore"
#$en addin ne# base stations to t$e net#or?" you cannot
eliminate t$e inter!erence by adFustin some c$annel numbers
only. >! usin M<@" $o#ever" you can easily solve t$is )roblem.
3.7 !etwor8 Capacity Co"parison
"or the comparison of the networ) capacity under various frequency reuse
patterns, see:
Com)arison o! t$e net#or? ca)acity under various !requency reuse )attern
Bandwidth
Frequency reuse
pattern
Frequency
reuse
degree
Base
station
configurati
on type
Loadabl
e traffic
volume
Admissible
subscribers
Capacit
y ratio
<-&z
.N/ 23 /$3$3 3A.: 2299 2
/N/ : /$/$/ /..= 2/9; 2.2<
.N/ I 2N/ A.= .$.$/ =/.= 32.; 2.9
-!J023, :, <1 : /$/$/ /..= 2/9; 2.2<
3 N < 23 3$3$3$3$3$3 .:.3 2:<9 2.<<
(7G: . N / I 3 N / : .$.$/ =/.= 32.; 2.9
A.3-&z
. N / 23 /$/$/ /..= 2/9; 2
/ N / : .$.$. <3 3.9; 2.9
. N / I 2 N / A.= =$=$= 92.: /3A< 3./A
-!J023, :, 9, A1 : .$.$. <3 3.9; 2.9
3 N < 23 /$/$/$3$3$3 <;.2 3.;. 2.A.
(7G: . N / I 3 N / : =$=$= 92.: /3A< 3./A
:.<-&z
. N / 23 .$.$. <3 3.9; 2
/ N/ : =$=$= 92.: /3A< 2./3
. N / I 2 N / A.= A$A$A 23/.< .:.. 2.::
-!J023,:,9,A,<,<1 9 <$<$< 2;..2 .2<. 2.A;
3 N < 23 .$.$.$.$.$. 23< =;.; 3.;/
(7G: . N / I 3 N / : A$A$A 23/.< .:.. 2.::
Note:
GoS = 0.02; a = 0.02 !rl.
4 GSM Parameter
Configuration and
Adjustment
Bhen operators prepare to construct a mobile communication networ), they
must predict coverage according to traffic prediction and local radio
propagation environment. This guides pro6ect design of the system and
parameter configuration of radio networ).
The pro6ect design includes the following aspects:
Eet#or? to)oloy desin
Selectin t$e location o! base station
Frequency )lannin
Cell )arameter con!iuration
The !" planning determines the coverage range of a cell, and the serving
range of the cell is determined based on the combination of !" planning and
cell parameter configuration. 8y this, the -, always en6oys optimal services
and maximum networ) capacity at the best cell.
This chapter discusses the meaning and effect of important parameters in
+,- radio communication. -astering the effect and impact of these
parameters helps to configure networ) parameters and optimize the networ)
in later stages.
(n a +,- networ), abundant radio parameters are configured according to
cells or partial areas@ however, the parameter configuration might affect
neighbor areas. Therefore, while configuring and ad6usting parameters, you
must pay attention to the impact of configuring parameters on other areas,
especially neighbor areas.
4.1 !etwor8 and Cell &9
4.1.1 Cell Glo5al &9
I. 9e4inition
+,- is a global cellular mobile communication system. To ensure that each
cell corresponds to a unique (* globally, the +,- system numbers the
following items:
;ac$ GSM net#or? in eac$ country
;ac$ location area
;ac$ base station
;ac$ cell
Numbering the previous items aims as follows:
An MS can identi!y t$e servin net#or? so t$at t$e MS can
select a net#or? in any environment.
+$e net#or? can obtain t$e )recise location o! t$e MS so t$at
t$e net#or? can )rocess various service requests involvin t$e
MS.
+$e MS can re)ort in!ormation about nei$bor cells to t$e
net#or? durin callin to avoid call dro).
The cell global identity 0#+(1 is a ma6or networ) identity parameter. #+(
consists of location area identity 0H(1 and cell identity 0#(1. H( includes
mobile country code 0-##1, mobile networ) code 0-N#1, and location area
code 0H#1.
The system transmits #+( information through system information 0,(1
transmitted by cell broadcast. Bhen an -, receives ,(, it demodulates ,( for
#+( information. The -, 6udge whether to camp on the cell according to the
-## and -N#. (t also 6udges whether the current location area changes to
determine updating location. Bhile updating location, the -, reports H(
information to the networ) so that the networ) can )now the location area of
the -,.
II. For"at
The #+( is -##4-N#4H#4#(, with details as follows:
MCC consists o! t$ree decimal diits" ranin !rom 000 to 999.
MEC consists o! t#o decimal diits" ranin !rom 00 to 99.
GAC ranes !rom 0 to 31121
C> ranes !rom 0 to 31121.
III. Con4iguration and &n4luence
s a globally unique mobile identity, the -## is uniformly distributed and
managed by international telecommunication union 0(T71. The -## for
#hina is .<; 0decimal1.
The -N# is uniformly distributed by state telecommunication management
organs. Now two +,- networ)s exist in #hina. The -N# for #hina -obile
is ;;. The -N# for #hina 7nicom is ;2.
The method for coding H# is ruled by each country accordingly. This caters
for #hina also 0refer to +,- system from -inistry of (nformation (ndustry1.
t the early stage of networ) construction, the H# is coded and distributed.
The H# is seldom changed in the later stages.
The coverage areas related to the H# is vital in the networ). Eou can
configure it as great as possible.
No special restriction is on the distribution of #(. The #( ranges from ; to
<==/= 0decimal1. (t must be ensured that two equivalent #(s exist in the same
location area. This is determined in the system design. Dxcept for special
situations 0such as constructing base stations1, the #( must not be changed
during the system operation.
IV. Precautions
Eou must pay attention to the following aspects:
+$e MEC is unc$aneable.
9$ile con!iurin t$e GAC" you must !ollo# related
reulations. ;quivalent GACs must not e:ist in t$e state net#or?.
;quivalent C>s must not e:ist in t$e same location area.
4.1.2 +ase Station &dentity Code
I. 9e4inition
(n a +,- networ), each base station corresponds to a distributed local color
code, called base station identity code 08,(#1. Bhen the -, receives
broadcast control channel 08##&1 carriers of two cells at the same time, with
same channel number, the -, distinguishes them by 8,(#.
(n networ) planning, the 8##& carriers of neighbor cells are different in
frequency to reduce intra4frequency interference. The cellular communication
system features that the 8##& carrier might be reused. Therefore, the 8,(#
of the cells with the same 8##& carrier must be different.
+$e system transmits BS>C on sync$roni(ation c$annel -SC'. o! eac$ cell. +$e
e!!ect o! BS>C is as !ollo#s*
+$e BS>C involves in decodin )rocess o! random access
c$annel -<AC'. to )revent base stations !rom connectin to t$e
<AC' sent to t$e nei$bor cells by t$e MS by error.
A!ter t$e MS receives SC' messaes" it Fudes t$at it $as been
sync$ronous to t$e cell. Decodin in!ormation on t$e do#nlin?
common sinalin c$annel correctly requires trainin sequence
code -+SC. used on common sinalin c$annel.
GSM reulations describe +SC in ei$t !i:ed !ormats" and t$e
sequence number o! t$em is 005. +$e cell BCC determines t$e
+SC used by t$e common sinalin c$annel o! a cell. +$ere!ore
t$e BS>C $el)s in!orm t$e MS o! t$e +SC used by t$e common
sinalin c$annel o! t$e servin cell.
>n a call" t$e MS must measure t$e level o! BCC' carrier o!
nei$bor cells and re)ort it to t$e base station accordin to
reulations to nei$bor cell list o! BCC'. Mean#$ile" t$e MS must
)rovide measured BS>C o! t$e carrier in t$e u)lin? measurement
re)orts. 9$en t$e nei$bor cells o! a cell include t#o or more
cells #it$ t$e same BCC' carrier" t$e base station can
distinuis$ t$e cells by BS>C to avoid incorrect $andover.
>n a call" t$e MS must measure sinals o! nei$bor cells" and
sends measurement re)orts to t$e net#or?. +$e measurement
re)ort can contain in!ormation about si: nei$bor cells only" so
t$e MS must be controlled to re)ort t$e cells actually related to
$andover. +$e !irst t$ree diits o! BS>C -namely" ECC. aims as
)reviously mentioned. A)erators control t$e MS to re)ort t$e
nei$bor cell in!ormation )ermitted by t$e servin cell ECC by
broadcast )arameters !CC per"itted.
II. For"at
The 8,(# is N##48##, with details as follows:
+$e ECC ranes !rom 0 to 5.
+$e BCC ranes !rom 0 to 5.
III. Con4iguration and &n4luence
7sually different +,- JH-Ns use the same frequency resource, but, to
some degree, their networ) planning is independent. The neighbor +,-
JH-Ns use different N##s according to regulations. This ensures that the
neighbor base stations with same frequency use different 8,(#s.
The 8## is part of the 8,(#. (t helps identify different base stations with
same 8##& carrier number in the same +,- JH-N. The values of 8##
must meet the previous requirements. ccording to +,- regulations, the
T,# of cell 8##& carrier must be same as that of cell 8##. The equipment
providers must ensure the T,# consistency.
IV. Precautions
The neighbor cells or cells nearby using the same 8##& carrier must use
different 8,(#s. Dspecially when two or more cells use the same 8##&
carrier in the neighbor cell list of a cell, theses cells must use different 8,(#.
Jay attention to cells at the bordering areas between provinces and cities, and
otherwise cross4cell handover might fail and abundant mista)en access
problems might occur.
.2 Paging and Access Control Para"eters
.2.1 !u"5er o4 Access Grant $eser0ed +loc8s :+S;AG;+L<;$=S or AG>
&. 9e4inition
The common control channel consists of access grant channel 0+#&1 and
paging channel 0J#&1.
"or different ###&s, each 8##& multiframe 0including =2 frames1 contains
###& message bloc)s different number. The ###& is shared by +#& and
J#&. ccording g to regulations, partial message bloc)s on ###& are
especially reserved for +#&. This avoids that the +#& messages are
bloc)ed when the J#& traffic is great.
The number of parameter access grant reserved bloc)s 0+1 refers to the
number of message bloc)s reserved for +#& on ###& in each 8##&
multiframe.
&&. For"at
The + ranges from ; to 3 when ###& shares physical channel
0###&O#GN" = 21 with stand4alone dedicated control channel 0,*##&1.
The + ranges from ; to = when ###& does not share physical channel
0###&O#GN"=;1 with stand4alone dedicated control channel 0,*##&1.
&&&. Con4iguration and &n4luence
Bhen the channel combination of the cell is fixed, the parameter + ad6usts
the ratio of +#& and J#& in ###&. Bhen the J#& is idle, it can send
immediate assignment messages. The +#& does not transmit paging
messages. Dquipment operators can balance +#& and J#& by ad6usting
+, with the following principles.
The principle for + value is that based on no overload of +#&, you must
reduce the parameter to shorten the time for -, to respond to paging, and to
improve system service performance. Bhen the immediate assignment
messages are superior to paging messages to be sent, configure + to ;.
The value of + is recommended as follows:
AG is 1 #$en t$e CCC' and SDCC' s$are a )$ysical c$annel.
AG is % or 2 in ot$er situations.
(n networ) operation, ta)e statistics of overload situations of +#& and
ad6ust + accordingly. 8y default the immediate assignment messages are
superior to paging messages to be sent in the networ), so you need not
reserve a channel for immediate assignment messages. (n this situation,
configure + to ;.
.2.2 Fra"e !u"5er Coding +etween &dentical Paging
"rame number coding between identical paging is 8,OJO-"!-, 0-"!
for short1.
&. 9e4inition
ccording to +,- regulations, each -, 0corresponding to an (-,(1 belongs
to a paging group 0for calculation of paging groups, see +,- regulation
;=.;31. Dach paging group in a cell corresponds to a paging subchannel.
ccording to its (-,(, the -, calculates the paging group that it belongs to,
and then calculates the location of paging subchannel that belongs to the
paging group. The -, only receives the signals of the paging subchannel that
it belongs to, and neglects that of other paging subchannels. (n addition, the
-, even powers off some hardware of itself during other paging subchannel
to lower power cost of itself.
The number of paging channel multiframe 0-"!1 is the number of
multiframes used in a period of paging subchannel. The -"! determines the
number of paging subchannels that the cell J#& is divided into.
&&. For"at
The -"! ranges from 3 to :, which respectively means that the same paging
group cycles in a period of 3 to : multiframes.
&&&. Con4iguration and &n4luence
ccording to the definition of ###&, +, and -"T, you can calculate the
number of paging channel in each cell.
9$en t$e CCC' and SDCC' s$are a )$ysical c$annel" t$ere
is -2 , AG. MF<s.
9$en t$e CCC' and SDCC' s$are a )$ysical c$annel" t$ere
is -9 , AG. MF<s.
ccording to the previous analysis, the greater the -"! is, the more the
paging channels of the cell are 0see the calculation of paging groups in +,-
regulation ;=.;31. Theoretically, the capacity of paging channels does not
increase with the increase of -"!. The number of buffers for buffering
paging messages on each base transceiver station 08T,1 increases. The
paging messages are sent more evenly both in time and space, so it seldom
occurs that the paging messages overflow in the buffers so call lost occurs
0related to functions by equipment providers1.
&owever, to en6oy the previous advantages, you will have a longer delay of
paging messages on the radio channels. The greater the -"! is, the greater
the delay of paging messages in the space is, and the lower the average
service performance of the system is. Therefore, the -"! is an important
parameter in networ) optimization.
The following principle caters for configuring -"!:
The configured strategy for buffers of each equipment provider is different,
so you must select the -"! properly so that the paging messages do not
overflow on J#&. 8ased on this, configure the parameter as small as
possible. (n addition, you must measurement the overflow situations of J#&
periodically while the networ) is running, and ad6ust -"! accordingly.
&?. Precautions
ny paging message of the same location area must be sent to all cells in the
location areas at the same time, so the J#& capacity of each cell in the
location area must be equivalent or close to each other. Gtherwise, you must
consider smaller J#& capacity as the evidence for designing location area.
.2.3 Co""on Control Channel Con4iguration :CCC)@C3!F>
&. 9e4inition
The ###& includes +#& and J#&. (t sends immediate assignment
messages and paging messages. (n each cell, all traffic channels 0T#&s1 share
###&. ccording to the T#& configuration and traffic model of the cell, the
###& can be one or more physical channels. (n addition, the ###& and
,*##& share a physical channel. The combination methods for ##& are
determined by ###& parameter ###&O#GN".
&&. For"at
+$e CCC'PCAEF consists o! t$ree bits" #it$ t$e codin met$ods listed in CCC'
con!iuration codin
CCC' con!iuration codin
###&O#GN" -eaning
Number of ###& message bloc)s in
a 8##& multiframe
;;;
Gne physical channel for used
for ###&, not shared with
,*##&
:
;;2
Gne physical channel for used
for ###&, shared with ,*##&
/
;2;
Two physical channels for used
for ###&, not shared with
,*##&
29
2;;
Three physical channels for
used for ###&, not shared with
,*##&
3A
22;
"our physical channels for used
for ###&, not shared with
,*##&
/<
&&&. Con4iguration and &n4luence
Bhen the ###& and ,*##& share one physical channel, the ###& has the
minimum channel capacity. Bhen the ###& and ,*##& do not share a
physical channel, the more physical channels that the ###& uses, the greater
the capacity is.
The ###&O#GN" is determined by the operators based on combination of
cell traffic model and paging capacity of the location area where a cell
belongs to. (t is determined in system design, and ad6usted in networ)
expansion. ccording to experiences, when the paging capacity in the
location area is not high and cell has one or two carriers, it is recommended
that the ###& uses one physical channel and share it with ,*##& 0in
combination ###& methods1. This spares a physical channel for paging.
Gtherwise, the method that ###& and ,*##& do not share one physical
channel is used.
Bhen the cell T!% exceeds < and ###& GPD!HG* occurs in the cell, it
is recommended that the ###& uses two or more basic physical channel and
does not share them with ,*##&.
&?. Precautions
The ###&O#GN" must be consistent with the actual configuration of cell
###&. (n addition, you must consider the influence on the access grant
reserved bloc)s.
.2. ='tended Trans"ission Slots :T6;&!T=G=$>
&. 9e4inition
(n a +,- networ), a random access channel 0!#&1 is an HG&. To reduce
the conflicting times on !#& when an -, accesses the networ), and to
increase !#& efficiency, +,- regulations 0sections /./.2.3 of ;..;91
prescribe the compulsory access algorithm for -,. The algorithm defines
three parameters as follows:
;:tended transmission slots +
Ma:imum retransmission times <;+
+
>t is t$e number o! slots bet#een t#o sendin #$en t$e MS ?ee)s
sendin multi)le c$annel request messaes.
S
>t is related to c$annel combination" and is an intermediate
variable o! access alorit$m. >t is determined by + and CCC'
con!iuration.
&&. For"at
The value of T is from / to 23, 2., 2<, 3;, 3=, /3, and =;.
The value of , ranges as listed in Palues of ,
Halues o! S
T
, in different ###& combination methods
The ###& and ,*##& does
not share a physical channel
The ###& and ,*##& share a
physical channel
/, 9, 2., =; == .2
., :, 2< A< =3
=, 2;, 3; 2;: =9
<, 22, 3= 2</ 9<
A, 23, /3 32A 22=

&&&. Con4iguration and &n4luence
To access the networ), the -, must originate an immediate assignment
process. To begin the process, the -, sends 0!DT I 21 channel request
messages on !#&. To reduce conflicts on !#&, the time for -, to send
channel request messages must meet the following requirements:
+$e number o! slots -not includin slots !or sendin
messaes. bet#een oriinatin immediate assinment )rocess by
MS and sendin t$e !irst c$annel request messaes is random. >ts
rane is U0" 1" e" MAO -+" 8. , 1V. 9$en t$e MS oriinates t$e
immediate assinment )rocess" it ta?es a value !rom t$e rane
accordin to even distribution )robability.
+$e number o! slots -not includin slots !or sendin
messaes. bet#een a c$annel request messae and t$e ne:t is
!rom US" S J 1" e" S J + , 1V accordin to even distribution
)robability.
ccording to previous analysis, the greater the T is, the larger the range of
intervals between one channel request message and the next, and the less the
!#& conflicting times is. The greater the , is, the greater the interval
between one channel request message and the next, the less the !#&
conflicting times is, and the more efficiently the ,*##& is used. &owever,
the increase of T and , leads to longer time for -, to access the networ), so
the access performance of the whole networ) declines. Therefore you must
configure T and , properly.
, is calculated by -, according to T and combination of ##&. Eou can
configure T freely and sends it to -, by system information. 7sually, you
need configure T properly to ma)e T I , as small as possible 0to reduce the
time for -, to access the networ)1@ meanwhile you must ensure an effective
assignment of ,*##& to avoid overload 0for all random access requests, the
system does not distinguish whether they are from the same -,, but assigns a
,*##&1. (n operation, you can ad6ust the value according to traffic
measurement of cell immediate assignment.
.2./ Mini"u" Access Le0el o4 $AC)
&. 9e4inition
The minimum access level of !#& is the level threshold for the system to
6udge whether there is a random access request.
&&. For"at
The minimum access level of !#& ranges from ; to </ 0corresponding to M
22; d8m to M.A d8m1.
The unit is level grade value.
&&&. Con4iguration and &n4luence
Bhen the access burst level of !#& is greater than the threshold, the 8T,
6udges that there is an access request. The 8T,, together with the parameter
random access error threshold, determines whether the random access burst is
valid. To configure the parameter properly, you must combine actual
sensitivity of the base station and the parameter minimum received level
permitted for MS to cce!!. This prevents the -, from failing in calling
though there are signals. The access burst level of !#& affects call drop
rate and access range 0coverage1, so you must pay attention to the influence
on access of -,.
.2.. $ando" Access =rror Threshold
&. 9e4inition
+,- protocols prescribe that by relativity of 6udgment training sequence 0.2
bits1 the system can 6udge whether the received signals are the random access
signals of -,.
&&. For"at
The value ranges from ; to 3==. The recommended value is 29;.
&&&. Con4iguration and &n4luence
The random access error threshold defines the relativity of training sequence.
(f the smaller it is, the more errors of random access signals permitted by the
networ) are, the easily the -, randomly accesses the networ), and the
greater the report error rate is. (f the greater the random access error threshold
is, the smaller the report error rate is, and the more difficult the access to the
networ) is when signals are wea). ,ee protocol ;.;9, ;=;3.
The system requires the random access error threshold transferred by current
bit of .2 bit training sequence.
:;M2;; //
2;2M23; /.
232M2.; /=
2.2M2<; /<
2<2M2A= /A
2A<M2:= /9
2:<M332 /:
333M3./ .;
3..M3=; .2
;M9: or 3=2M
3==
/9

The two parameters random access error threshold and minimum access level
of !#& determine the validity of random access burst.
.2.A Access Control Class :ACC>
&. 9e4inition
+,- regulations 0;3.221 prescribe that each +,- user 0common user1
corresponds to an access class, ranging from class ; to class :. The access
class is stored in ,(- of mobile users. "or special users, +,- regulations
reserves five special access classes, ranging from class 22 to class 2=. Theses
classes are prior to other classes in accessing. ,pecial users might have one or
more access classes 0between 22 and 2=1, which are also stored in user ,(-.
7sers of class 22 to 2= are prior to that of class ; to :. &owever, the class
between ; and : or between 22 and 2= does not mean priority.
The access class is distributed as follows:
Class 009* common users
Class 11* users !or @GME manaement
Class 1%* users !or security de)artments
Class 12* common business de)artments -in c$are o! #ater"
as.
Class 14* emerency services
Class 11* @GME sta!!
7sers of class ;M: have its access rights catering for home JH-N and visited
JH-N. 7sers of class 22 and 2= have its access rights catering for visited
JH-N only. 7sers of class 23, 2/, and 2. have its access rights catering for
in the country where home JH-N belongs to.
&&. For"at
The access control class consists of two parts:
Common access control class
Halue rane* a c$ec? o)tion" includin class 0 disabled" e" class
9 disabled.
<ecommended value* all 0.
S)ecial access control class
Halue rane* a c$ec? o)tion" includin class 11 disabled" e" class
11 disabled.
<ecommended value* all 0.
(f a class is configured to 2, it means that access is forbidden. "or example, a
common access class is configured to 2;;;;;;;;;@ common users excluding
class ; users can access the networ).
&&&. Con4iguration and &n4luence
#;M#2= 0excluding #2;1 are set by equipment room operators. 7sually these
bits are configured to 2. Jroper configuration contributes to networ)
optimization as follow:
9$en installin a base station" startin a base station" or
maintainin and testin in some cells" con!iure C00C11
-e:cludin C10. to 1. >n t$is #ay" di!!erent users are )revented
!rom accessin t$e net#or?" so t$e installin and maintenance is
less in!luenced.
Durin busy $ours o! cells #it$ $i$ tra!!ic" conestion
occurs" <AC' con!lictin time increase" AGC' tra!!ic overloads"
and Abis inter!ace tra!!ic overloads. 9$en you con!iure class o!
some users to 1" you can reduce t$e tra!!ic o! t$e cell.
.2.7 Ma'i"u" $etrans"ission Ti"es :$=T>
&. 9e4inition
,ee +,- regulation ;..;9. Bhen an -, originates an immediate assignment
process, it sends a channel request message to the networ) on !#&. The
!#& is an HG&, so the -, can send multiple channel request messages
before receiving immediate assignment messages, to increase access success
rate of -,. The maximum retransmission times - 0!DT1 is determined by
equipment room operators, and sent to -, by ,(.
&&. For"at
The maximum retransmission times consists of two bits, with the meanings
listed in #oding of maximum transmission times -
Codin o! ma:imum transmission times M
- maximum transmission times
;; 2
;2 3
2; .
22 A

&&&. Con4iguration and &n4luence
The greater the - is, the higher the success rate of call attempt is, and the
higher the connection rate is, but the load of !#&, ###&, and ,*##&
increase. (n cell with high traffic, if the !DT is over great, overload of radio
channels and congestion occur, so the connection rate and radio resource
utilization declines sharply. (f the !DT is over small, the call attempt times of
-, reduces, success rate reduces, so the connection rate reduces. Therefore,
proper configuration of !DT for each cell help utilize networ) radio resources
and improve connection rate.
"or configuration of !DT -, refer to the following methods:
For areas #it$ lo# tra!!ic" suc$ as in suburban or rural
areas" con!iure <;+ to 5 to increase t$e access success rate o!
MS.
For areas #it$ averae tra!!ic" suc$ as common urban areas"
con!iure <;+ to 4.
"or microcell with high traffic and of apparent congestion, configure !DT to
2.
.2.- Control Class o4 MS Ma'i"u" Trans"it Power :MS@T6PB$@MA6@CC)>
&. 9e4inition
-,4T%JB!4-%4##& is sent in 8##& ,(s. (t affects behavior of -, in
idle mode. (t is also used in calculating #2 and #3, and determines cell
selection and reselection.
C1 N <GAPC , <OG;HPACC;SSPM>E ,
MAO--MSP+O@9<PMAOPCC' , @." 0.
<GAPC* averae received level by MS
<OG;HPACC;SSPM>E* minimum received level )ermitted !or
MS to access
MSP+O@9<PMAOPCC'* ma:imum )o#er level o! control
c$annel -control class o! MS ma:imum transmit )o#er.
@* Ma:imum transmit )o#er level o! MS
&&. For"at
The range of -,4T%JB!4-%4##& is ;M/2. "or cells of +,-:;; and
+,-29;;, the d8m values corresponding to the control class are different.
>n a GSM900 net#or?" t$e 2% control class o! ma:imum
transmit )o#er corres)ondin to 0021 is as !ollo#s*
U29" 29" 29" 25" 21" 22" 21" %9" %5" %1" %2" %1" 19" 15" 11" 12" 11"
9" 5" 1" 1" 1" 1" 1" 1" 1" 1" 1" 1" 1" 1" 1V
>n a GSM1800 net#or?" t$e 2% control class o! ma:imum
transmit )o#er corres)ondin to 0021 is as !ollo#s*
U20" %8" %3" %4" %%" %0" 18" 13" 14" 1%" 10" 8" 3" 4" %" 0" 0" 0" 0"
0" 0" 0" 0" 0" 0" 0" 0" 0" 0" 23" 24" 2%V
!ecommended values are = for +,-:;; and ; for +,-29;;.
&&&. Con4iguration and &n4luence
-,4T%JB!4-%4##& determines the power class used before -,
receives power control messages. "or details, see "rotocol 00#.
The smaller it is, the greater the output power of -, is. The -, near the base
station interferes with neighbor channels of the cell, so the access to the
networ) by other -,s and communication quality are influenced. The greater
it is, the smaller the output power of -, is, and the lower the access success
rate of -, at cell borders is. Eou must configure -,4T%JB!4-%4##&
properly according to the serving range of the cell.
.2.1( Power 344set :P3B=$3FFS=T>
&. 9e4inition
Bhen the -, accesses the networ) and before it receives the initial power
control messages, all +,-:;; -,s and type 2 and type 3 *#,29;; -,s use
-,OT%JB!O-%O##& of 8##&. (f the -,OT%JB!O-%O##& exceeds
the maximum transmit power of -,, the -, uses the closest power.
The parameter JGBD!G"",DT is effective to type / *#,29;; -,s. Bhen
the type / *#,29;; -, accesses the networ), it use total power of
-,OT%JB!O-%O##& I JGBD!G"",DT before receiving the initial
power control message. ,ee "rotocol GS$00#.
&&. For"at
The values of ;M/ correspond to ; d8, 3 d8, . d8, and < d8.
The recommended value is 3.
&&&. Con4iguration and &n4luence
The greater the parameter is, the more easily the type / *#,29;; -,
accesses the networ). great JGBD!G"",DT enables -, to access the
networ) afar, but does not help control cross4cell interference, so the networ)
quality is influenced.
.2.11 &MS& Attach%9etach Allowed
&. 9e4inition
The (-,( detach means that the -, informs the networ) of itself wor) state
changing from wor)ing to non4wor)ing. 7sually it refers to when the -,
powers off or the ,(- is ta)en off -,. fter receiving the inform from -,,
the networ) sets the (-,( as in non4wor)ing state.
The (-,( attach is opposite of (-,( detach. (t means that -, informs the
networ) of itself wor) state changing to wor)ing. 7sually it refers to when
the -, powers on or the ,(- is put into -, again. fter the -, turns to
wor)ing state again, it detects whether the current location areas 0H(1 is the
same as that recorded in -, at last.
>! yes" t$e MS starts >MS> attac$ )rocess -t$is is one o!
location u)datin..
>! no" t$e MS starts location u)datin )rocess o! cross
location area.
fter receiving the location updating message or (-,( message from -,, the
networ) sets the (-,( as in wor)ing state.
The parameter (-,( attach$detach allowed 0TT1 is used for informing -,
of the (-,( attach$detach process.
&&. For"at
The value of TT includes ED,$NG. NG means that starting (-,(
attach$detach process by -, is forbidden. ED, means that starting (-,(
attach$detach process by -, is compulsory.
&&&. Con4iguration and &n4luence
7sually configure TT to ED, so that the networ) will not process the
proceeding of the -, after the -, powers off. This frees system resources
0such as J#&1.
&?. Precautions
The TT of different cells in the same location area must be the same to
avoid abnormalities while the -, is called. "or example, in a cell with ED,
as the value of TT, when the -, powers off, it starts (-,( detach process.
Therefore the networ) records that the -, is in non4wor)ing state, so it does
not page the -,. (n a cell with No as the value of TT and the cell being
different from the one where the -, powers off, when the -, powers on
again in the cell, the -, does not start (-,( attach process. (n this situation,
the -, cannot be called normally until it starts location updating process.
.2.12 9irect $etry :9$>
&. 9e4inition
*uring the assignment process of call setup, congestion might cause
assignment failure. The assignment failure causes failure of the whole call.
+,- networ)s has a function to avoid such failures, namely, *!. The *! is
that the 8,, directly assign -, to T#& of neighbor cells. The parameter is
used by system to set whether to allow direct retry function.
&&. For"at
The value of *! includes ED, and NG. ED, means that the system allows
directional retry. NG means that the system does not support direction retry
function.
&&&. Con4iguration and &n4luence
*! improves call success rate. (f conditions are ready, start *!. Gn the
contrary, *! is that the 8,, directly assign -, to T#& of neighbor cells
when congestion occurs in the cell where the -, camps, so the -, can
originates a call in the non4best cell with lowest received level, and extra
interference might be brought about in frequency reuse networ)s. Therefore,
you must use the function properly according to comprehensive networ)
situations.
.3 Serial Para"eters o4 Cell Selection and
$eselection
.3.1 cell;5ar;access
&. 9e4inition
(n the ,( broadcasted in each cell, a bit indicates whether the -, is allowed
to access the networ) in the cell, namely, cellObarOaccess.
&&. For"at
The value of cellObarOaccess includes 2 and ;. The value ; indicates that -,
is allowed to access the networ) from the cell. The value 2 indicates that the
-, is barred to access the networ) from the cell. ctually whether to allow
-, to access the networ) from the cell is determined by both cellObarOaccess
and cellObarOqualify.
&&&. Con4iguration and &n4luence
+$e cellPbarPaccess is con!iured by equi)ment room o)erators. /sually t$e MS
is allo#ed to access t$e net#or? !rom all t$e cells" so cellPbarPaccess is
con!iured to 0. >n s)ecial situations" t$e o)erators #ant some cell !or
$andover service only" so cellPbarPaccess is con!iured to 1
+$e MS usually #or?s in microcells -you can con!iure t$e )riority o! cells and
reselection )arameters to enable t$is.. 9$en t$e MS is callin #$ile movin
!ast" t$e net#or? !orce MS to $and over to t$e base station G. +$e sinals o!
base station G are stroner t$an microcell base station in most areas. 9$en t$e
call terminates" t$e MS Fust cam)s near base station G and at ede o! microcell
cells" t$e MS #ill not reselect a cell accordin to GSM reulations" t$ere!ore
t$e MS cannot return to microcell.
The capacity of base station + is usually small, so the previous phenomenon
leads to congestion of base station +. To solve the problem, you can
configure the cellObarOaccess to 2, namely, to forbid -, directly accessing
base station +. (n area , handover is allowed to base station +.
&?. Precautions
The cellObarOaccess is used only in some special areas. "or common cells, it
is configured to ;.
.3.2 cell;5ar;#uali4y
&. 9e4inition
The cellObarOqualify determines the priority of cells, namely, it enables -,
to select some cell by preference.
&&. For"at
The value of cellObarOqualify includes 2 and ;. The cellObarOqualify and
cellObarOaccess determine the priority state of cells, as listed in Table A42
#ell priorit.
Table 7-1 Cell )riorities
cellObarOqualify cellObarOaccess #ell selection priority #ell reselection state
; ; Normal Normal
; 2 8arred 8arred
2 ; How Normal
2 2 How Normal
n exception is that the cell selection priority and cell reselection state are
normal when the following conditions are met:
+$e cell belons to t$e @GME #$ic$ t$e MS belons to.
+$e MS is in cell test o)eration mode.
+$e cellPbarPaccess is 1.
+$e cellPbarPquali!y is 0.
+$e access control class 11 is disabled.
&&&. Con4iguration and &n4luence
The priority of all the cells are usually configured to norml, namely,
cellObarOqualify = ;. (n microcell and dualband networ)ing, operators might
want -, to camps on the cell of some type by preference. (n this situation,
the equipment room operators can configure the priority of these cells to
norml and other cells to lo".
*uring cell selection, when the proper cells with normal as the priority is not
present 0proper cells means that all parameters meet the conditions for cell
selection, namely, #2 Q ;, and the cell is allowed to access1, the -, will
select cells with low priority.
&?. Precautions
Jay attention to the following aspects:
9$en cell )riority is used as a met$od to o)timi(e net#or?"
t$e cellPbarPquali!y only a!!ects cell selection" #it$out any
in!luence on cell reselection. Rou must o)timi(e t$e net#or? by
combinin cellPbarPquali!y and C%.
Durin cell selection" #$en t$e )ro)er cells #it$ normal as
t$e )riority is not )resent" t$e MS #ill select cells #it$ lo#
)riority. +$ere!ore #$en t$e level o! t$e cell #it$ normal
)riority is lo#" and cells #it$ lo# )riority and $i$ level are
)resent" t$e MS #ill access t$e net#or? slo#ly #$ile )o#erin
on.
.3.3 Mini"u" $ecei0ed Le0el Allowing MS to Access :$6L=?;ACC=SS;M&!>
&. 9e4inition
To avoid bad communication quality, call drop, and a waste of networ) radio
resources due to -, accessing the networ) at low received signal level, +,-
regulations prescribe that when an -, accesses the networ) the received
level must be greater than the threshold level, namely, the minimum received
level allowing -, to access.
&&. For"at
The value range of !%HDPO##D,,O-(N is from M22; d8m to M.A d8m.
&&&. Con4iguration and &n4luence
The recommended !%HDPO##D,,O-(N needs to be approximately equal
to the receiving sensitivity of -,. The !%HDPO##D,,O-(N affects cell
selection parameter #2, so it is important to traffic ad6ustment and networ)
optimization.
"or cells with over high traffic and severe congestion, you can increase
!%HDPO##D,,O-(N. (n this way, the #2 and #3 of the cells decrease,
and the effective coverage range decreases. Eou must not configure
!%HDPO##D,,O-(N over great, because this might cause non4seamless
coverage and complaints for signal fluctuation. (t is recommended that the
!%HDPO##D,,O-(N is smaller than or equal to M:; d8m.
&?. Precautions
Dxcept for areas of high density of base stations and of qualified coverage,
ad6usting cell traffic by !%HDPO##D,,O-(N is not recommended.
.3. Additional $eselection Para"eter &ndicator
&. 9e4inition
The cell selection and reselection by -, depends on the parameters #2 and
#3. Bhether #3 is the cell reselection parameter is determined by networ)
operators. dditional reselection parameter indicator 0**(T(GNH
!D,DHD#T1 informs -, of whether to use #3 in cell reselection.
&&. For"at
**(T(GNH !D,DHD#T consists of 2 bit. (n ,(/, it is meaningless, and
equipment manufacturers configure it to N. The -, uses **(T(GNH
!D,DHD#T of ,(..
9$en ADD>+>AEAG <;S;G;C+ is con!iured to E" t$e meanin
is* i! t$e rest bytes o! S>4 -S>4<estActets. are )resent" t$e MS
must abstract and calculate )arameters related to C% and
related cell reselection )arameter @>.
9$en ADD>+>AEAG <;S;G;C+ is con!iured to R" t$e meanin
is t$at t$e MS must abstract and calculate )arameters related to
C% and related cell reselection )arameter @>.
&&&. Con4iguration and &n4luence
#ells seldom use ,(A and ,(9, so you can configure **(T(GNH
!D,DHD#T to N. Bhen cells use ,(A and ,(9, and the parameter #3 is used
in cell reselection, you can configure **(T(GNH !D,DHD#T to E.
.3./ Cell $eselection Para"eter &ndicator
&. 9e4inition
The cell reselection parameter indicator 0#DHHO!D,DHD#TOJ!-O(N*1
is used in informing -, of whether #3 is a cell reselection parameter and
whether #3 is present.
&&. For"at
The value of #DHHO!D,DHD#TOJ!-O(N* includes E and N, with the
meanings as follows:
R* +$e MS must calculate C% by abstractin )arameters !rom
S>s o! cell broadcast" and set C% as t$e standard !or cell
reselection.
E* +$e MS must set C1 as t$e standard" namely" C% N C1.
&&&. Con4iguration and &n4luence
The equipment room operators determine the value of J(. #onfigure J( to E if
related cells set #3 as the standard for cell reselection@ otherwise, configure it
to N.
.3.. Cell $eselection 344setC Te"porary 344setC and Penalty Ti"e
&. 9e4inition
fter the -, selects a cell, without great change of all the conditions, the -,
will camp on the selected cell. -eanwhile, it does as follow:
Starts measurin sinals level o! BCC' carrier in nei$bor
cells.
<ecords t$e 3 nei$bor cells #it$ reatest sinal level.
Abstract various S> and control in!ormation o! eac$ nei$bor
cell !rom t$e 3 cells.
Bhen conditions are met, the -, hands over from the selected cell to
another. This process is called cell reselection. The conditions include:
Cell )riority
9$et$er t$e cell is barred to access
<adio c$annel level -im)ortant.
Bhen the signal level of neighbor cells exceeds that of the serving cell, cell
reselection occurs. The channel level standard used in cell reselection is #3,
with the calculation as follows:
1. 9$en @;E;G+RP+>M; W 11111*
C% N C1 J C;GGP<;S;G;C+PAFFS;+ , +;M@A<A<RPAFFS;+ M '
-@;EAG+RP+>M; , +.
9$erein" i! @;EAG+RP+>M; , + -:. T 0" t$e !unction '-:. N 08 i! :
c 0" '-:. N 1.
%. 9$en @;E;G+RP+>M; N 11111*
C% N C1 , C;GGP<;S;G;C+PAFFS;+
T is a timer, with ; as the initial value. Bhen a cell is listed by -, in the list
of cells with maximum signal level, start T with step of ..<3ms 0a T*-
frame1. Bhen the cell is removed from the list, the associated T is reset.
fter cell reselection, the T of original cell wor)s as JDNHTEOT(-D.
Namely, temporary offset is not performed on the original cell.
#DHHO!D,DHD#TOG"",DT 0#!G1 modifies cell reselecting time #3.
TD-JG!!EOG"",DT 0TG1 is supplemented to #3 from starting wor)ing
of T to the prescribed time.
JDNHTEOT(-D is the time for TD-JG!!EOG"",DT having effect on
#3. Bhen JDNHTEOT(-D = 22222, the -, is informed of using #3 = #2
M #!G.
#DHHO!D,DHD#TOG"",DT, TD-JG!!EOG"",DT, and
JDNHTEOT(-D are cell reselection parameters.
9$en t$e cell reselection )arameter @> is 1" t$e MS is
in!ormed o! receivin values o! t$ree )arameters on BCC'.
>! @> is 0" t$e MS Fudes t$at t$e )revious t$ree )arameters
are 0" namely C% N C1.
(f the #3 of a cell 0in the same location area as the serving cell1 calculated by
-, is greater than the #3 of the cell where -, camps, and this lasts for over
=s, the -, reselects to camp on the cell.
(f the #3 of a cell 0in different location area as the serving cell1 calculated by
-, is greater than the sum of #3 of the cell where -, camps and cell
reselect hysteresis, and this lasts for over =s, the -, reselects to camp on the
cell.
The interval between two reselections is at least 2=s, and this avoids frequent
cell reselection by -,.
#3 is formed on the combination of #2 and artificial offset parameters. The
artificial offset parameters help -, camp on or prevent -, from camping on
some cell. This balances the traffic of the networ).
&&. For"at
1. +$e cell reselection o!!set -C<A. is in decimal" #it$ unit o!
dB. >t ranes !rom 0 to 32" #$ic$ means 0 to 1%3 dB -% dB as t$e
ste).. +$e recommended value is 0.
%. +$e tem)orary o!!set -+A. is in decimal" #it$ unit o! dB. >t
ranes !rom 0 to 5" #$ic$ means 0 to 50 dB -10 dB as t$e ste)..
+$e recommended value is 0.
2. +$e )enalty time -@+. is in decimal" #it$ unit o! second. >t
ranes !rom 0 to 21. +$e value 0 to 20 means %0s to 3%0s -%0s as
t$e ste).. +$e value 21 is reserved !or c$anin t$e e!!ect
direction o! C% by C<A. +$e recommended value is 0.
&&&. Con4igurationa and &n4luence
The previous parameters can be ad6usted accordingly in the following three
situations:
1. 9$en t$e communication quality is bad due to $eavy tra!!ic
or ot$er causes" c$ane t$e )arameters to enable MS not cam)s
on t$e cell -t$e cell is e:clusive !rom t$e MS.. For t$is situation"
con!iure @+ to 21" so +A is ine!!ective. C% N C1 0 C<A. +$e C% is
arti!icially lo#ered. So t$e )robability !or MS to reselect t$e cell
decreases. >n addition" t$e equi)ment room o)erators can
con!iure C<A to a )ro)er value accordin to t$e e:clusive level
o! t$e cell by MS. +$e reater t$e e:clusion is" t$e reater t$e
C<A is.
%. For cells #it$ lo# tra!!ic and equi)ment o! lo# utili(ation"
c$ane t$e )arameters to enable MS to cam) on t$e cell -t$e
cell is )rior.. >n t$is situation" con!iure C<A to 00%0 dB
accordin to t$e )riority. +$e $i$er t$e )riority is" t$e reater
t$e C<A is. +A is con!iured t$e same as or a little reater t$an
C<A. @+ $el)s avoid over !requent cell reselection" t$e
recommended value o! @+ is %0s or 40s.
2. For cell #it$ averae tra!!ic" con!iure C<A to 0" @+ to
11111 so t$at C% N C1. Eo arti!icial in!luence is on t$e cell.
&?. Precautions
(n whatever situations, the #!G must not be greater than /; d8, because over
great #!G leads to unstable networ), such as complaints about signal
fluctuation.
.3.A Cell $eselection )ysteresis :C$)>
&. 9e4inition
#!& affects cell reselection of cross location area. The -, starts cell
reselection if the following conditions are met:
+$e sinal level o! nei$bor cell -in di!!erent location area.
is reater t$an t$at o! t$e servin cell.
+$e di!!erence bet#een t$e sinal levels o! t$e nei$bor
cell and t$e servin cell must be reater t$an t$e value
)rescribed by cell reselection $ysteresis.
The difference is based on the cell reselection methods used by -,. (f the
-, reselects a cell with #3, then compare values of #3.
&&. For"at
#!& is in decimal, with unit of d8. The range is ; to 2., with step of 3 d8.
The recommended value is ..
&&&. Con4iguration and &n4luence
(f the original cell and target cell belongs to different location areas, the -,
must originate a location updating process after cell reselection. *ue to the
attenuation feature of radio channels, the #3 of two cells measured at the
bordering area of neighbor cells fluctuates much, so the -, reselect cells
frequently. The interval between two reselections is over 2=s, which is rather
short for location updating. The signal flow of networ) increases sharply,
radio resources cannot be fully utilized.
*uring location updating, the -, cannot respond to paging, so the
connection rate decreases. d6ust #!& according to signal flow and
coverage. Bhen signal flow overloads or location updating of cross location
area is frequent, the cell reselection hysteresis is increased as recommended.
Eou must avoid abnormal coverage due to over large location area.
&?. Precautions
*o not configure #!& to ; d8.
. Para"eters A44ecting !etwor8 Functions
..1 !ewly =sta5lished Cause &ndicator :!=C&>
&. 9e4inition
(n a +,- networ), the traffic channel 0T#&1 consists of full4speed T#& and
half4speed T#&. Bhen the networ) supports half4speed T#&, the -, is
informed of whether the area supports half4speed T#& by ND#(.
&&. For"at
The value of ND#( includes E and N, with the meaning as follows:
R means t$at t$e area su))ort $al!,s)eed +C'.
E means t$at t$e area cannot su))ort $al!,s)eed +C'.
&&&. Con4iguration and &n4luence
&alf4speed T#&s enable each carrier to support more traffic channel, but you
must confirm whether the system support half4speed T#&.
..2 Power Control &ndicator :PB$C>
&. 9e4inition
The JB!# informs -, of whether to ta)e statistics of downlin) level of
8##& carrier slot for measuring average value when the 8##& frequency
participates in frequency hopping. The causes to configuring JB!# are as
follows:
GSM reulations allo# !requency $o))in c$annels to use
BCC' -!requency $o))in not in BCC' slots. .
GSM reulations allo# do#nlin? )o#er control over
!requency $o))in c$annels.
+$e MS needs sinal level o! t$e measured nei$bor cells" so
t$e )o#er o! eac$ slot on BCC' !requency is )ro$ibited to
c$ane. +$e do#nlin? )o#er control does not involve carrier
slots !or BCC' #$ic$ includes t$e !requency $o))in.
"or previous causes, when the -, measures the average downlin) channel
level with common methods, the measurement result is inaccurate for power
control because the average value includes the downlin) received level of
8##& carriers the power of which are not controlled, so the measurement
report is inaccurate for power control.
To avoid the influence on power control, when the -, calculates average
received level during frequency hopping, the received level obtained from
8##& carrier slot must be removed 0see +,- regulations ;=.;91.
&&. For"at
The value of JB!# includes ; and 2, with meanings as follows:
9$en @9<C is 0" t$e measurement result by MS includes
BCC' carrier.
9$en @9<C is 1" t$e measurement result by MS does not
include BCC' carrier.
&&&. Con4iguration and &n4luence
The JB!# is usually configured to ;. #onfigure it to 2 if all the following
conditions are met:
C$annels $ave !requency $o))in on t#o or more
!requencies.
Ane o! t$e !requency is BCC' carrier !requency.
+$e system uses do#nlin? )o#er control.
&?. Precautions
The value of JB!# depends actually on the following parameters:
9$et$er to use !requency $o))in.
9$et$er t$e $o))in !requency includes BCC' carrier.
9$et$er t$e system uses do#nlin? )o#er control.
..3 9iscontinuous Trans"it o4 2plin8
&. 9e4inition
*iscontinuous transmit of uplin) 0*T%71 refers to the process for -, not to
transmit signals during silent period 0see description about *T% in #hapter
31.
&&. For"at
Bhether the networ) allows uplin) to use discontinuous transmit 0*T%1 is
set by equipment room operators. *T% ranges from ; to 3, with the following
meanings:
0* MS can use D+O/.
1* MS must use D+O/.
%* MS cannot use D+O/.
&&&. Con4iguration and &n4luence
7sing uplin) *T% affects call quality, but it is helpful in the following
aspects:
Go#er inter!erence to radio c$annels.
Due to t$is" t$e averae call quality o! net#or? is im)roved.
Cut )o#er consum)tion by MS
"or the previous advantages, *T% is recommended to use.
.. 9iscontinuous Trans"it o4 9ownlin8
&. 9e4intion
*iscontinuous transmit of downlin) 0*T%*1 means the networ) does not
transmit signals during silent period.
&&. 9e4inition
*T%* is in string, and the range is ED, and NG. The meanings are as
follows:
R;S* Do#nlin? uses D+O.
EA* Do#nlin? does not use D+O.
&&&. Con4iguration and &n4luence
7sing downlin) *T% affects call quality in a limit scale, but it is helpful in
the following aspects:
Go#er inter!erence to radio c$annels.
Due to t$is" t$e averae call quality o! net#or? is im)roved.
<educe load o! base station C@/
Therefore, if possible, you use *T%.
&?. Precautions
ccording to +,- regulations, downlin) *T% is optional. (f the base station
equipment supports *T%*, then use it. &owever, you must ensure that voice
transcoder is available to support *T%*.
../ Call $esetup Allowed
&. 9e4inition
Bhen coverage voids cause radio lin) failure, consequently call drop, the -,
starts to resetup the call for recovery. Bhether resetting up the call is allowed
depends on the parameter call resetup allowed 0!D1.
&&. For"at
The values of call resetup allowed are 2 and ;, with meanings as follows:
1* Call resetu) is allo#ed in t$e cell.
0* Call resetu) is !orbidden in t$e cell.
&&&. Con4iguration and &n4luence
Bhen a connected -, passes coverage voids, call drop occurs easily. (f call
resetup is allowed, the average call drop rate 0#*!1 is lowered. &owever,
call resetup ta)es longer time, and most users disconnects before completion
of call resetup. Therefore call resetup is difficult to achieve, and even wastes
abundant radio resources. (n a word, call resetup is disabled.
... ="ergency Call Allowed
&. 9e4inition
The following -,s cannot en6oy various services:
MS #it$out S>M
MS #it$ ACC as one o! C0 to C9 and #it$ cellPbarPaccess
The parameter emergency call allowed 0D#1 determines whether the -, is
allowed for emergency calls, such as police emergency call.
&&. For"at
D# consists of 2 bit. "or the -, with ## of #; to #: or without ,(-, the
D# is NG, meaning emergency call forbidden. ED, means emergency call
allowed. "or the -, with ## of #22 to #2=, when both the access control
bit and D# are configured to forbidden, it is forbidden for emergency calls.
&&&. Con4iguration and &n4luence
ccording to the +,- regulations, the emergency number is 223, different
from that in #hina. The #hinese emergency call cannot function as prescribed
in +,- regulations. "or international roaming users, set 223 to answerphone
to inform users of various special service numbers. Therefore, setting
emergency call must be allowed through configuring radio parameters,
namely, configure D# to 2.
..A =arly Class"ar8 Sending Control
&. 9e4inition
(n a +,- networ), the -, classmar) mar)s the following aspects:
Service ca)acity
Su))orted !requency band
@o#er ca)acity
;ncry)tion ca)acity
#lassmar) consists of classmar)2, classmar)3, and classmar)/. +,- -,.
(n a +,- networ), the -, reports #lassmar)2 or #lassmar)3 information
immediately after D,T(N*R#- ,D!P !DSQ 0corresponding to H34,8-
at 7m interface1 is allocated. #lassmar)/ 0#-/1 information includes power
information of various frequency band of multi4frequency -,.
*uring handover between different bands, the power class must be correctly
described. Bhen the +,- system pages and transmits 83 in different
bands, it must )now the #-/ message. (n +,- regulation Jhase3plus, early
classmar) sending control 0D#,#1 is added. D#,# means that by ,( the
system informs -, of reporting #lassmar)/ after lin) setup. This avoids
querying process by networ).
&&. For"at
The values of D#,# are E and N, with the following meanings:
R* +$e MS re)orts Classmar?2 to t$e net#or? immediately
a!ter lin? setu).
E* +$e MS is !orbidden to re)ort its Classmar?2 to net#or?
initiatively.
&&&. Con4iguration and &n4luence
The ma6or information of #lassmar)/ is for dualband networ), so do as
follows:
Con!iure ;CSC to E in sinle !requency GSM a))lication
areas.
Con!iure ;CSC to R in dualband GSM a))lication areas.
&?. Precautions
(n a dualband networ), configure the parameter of all cell to the same value.
#onfiguring the parameter to different values in one or more cells is
forbidden@ otherwise, the networ) quality declines.
./ Fre#uency )opping Para"eters
./.1 Fre#uency )opping Se#uence !u"5er
&. 9e4inition
(n a +,- networ), the cell allocation 0#1 means the set of carriers used by
each cell, recorded as >!;, !2, L, !n 4 2?. Bherein, !i indicates the
absolute channel number. "or each communication process, the set of carriers
used by base station and -, is mobile allocation 0-1, recorded as >-;,
-2, L, -n 4 2?. Bherein, -i indicates the absolute channel number.
Gbviously - is a subset of #.
*uring a communication process, the air interface uses a carrier number, one
element of -. The variable mobile allocation index 0-(1 determines an
exact element of -. ccording to the frequency hopping algorithm in +,-
regulation ;=.;3, the -( is the T*- frame number 0!N1 or reduced
frame number 0!"N1, frequency hopping sequence number 0&,N1, and
mobile allocation index offset 0-(G1.
Bherein, the &,N determines two aspects:
+rac? o! !requency )oints durin !requency $o))in
+$e async$ronous nei$bor cells usin t$e same MA can
avoid continuous !requency collision durin !requency $o))in by
usin di!!erent 'SEs.
&&. For"at
&,N is in decimal, ranging from ; to </, wherein:
0* cyclic !requency $o))in
1032* )seudo !requency $o))in
&&&. Con4iguration and &n4luence
Eou can choose any &,N in cells using frequency hopping, but you must
ensure that the cells using same frequency group must use different &,N. The
following paragraph is an exception:
(n an 2%2 networ), three cells under a base station use the same frequency
group, but they are synchronous cells because of same "N. Therefore the
three cells use the same &,N. Eou must plan -(G properly to avoid
frequency collision of the three cells under the same base station.
./.2 Mo5ile Allocation
&. 9e4inition
The mobile allocation 0-1 in the +,- networ) indicates a frequency set for
frequency hopping. Namely, when the - of a cell is fixed, the
communication frequency points of the cell performs transient in the set by
- according to rules.
The parameter - determines all the elements in -.
&&. For"at
- is a set, with all +,- frequency points as its element, namely:
For GSM900 net#or?s* 101%4 and 951010%2.
For GSM1800 net#or?s* 11%0881
&&&. Con4iguration and &n4luence
- is configured according to networ) designing requirements.
&?. Precautions
#hinese +,- networ)s do not cover all available frequency bands of +,-
system, so configure - in available frequency bands.
The number of elements in each - set cannot exceed </.
The - cannot include 8##& carriers.
The number of - must not be multiples of 2/ if all the following conditions
are met:
/sin D+O
'SE N 0 -cyclic !requency $o))in.
Eou must avoid ,##& to appear usually at the same frequency point.
./.3 Mo5ile Allocation &nde' 344set
&. 9e4inition
*uring communication, the air interface uses a carrier frequency, one element
of - set. -(G determines an exact element of - set. ccording to the
frequency hopping algorithm in +,- regulation ;=.;3, the -( is the
T*- frame number 0!N1 or reduced frame number 0!"N1, frequency
hopping sequence number 0&,N1, and mobile allocation index offset
0-(G1. -(G is an initial offset of -(, and it aims to avoid multiple
channels to use the same frequency carrier in the same time.
&&. For"at
-(G ranges from ; to </.
&&&. Con4iguration and &n4luence
-(G is configured by equipment room operators.
&?. Precautions
The different cells using same group of - must use consistent -(G.
7sing different -(Gs enables different sectors in the same location to use
the same frequency group 0-1 without frequency collision.
.. 9istance Control Para"eters
...1 Call Clearing
&. 9e4inition
#all clearing 0#all#learing1 means that the maximum allowed distance
threshold is cleared between -, and base station in tal).
&&. For"at
#all#learing ranges from ; to </, with unit of T.
&&&. Con4iguration and &n4luence
#onfigure #all#learing according to actual coverage range of a cell. Jroper
configuration of #all#learing helps chec) whether the handover threshold of
the cell is properly defined, especially for urban cells.
(f the call is frequently cleared after #all#learing threshold is defined
according to cell radium, probably the handover threshold is improperly
configured. This is due to that the -, cannot hand over to the best server cell
after exceeding designed coverage range.
*efine #all#learing according to ms!ange-ax, namely, #all#learing Q
ms!ange-ax.
(n actual networ) operation, call clearing is unusually performed, because
radio lin) fails due to over poor coverage before call clearing. *efining
#all#learing aims to restrict the distance between -, and base station and to
avoid -,s in allowed coverage range to interfere other -,s, especially in
areas with complex landform.
The cell coverage range is irregular, so island effect might occur. "or this
phenomenon, define #all#learing to clear calls in island areas.
...2 TA )ando0er Threshold :MS$A!G=MA6>
&. 9e4intion
Bhen the distance between -, and base station reaches or exceeds
-,!N+D-%, distance handover is triggered.
&&. For"at
-,!N+D-% ranges from ; to </, with unit of T. The reference is </.
&&&. Con4iguration and &n4luence
-,!N+D-% must be smaller than #all#learing, and otherwise the
handover function will be actually unavailable. Bhile configuring
-,!N+D-%, you must ad6ust the threshold of other types of handover@
otherwise ping4pong handover occurs. one occasion might be as follows:
The distance between -, and the serving cell exceeds the threshold, but the
signals of target cell are wea)er than that of original cell. #onsequently the
Jower8udget handover is triggered immediately after distance handover is
triggered.
...3 TA $estriction :MS;+S;9&ST;2S=9>
&. 9e4inition
The maximum allowed access distance between base station and -,. (f the
distance between an -, and base station exceeds the maximum allowed
access distance, the -, is forbidden to access cells.
&&. For"at
The range is ; to </, with unit of T. The reference is </.
&&&. Con4iguration and &n4luence
"or its configuration, refer to the method for configuring #all#learing. d6ust
the parameter to enable it consistent with the geographic coverage range of
the cell. ,et a proper threshold to filter pseudo !#& requests to avoid
unnecessary assigning ,*##&.
ccording to tests, for mountain4mounted base stations, the coverage and
interference is difficult to control. (f you define the maximum allowed access
distance to </, the !#& mis6udgment increases 0the system demodulates
interference to !#& bursts by mista)e1. Therefore the radio performance
and traffic measurement indexes of the cell are affected.
.A $adio Lin8 Failure Process and Para"eters
The radio lin) failure is detected from uplin) and downlin). The -,
completes downlin) detection, while the base station completes uplin)
detection.
.A.1 $adio Lin8 Failure Counter :$LC or $adio Lin8 Ti"eout>
&. 9e4inition
The -, originates call resetup or disconnects by force if all the following
conditions are met:
+$e voice or data quality is too )oor to be received.
@o#er control and $andover cannot $el) to im)rove t$e
quality.
disconnection by force actually brings about a call drop, so the -,
considers it a radio lin) failure that the voice or data service is actually too
poor to be received. +,- regulations provide solutions to the previous
problems as follows:
,et a counter , in the -,. The initial value of , is provided at the beginning
of tal), and it is the value of the parameter radio lin) failure counter. ,
changes as follows:
S decreases by 1 i! t$e MS !ails in decodin a correct SACC'
messae #$en t$e MS s$ould receive t$e SACC' messae.
S increases by % i! t$e MS succeed in decodin a correct
SACC' messae.
, cannot exceed the value for radio lin) failure counter. Bhen , equals to ;,
the -, originates call resetup or disconnects by force.
&&. For"at
The step from . to <. is ., with unit of ,##& period as follows:
For +C'" t$e SACC' )eriod is 480ms.
For SDCC'" t$e SACC' )eriod is 450ms.
&&&. Con4iguration and &n4luence
The value of the parameter radio lin) failure counter affects #*! and
utilization of radio resources.
ssume that cell is a neighbor cell to cell 8 and the bordering coverage is
poor. Bhen an -, moves from J to S while in tal),
>! t$e radio lin? !ailure counter is over small" call dro)
occurs be!ore cross,cell $andover.
>! t$e radio lin? !ailure counter is over reat" t$e net#or?
releases related resources until radio lin? e:)ires" t$ou$ t$e
voice quality is too )oor #$en MS cam)s on cell B near @.
+$ere!ore" t$e utili(ation o! radio resources declines.
Jroper configuration of radio lin) failure counter is important, and is related
to the actual situations. To configure radio lin) failure counter, refer to the
following rules:
Con!iure it to bet#een 1% and 34 in areas #it$ over lo#
tra!!ic.
Con!iure it to bet#een 23 and 48 in areas #it$ lo# tra!!ic
and reat coverae radium
Con!iure it to bet#een %0 and 2% in areas #it$ $eavy
tra!!ic.
&?. Precautions
#onfigure radioHin)Timeout to smaller than T/2;:. This contributes to
success of call resetup and avoids the following situation effectively:
8efore the -, releases radio resources due to expiration, the networ) side
completes releasing channels resources and reallocates resources to other
-,s. Therefore two -,s might use the same slot and this causes
interferences even call drop.
.A.2 SACC) Multi4ra"e :$LT3;+S>
&. 9e4inition
!efer to the description of radio lin) failure counter. counter is set
accordingly to radio lin) at base station side for managing radio lin) failures.
The solutions vary due to different equipment providers, but a general method
is as follows:
,et a counter , in the base station. The initial value of , is provided at the
beginning of tal), and it is the value of the parameter radio lin) failure
expiration. , changes as follows:
S decreases by 1 i! t$e MS !ails in decodin a correct SACC'
messae #$en t$e MS s$ould receive t$e SACC' messae.
S increases by % i! t$e MS succeed in decodin a correct
SACC' messae.
, cannot exceed the value for radio lin) expiration of base station. Bhen ,
equals to ;, the -, originates call resetup or disconnects by force, as shown
in Drror: !eference source not found.
&&. For"at
!HT;O8, ranges from . to <..
&&&. Con4iguration and &n4luence
Jroper configuration of radio lin) expiration of base station affects #*! and
utilization of radio resources. (t is related to the actual situations. To
configure radio lin) failure counter, refer to the following rules:
Con!iure it to bet#een 1% and 34 in areas #it$ over lo#
tra!!ic.
Con!iure it to bet#een 23 and 48 in areas #it$ lo# tra!!ic
and reat coverae radium
Con!iure it to bet#een %0 and 2% in areas #it$ $eavy
tra!!ic.
Con!iure it to a reater value in areas #it$ a))arent voids
or #$ere call dro) occurs !requently #$ile t$e MS moves.
&?. Precautions
!HT;O8, and !H# must be consistent.
.7 )ando0er and $elated Para"eters
.7.1 P+GT )ando0er Threshold :)oMargin>
&. 9e4inition
The J8+T handover threshold is power handover tolerance 0handover in
serving areas1. Bhen the signal level of neighbor cell is ho-argin 0d81
higher than that of the serving cell, handover occurs. #omplex radio
propagation conditions cause fluctuation of signal level. 7sing handover
tolerance avoids frequent handover at bordering areas. The J8+T handover
threshold is similar to &GO-!+(N 0+,- ;=.;91.
&&. For"at
The J8+T handover threshold ranges from ; to 23A, corresponding to M<.
d8 to I</ d8. The reference value for suburban areas is <9. The reference
value for urban areas is A; to A3.
&&&. Con4iguration and &n4luence
The J8+T handover threshold aims to ad6ust handover difficulty properly,
and to avoid ping4pong handover. (f it is configured over great, the handover
is delayed and handover is less efficient. Bhen it is smaller than <., the -,
hands over from the serving cell to the neighbor cell with lower level.
.7.2 Mini"u" 9ownlin8 Power o4 )ando0er Candidate Cells :r'Le0MinCell>
&. 9e4inition
(t is the minimum allowed access level for a cell to be a neighbor cell. Bhen
the cell level measured by -, is greater than the threshold, the 8,, list the
cell into candidate cell list for handover 6udgment.
&&. For"at
(t ranges from M22; d8m to M.A d8m.
&&&. Con4iguration and &n4luence
(t is helpful in the following two aspects:
>t uarantees communication quality.
For a common sinle layer net#or? structure" t$e value ranes
!rom 090 dBm to 080 dBm.
>t $el)s allocate tra!!ic bet#een cells averaely.
;s)ecially in multi,layer net#or? structure" to maintain MS in a
net#or? layer" you can increase t$e level o! t$e cell o! t$e
net#or? layer -suc$ as 050 dBm." and also decrease t$at in ot$er
cells.
&?. Precautions
Eou cannot configure rxHev-in#ell over great 0over M<= d8m1 or over small
0lower than M:= d8m1, and otherwise communication quality is affected.
.7.3 )ando0er Threshold at 2plin8 =dge
&. 9e4inition
(f the uplin) received level )eeps being smaller than the handover threshold
at uplin) edge for a period, edge handover can be performed.
&&. For"at
(t ranges from ; to </, corresponding to M22; d8m to M.A d8m. The
recommended values are as follows:
Con!iure it to %1 in urban areas #it$out @BG+ $andover.
Con!iure it to %0 in sinle site o! suburban areas.
Con!iure it to %0 in urban areas #it$ @BG+ $andover
&&&. Con4iguration and &n4luence
Bhen J8+T handover is enabled, the corresponding edge handover threshold
can be lowered. Bhen J8+T handover is disabled, and the edge handover
threshold is over low, an artificial cross4cell non4handover occurs. Therefore
call drop occurs or intra4frequency and side interference occur due to cross4
cell tal).
.7. )ando0er Threshold at 9ownlin8 =dge
&. 9e4inition
(f the downlin) received level )eeps being smaller than the handover
threshold at downlin) edge for a period, edge handover can be performed.
&&. For"at
(t ranges from ; to </, corresponding to M22; d8m to M.A d8m. The
recommended values are as follows:
Con!iure it to 20 in urban areas #it$out @BG+ $andover.
Con!iure it to %1 in sinle site o! suburban areas.
Con!iure it to %1 in urban areas #it$ @BG+ $andover
&&&. Con4iguration and &n4luence
Bhen J8+T handover is enabled, the corresponding edge handover threshold
can be lowered. Bhen J8+T handover is disabled, and the edge handover
threshold is over low, an artificial cross4cell non4handover occurs. Therefore
call drop occurs or intra4frequency and side interference occur due to cross4
cell tal).
.7./ 9ownlin8 Duality $estriction o4 ="ergency )ando0er
&. 9e4inition
(f the downlin) received quality is lower than the threshold of downlin)
quality restriction of emergency handover, the quality difference emergency
handover occurs.
&&. For"at
(t ranges from ; to A;, corresponding to !S 0So, ; to A1 x 2;.
The recommended value is =;.
&&&. Con4iguration and &n4luence
Bhen frequency hopping is enabled, the voice quality is better with the same
!S, you can configure it to <; or A;. Bhen emergency handover occurs, the
intracell handover occurs first. (f there are no other candidate cells, and the
intracell handover is enabled, the intracell handover occurs.
.7.. 2plin8 Duality $estriction o4 ="ergency )ando0er
&. 9e4inition
(f the uplin) received quality is lower than it, quality difference emergency
handover is triggered.
&&. For"at
(t ranges from ; to A;, corresponding to !S 0So, ; to A1 x 2;.
The recommended value is =;.
&&&. Con4iguration and &n4luence
Bhen frequency hopping is enabled, the voice quality is better with the same
!S, you can configure it to <; or A;. Bhen emergency handover occurs, the
intracell handover occurs first. (f there are no other candidate cells, and the
intracell handover is enabled, the intracell handover occurs.
.7.A 2plin8 Duality Threshold o4 &nter4erence )ando0er
&. 9e4inition
(t is the uplin) received quality threshold of the serving cell that triggers
interference handover. The interference handover is triggered if all the
following conditions are met:
+$e u)lin? received level is $i$er t$an t$e u)lin? received
)o#er t$res$old o! inter!erence $andover.
+$e u)lin? received quality is lo#er t$an t$e u)lin? quality
t$res$old o! inter!erence $andover.
Bhen handover switch is enabled, the interference handover occurs within
the cell by preference.
&&. For"at
(t ranges from ; to A;, corresponding to !S 0So, ; to A1 x 2;.
The recommended value is =;.
&&&. Con4iguration and &n4luence
Bhen frequency hopping is enabled, the voice quality is better with the same
!S, you can configure it to <; or A;. Bhen interference handover is
triggered, select the candidates according to the sorted result. (f the serving
cell ran)s first and its intracell handover is enabled, the -, selects the
serving cell@ otherwise it selects the second candidate cell.
.7.7 9ownlin8 Duality Threshold o4 &nter4erence )ando0er
&. 9e4inition
(t is the downlin) received quality threshold of the serving cell that triggers
interference handover. The interference handover is triggered if all the
following conditions are met:
+$e do#nlin? received level is $i$er t$an t$e do#nlin?
received )o#er t$res$old o! inter!erence $andover.
+$e do#nlin? received quality is lo#er t$an t$e do#nlin?
quality t$res$old o! inter!erence $andover.
Bhen handover switch is enabled, the interference handover occurs within
the cell by preference.
&&. For"at
(t ranges from ; to A;, corresponding to !S 0So, ; to A1 x 2;.
The recommended value is =;.
&&&. Con4iguration and &n4luence
Bhen frequency hopping is enabled, the voice quality is better with the same
!S, you can configure it to <; or A;. Bhen interference handover is
triggered, select the candidates according to the sorted result. (f the serving
cell ran)s first and its intracell handover is enabled, the -, selects the
serving cell@ otherwise it selects the second candidate cell.
&?. Precautions
The interference handover quality must be better than emergency handover
quality.
.7.- 2plin8 $ecei0ed Power Threshold o4 &nter4erence )ando0er
&. 9e4inition
(f interference handover occurs due to uplin) quality, the serving cell must
reach the minimum uplin) received power threshold. (f this is met, the system
6udges that uplin) is interfered, so interference handover is triggered.
The interference handover is triggered if all the following conditions are met:
+$e u)lin? received level is $i$er t$an t$e u)lin? received
)o#er t$res$old o! inter!erence $andover.
+$e u)lin? received quality is lo#er t$an t$e u)lin? quality
t$res$old o! inter!erence $andover.
Bhen handover switch is enabled, the interference handover occurs within
the cell by preference.
&&. For"at
(t ranges from ; to </, corresponding to M22; d8m to M.A d8m.
The recommended value is 3=.
&&&. Con4igurationa and &n4luence
Bhen interference handover is triggered, select the candidates according to
the sorted result. (f the serving cell ran)s first and its intracell handover is
enabled, the -, selects the serving cell@ otherwise it selects the second
candidate cell.
.7.1( 9ownlin8 $ecei0ed Power Threshold o4 &nter4erence )ando0er
&. 9e4inition
(f interference handover occurs due to uplin) quality, the serving cell must
reach the minimum downlin) received power threshold. (f this is met, the
system 6udges that downlin) is interfered, so interference handover is
triggered.
The interference handover is triggered if all the following conditions are met:
+$e do#nlin? received level is $i$er t$an t$e do#nlin?
received )o#er t$res$old o! inter!erence $andover.
+$e do#nlin? received quality is lo#er t$an t$e do#nlin?
quality t$res$old o! inter!erence $andover.
Bhen handover switch is enabled, the interference handover occurs within
the cell by preference.
&&. For"at
(t ranges from ; to </, corresponding to M22; d8m to M.A d8m.
The recommended value is /;.
&&&. Con4igurationa and &n4luence
Bhen interference handover is triggered, select the candidates according to
the sorted result. (f the serving cell ran)s first and its intracell handover is
enabled, the -, selects the serving cell@ otherwise it selects the second
candidate cell.
.7.11 Ma'i"u" $epeated Ti"es o4 Physical Messages :!E1>
&. 9e4inition
(n asynchronous handover process of +,- system, when the -, receives
handover messages of the networ), it sends handover access messages on the
target channel. fter the networ) receives the message, it does as follows:
1. Calculate related <F !eatures.
%. Send )$ysical messaes -it t$e c$annel messaes are
encry)ted" start encry)tion and decry)tion alorit$m. in unit
data to MSs.
2. Start timer +2101.
(f the networ) does not receive correct layer 3 frames sent by -, until
expiration of T/2;=, the networ) will resend the physical message and restart
T/2;=. The maximum times for resending physical messages is determined
by the parameter maximum repeated times of physical messages 0NE21
&&. For"at
NE2 ranges from ; to 3=..
The recommended value is 3;.
&&&. Con4iguration and &n4luence
Bhen the networ) receives the handover access messages sent by -,, the
physical channel 0J#&1 needs to be synchronous. (f the communication
quality on channels is guaranteed, the -, can receive physical messages
correctly and send layer 3 frames to the networ).
(f the physical messages are sent multiple times, and the networ) cannot
receive layer 3 frames sent by -,, the J#& is too poor to communicate
normally. Though lin) is setup after multiple trials, the communication
quality is not guaranteed. This lowers the utilization of radio resources.
Therefore configure NE2 to a smaller value.
&?. Precautions
#onfiguring NE2 is affected by T/2;=. (f T/2;= is configured to a short
value, then the NE2 needs to be increased accordingly.
(f a handover trial fails before the original cell receives the &N*GPD!
"(H7!D message, and the T/2;= of the target cell expires for Ny times, the
target 8T, sends a #GNND#T(GN "(H7!D (N*(#T(GN message to the
target 8,#. Though the -, might return to the original channel, the traffic
measurement counters from multiple vendors will ta)e statistics of
connection failure.
To avoid the previous phenomenon, configure T/2;= as follows:
Ny T T/2;= Q T/23. I delta 0delta: the time between expiration of T/23.
and receiving &N*GPD! "(H7!D message by original 8T,1
.7.12 Multi5and &ndicator :"ulti5and;reporting>
&. 9e4inition
(n a single band +,- networ), when the -, send measurement reports of
neighbor cells to the networ), it needs to report the content of the six
neighbor cells with strongest signals.
(n a multiband networ), operators wish that -, uses a band by preference in
cross4cell handover. Therefore the -, sends measurement reports according
to signal strength and signal band. The parameter multiband indicator
indicates -, to report content of multiband neighbor cells.
&&. For"at
The multiband indicator ranges from ; to /, with meanings as follows:
0* Accordin to sinal strent$ o! nei$bor cells" t$e MS
must re)ort si: allo#ed measurement re)orts o! nei$bor cells
#it$ stronest sinals and ?no#n ECC" #it$ t$e nei$bor cells in
#$atever band.
1* +$e MS must re)ort t$e allo#ed measurement re)ort o! a
nei$bor cell #it$ ?no#n ECC and #it$ stronest sinals at eac$
band e:)ect !or t$e band used by t$e servin cell. +$e MS must
also re)ort t$e nei$bor cells o! t$e band used by t$e servin
cell in rest locations. >! t$ere are ot$er rest locations" t$e MS
must re)ort conditions o! ot$er nei$bor cells in any band.
%* +$e MS must re)ort t$e allo#ed measurement re)ort o!
t#o nei$bor cells #it$ ?no#n ECC and #it$ stronest sinals at
eac$ band e:)ect !or t$e band used by t$e servin cell. +$e MS
must also re)ort t$e nei$bor cells o! t$e band used by t$e
servin cell in rest locations. >! t$ere are ot$er rest locations"
t$e MS must re)ort conditions o! ot$er nei$bor cells in any
band.
/: +$e MS must re)ort t$e allo#ed measurement re)ort o!
t$ree nei$bor cells #it$ ?no#n ECC and #it$ stronest sinals
at eac$ band e:)ect !or t$e band used by t$e servin cell. +$e
MS must also re)ort t$e nei$bor cells o! t$e band used by t$e
servin cell in rest locations. >! t$ere are ot$er rest locations"
t$e MS must re)ort conditions o! ot$er nei$bor cells in any
band.
&&&. Con4iguration and &n4luence
(n multiband networ)s, it is related to traffic of each band. "or configuration,
refer to the following rules:
>! t$e tra!!ic o! eac$ band is a))ro:imately equal" and
o)erators do not select a band intentionally" you can con!iure
t$e multiband indicator to 0
>! t$e tra!!ic o! eac$ band is obviously di!!erent" and
o)erators #ant MS to select a band by )re!erence" you can
con!iure t$e multiband indicator to 2.
For situations bet#een t$e )revious t#o" con!iure
multiband indicator to 1 or %.
.7.13 Per"itted !etwor8 Color Code :ncc per"itted>
&. 9e4inition
*uring a tal), the -, must report the measured signals of neighbor cells to
the base station, but each report includes only six neighbor cells. Therefore
the -, is configured to report the potential handover target neighbor cells,
instead of reporting unselectively and according to signal level.
To enable previous functions, restrict -, to measure the cells with the fixed
networ) color code 0N##1. The NN# allowed by parameters list the N##s of
the cells to be measured by -,. The -, compares the measured N## of
neighbor cells and N##s set allowed by parameters. (f the measured N## is
in the set, the -, reports the N## to the base station@ otherwise, the -,
discard the measurement report.
&&. For"at
The parameter ncc permitted is a bit mapping value, consisting of 9 bits. The
most significant bit is bit A while the least significant bit is bit ;. Dach bit
corresponds to an N## code ; to A 0see +,- regulations ;/.;/ and ;..;91.
(f the bit N is ; 0N ranges from ; to A1, the -, needs not to measure the level
of the cell with N## of N. Namely, it only measures the signal quality and
level of the cells corresponding to bit number of 2 in N## and ncc permitted
configuration.
&&&. Con4iguration and &n4luence
Dach area is allocated with one or more N##s. (n the parameter ncc permitted
of the cell, the local N## is absolutely and only included. (f excluded,
abnormal handover and call drop occur. "or normal roaming between areas,
the N## of neighbor areas must be included in the edge cells of an area.
&?. Precautions
(mproper configuration of the parameter causes normal handover and even
call drop. The parameter only affects behaviors of -,.
.- Power Control and $elated Para"eters
.-.1 Ma'i"u" Trans"it Power o4 MS :MST6PB$M6>
&. 9e4inition
The transmit power of -, in communication is controlled by 8T,.
ccording to the uplin) signal strength and quality, power budget result, the
8T, controls -, to increase or decrease its transmit power.

& !ote,
(n any situation, power control is prior to related handover for 8,,. Gnly
when the 8,, fails to improve uplin) signal strength and voice quality to the
prescribed level, it starts handover.

To reduce interference between neighbor cells, the power control of -, is
restricted. Namely, the 8T, controls -, to transmit power within the
threshold.
-,T%JB!-% is the maximum transmit power of -, controlled by 8T,.
&&. For"at
-,T%JB!-% ranges from ; to /2.
The d8m values corresponding to +,-:;; and +,-29;; cells are different:
+$e 2% ma:imum transmit )o#er control classes !or GSM900
are U29" 29" 29" 25" 21" 22" 21" %9" %5" %1" %2" %1" 19" 15" 11" 12"
11" 9" 5" 1" 1" 1" 1" 1" 1" 1" 1" 1" 1" 1" 1" 1V
+$e 2% ma:imum transmit )o#er control classes !or GSM900
are U20" %8" %3" %4" %%" %0" 18" 13" 14" 1%" 10" 8" 3" 4" %" 0" 0" 0"
0" 0" 0" 0" 0" 0" 0" 0" 0" 0" 0" 23" 24" 2%V
&&&. Con4iguration and &n4luence
#onfiguring -,T%JB!-% helps control interferences between neighbor
cells, because:
>! MS+O@9<MO is over reat" t$e inter!erence bet#een
nei$bor cells increases.
>! MS+O@9<MO is over small" t$e voice quality declines and
im)ro)er $andover mi$t occur.
.-.2 $ecei0ed Le0el Threshold o4 9ownlin8 Power &ncre"ent :L9$#
&. 9e4inition
Bhen the downlin) received level of the serving cell is smaller than a
threshold, the networ) must start power control to increase the transmit power
of base station and to guarantee communication quality of -,.
The received level threshold of downlin) power increment defines the
downlin) received level threshold. Bhen the downlin) level received by -,
is smaller than it, the base station starts power control to increase its transmit
power.
The parameter N2 means that at lease N2 sampling points must be measured
before starting handover algorithm.
The parameter J2 means the level of at least J2 sampling points in N2
sampling points is smaller than the threshold prescribed by received level
threshold of downlin) power increment.
&&. For"at
(t ranges from M22; d8m to M.A d8m.
N2 ranges from 2 to /3.
J2 ranges from 2 to /3.
&&&. Con4iguration and &n4luence
The received level is between M<; d8m and M9; d8m in a +,- networ), so
configure received level threshold of downlin) power increment to M9= d8m.
N2 is related to propagation quality of radio channels within cell coverage
range. To reduce influence by attenuation, configure N2 to between / and =.
#onfigure J2 to about 3$/ of N2.
.-.3 $ecei0ed Le0el Threshold o4 2plin8 Power &ncre"ent :L2$#
&. 9e4inition
Bhen the uplin) received level of the serving cell is smaller than a threshold,
the networ) must start power control to increase the transmit power of -,
and to guarantee communication quality of -,.
The received level threshold of uplin) power increment defines the uplin)
received level threshold. Bhen the uplin) level received by -, is smaller
than it, the base station starts power control to increase -, transmit power.
The parameter N2 means that at lease N2 sampling points must be measured
before starting handover algorithm.
The parameter J2 means the level of at least J2 sampling points in N2
sampling points is smaller than the threshold prescribed by received level
threshold of uplin) power increment.
&&. For"at
(t ranges from M22; d8m to M.A d8m.
N2 ranges from 2 to /3.
J2 ranges from 2 to /3.
&&&. Con4iguration and &n4luence
The received level is between M<; d8m and M9; d8m in a +,- networ), so
configure received level threshold of uplin) power increment to M9= d8m.
N2 is related to propagation quality of radio channels within cell coverage
range. To reduce influence by attenuation, configure N2 to between / and =.
#onfigure J2 to about 3$/ of N2.
.-. $ecei0ed Duality Threshold o4 9ownlin8 Power &ncre"ent :L9$#
&. 9e4inition
Bhen the downlin) received quality of the serving cell is smaller than a
threshold, the networ) must start power control to increase the transmit power
of base station and to guarantee communication quality.
The received quality threshold of downlin) power increment defines the
downlin) received level threshold. Bhen the downlin) quality received by
-, is smaller than it, the base station starts power control to increase its
transmit power.
The parameter N/ means that at lease N/ sampling points must be measured
before starting handover algorithm.
The parameter J/ means the quality of at least J/ sampling points in N/
sampling points is smaller than the threshold prescribed by received quality
threshold of downlin) power increment.
&&. For"at
(t ranges from ; to A, the voice quality grade.
N/ ranges from 2 to /3.
J/ ranges from 2 to /3.
&&&. Con4iguration and &n4luence
The received quality is ; to 3 of quality grade in a +,- networ), so
configure received quality threshold of downlin) power increment to M9=
d8m.
N/ is related to propagation quality of radio channels within cell coverage
range. To reduce influence by attenuation, configure N/ to between / and =.
#onfigure J/ to about 3$/ of N/.
.-./ $ecei0ed Duality Threshold o4 2plin8 Power &ncre"ent :L2$#
&. 9e4inition
Bhen the uplin) received quality of the serving cell is smaller than a
threshold, the networ) must start power control to increase the transmit power
of -, and to guarantee communication quality.
The received quality threshold of uplin) power increment defines the uplin)
received quality threshold. Bhen the uplin) quality received by -, is smaller
than it, the base station starts power control to increase transmit power of -,.
The parameter N/ means that at lease N/ sampling points must be measured
before starting handover algorithm.
The parameter J/ means the quality of at least J/ sampling points in N/
sampling points is smaller than the threshold prescribed by received quality
threshold of uplin) power increment.
&&. For"at
(t ranges from ; to A, the voice quality grade.
N/ ranges from 2 to /3.
J/ ranges from 2 to /3.
&&&. Con4iguration and &n4luence
The received quality is ; to 3 of quality grade in a +,- networ), so
configure received quality threshold of uplin) power increment to /.
N/ is related to propagation quality of radio channels within cell coverage
range. To reduce influence by attenuation, configure N/ to between / and =.
#onfigure J/ to about 3$/ of N/.
.-.. $ecei0ed Le0el Threshold o4 9ownlin8 Power 9ecre"ent :29$#
&. 9e4inition
Bhen the downlin) received level of the serving cell is greater than a
threshold, the networ) must start power control to decrease the transmit
power of base station and to decrease interference to radio channels.
The received level threshold of downlin) power decrement defines the
downlin) received level threshold. Bhen the downlin) level received by -,
is greater than it, the base station starts power control to decrease its transmit
power.
The parameter N3 means that at lease N3 sampling points must be measured
before starting handover algorithm.
The parameter J3 means the level of at least J3 sampling points in N3
sampling points is greater than the threshold prescribed by received level
threshold of downlin) power decrement.
&&. For"at
(t ranges from M22; d8m to M.A d8m.
N2 ranges from 2 to /3.
J2 ranges from 2 to /3.
&&&. Con4iguration and &n4luence
The received level is between M<; d8m and M9; d8m in a +,- networ), so
configure received level threshold of downlin) power decrement to M9= d8m.
N3 is related to propagation quality of radio channels within cell coverage
range. To reduce influence by attenuation, configure N3 to between / and =.
#onfigure J3 to about 3$/ of N3.
.-.A $ecei0ed Le0el Threshold o4 2plin8 Power 9ecre"ent :22$#
&. 9e4inition
Bhen the uplin) received level of the serving cell is greater than a threshold,
the networ) must start power control to decrease the transmit power of -,
and to decrease interference to radio channels.
The received level threshold of uplin) power decrement defines the uplin)
received level threshold. Bhen the uplin) level received by -, is greater
than it, the base station starts power control to decrease transmit power of
-,.
The parameter N3 means that at lease N3 sampling points must be measured
before starting handover algorithm.
The parameter J3 means the level of at least J3 sampling points in N3
sampling points is greater than the threshold prescribed by received level
threshold of uplin) power decrement.
&&. For"at
(t ranges from M22; d8m to M.A d8m.
N3 ranges from 2 to /3.
J3 ranges from 2 to /3.
&&&. Con4iguration and &n4luence
The received level is between M<; d8m and M9; d8m in a +,- networ), so
configure received level threshold of uplin) power decrement to M<; d8m.
N3 is related to propagation quality of radio channels within cell coverage
range. To reduce influence by attenuation, configure N3 to between / and =.
#onfigure J3 to about 3$/ of N3.
.-.7 $ecei0ed Duality Threshold o4 9ownlin8 Power 9ecre"ent :29$#
&. 9e4inition
Bhen the downlin) received quality of the serving cell is greater than a
threshold, the networ) must start power control to decrease the transmit
power of base station and to decrease space interference.
The received quality threshold of downlin) power decrement defines the
downlin) received quality threshold. Bhen the downlin) quality received by
-, is greater than it, the base station starts power control to decrease transmit
power of -,.
The parameter N. means that at lease N. sampling points must be measured
before starting handover algorithm.
The parameter J. means the quality of at least J. sampling points in N3
sampling points is greater than the threshold prescribed by received quality
threshold of downlin) power decrement.
&&. For"at
(t ranges from ; to A, the voice quality grade.
N. ranges from 2 to /3.
J. ranges from 2 to /3.
&&&. Con4iguration and &n4luence
The received quality is ; to 3 of quality grade in a +,- networ), so
configure received quality threshold of downlin) power decrement to ;.
N. is related to propagation quality of radio channels within cell coverage
range. To reduce influence by attenuation, configure N. to between / and =.
#onfigure J. to about 3$/ of N..
.-.- $ecei0ed Duality Threshold o4 2plin8 Power 9ecre"ent :22$>
&. 9e4inition
Bhen the uplin) received quality of the serving cell is greater than a
threshold, the networ) must start power control to decrease the transmit
power of -, and to decrease space interference.
The received quality threshold of uplin) power decrement defines the uplin)
received quality threshold. Bhen the uplin) quality received by -, is greater
than it, the base station starts power control to decrease transmit power of
-,.
The parameter N. means that at lease N. sampling points must be measured
before starting handover algorithm.
The parameter J. means the quality of at least J. sampling points in N.
sampling points is greater than the threshold prescribed by received quality
threshold of uplin) power decrement.
&&. For"at
(t ranges from ; to A, the voice quality grade.
N. ranges from 2 to /3.
J. ranges from 2 to /3.
&&&. Con4iguration and &n4luence
The received quality is ; to 3 of quality grade in a +,- networ), so
configure received quality threshold of uplin) power decrement to ;.
N. is related to propagation quality of radio channels within cell coverage
range. To reduce influence by attenuation, configure N. to between / and =.
#onfigure J. to about 3$/ of N..
.-.1( Power Control &nter0al :&!T>
&. 9e4inition
(t ta)es a period from beginning of power control to detection of effect of
power control. Therefore an interval must exist between continuous two
power controls@ otherwise the system becomes unstable and even call drop
occurs.
The parameter power control interval 0(NT1 configures the minimum interval
between two continuous times of power control.
&&. For"at
(t ranges from ; to /2s.
&&&. Con4iguration and &n4luence
ccording to frame structure of +,- networ), configure (NT to about /s.
&?. Precautions
(NT cannot be smaller than 2s, and otherwise the system becomes unstable.
.-.11 Power &ncre"ent Step :&!C>
&. 9e4inition
The (N# indicates the power increment of -, or base station in power
control.
&&. For"at
The range of (N# is 3 d8, . d8, or < d8.
&&&. Con4iguration and &n4luence
The recommended value is . d8.
.-.12 Power 9ecre"ent Step :$=9>
&. 9e4inition
The !D* indicates the power decrement of -, or base station in power
control.
&&. For"at
The range of !D* is 3 d8 or . d8.
&&&. Con4iguration and &n4luence
The recommended value of !D* is 3 d8.
.1( Syste"atic &"portant Ti"ers
.1(.1 T31(1
&. 9e4inition
T/2;2 is the 8,# timer controlling time of immediate assignment process.
&&. For"at
T/2;2 ranges from ; to 3==s. The recommended value is /s.
&&&. Con4iguration and &n4luence
(n an immediate assignment process, the 8,# requires 8T, to provide
,*##& to set up signaling channel. Bhen the 8,# sends a channel
activation message, T/2;2 starts timing. Bhen the 8,# receives the setup
instruction sent by 8T,, T/2;2 stops timing. Bhen T/2;2 expires, the
system releases corresponding ,*##& resources. Jroper configuration of
T/2;2 reduces congestion due to dual assignment ,*##& effectively.
The greater the T/2;2 is, the longer the inefficient time for using signaling
resources is. "or example, if the extended transmission delay is improperly
configured 0usually the sum of T and , is over small1, the -, fails in
responding to the networ) side, so the -, resends the random access request
message.
Therefore, the networ) side will assign ,*##& 0the networ) cannot
distinguish the repeated sending access request from the first send1. "or better
use of signaling resources, especially in activating queue function, you must
configure T/2;2 to a smaller value. The minimum interval for sending
channel activation message and receiving setup indicator is <;;ms. "or non4
overload 8,,, the maximum interval is 2.9s.
.1(.2 T31(3
&. 9e4inition
(n inter4 and intra48,, handover, the 8,# determines the time for )eeping
T#& both in handover4originated cell and target cell. Bhen the time receives
handover completion 0intra48,#1 or clearing 0inter48,#1 message, T/2;/
stops.
&&. For"at
T/2;/ ranges from ; to 3==s. The recommended value is =s.
&&&. Con4iguration and &n4luence
The following paragraph is an example of inter48,, handover.
Bhen T/2;/ receives the handover command, it is reset and starts timing.
Bhen it receives clearing command, it is reset. This means that T/2;/
reserves two channels when it is timing, one channel for source 8,#, and one
channel for target 8,#. (f it is over long, two channels are occupied for a
long time and resources might be wasted.
ccording to the tests, if the N,, timer is properly configured, the handover
process occurs within =s. Therefore, the recommended value is =s.
.1(.3 T31(/
&. 9e4inition
,ee the protocol ;.;9 and ;9=9. Bhen sending physical information, the
networ) starts T/2;=. (f the timer expires before receiving any correct frames
from -,, the networ) resends physical information and restarts the T/2;=.
The maximum repeated times is Ny2.
&&. For"at
T/2;= ranges from ; to 3==, with unit of 2;ms.
&&&. Con4iguration and &n4luence
The physical information is sent on "##&. The time for sending four
T*- in a time on "##& is about 29ms. (f the next physical information
is 6ust sent 29ms after the first one, probably the first physical information is
still being sent. The minimum time for sending physical information
continuously and most quic)ly is 3;ms.
&?. Precautions
T/2;= is related to the timer NE2. (f T/2;= is small, configure NE2 to a
greater value. (f a handover trial fails and the T/2;= of the target cell expires
for Ny times before the original cell receives the &N*GPD! "(H7!D
message, the target 8T, sends the #GNND#T(GN "(H7!D (N*(#T(GN
message to the target 8,#.
The counter of target 8,# is renewed though -, might return to the original
channel. To avoid this, the T/2;= must meet the following foulard:
Ny T T/2;= Q T/23. I delta
Bherein, delta is the time between expiration of T/23. and receiving
&N*GPD! "(H7!D message by original 8,#.
.1(. T31(A
&. 9e4inition
T/2;A is a 8,# timer, restricting the time for executing T#& assignment
instruction. (t caters for T#& assignment of intracell handover and channel
assignment of calling.
&&. For"at
T/2;A ranges form ;s to 3==s. The recommended values are as follows:
10s #$en c$annel resources are enou$.
1s #$en c$annel resources are limited.
&&&. Con4iguration and &n4luence
T/2;A starts after the 8,# sends the ,,O#-* message to 8T,. (t stops
after the 8,# receives the ,,O#-J or ,,O"(H message sent by 8T,. (f
T/2;A expires, the system 6udges that the -, disconnects to the networ), so
the occupied resource is released to other -,s. ccording to the measured
statistics result of networ), the channel assignment is complete within 3s. (f
the 8,# does not receive ,,O#-J message after 3s, the assignment
command fails.
(f the radio lin) is bad and some information must be resent, the process
might be prolonged to =s. To avoid premature disconnection, configure
T/2;A to 2;s. (n this way, the -, can reuse the original channel when
handover or assignment fails. Therefore the call drop due to intracell
handover decreases or the system service quality of re4assignment is
improved 0if the system supports re4assignment function1. &owever, the
channel resource might be wasted for several seconds. Bhen the networ)
capacity is limited, you must save the resource as possible.
.1(./ T31(-
&. 9e4inition
The 8,# restricts the releasing resource of ,##& by T/2;:.
&&. For"at
T/2;: ranges from /s to /.s. The recommended T/2;: is as follows:
T/2;: = a I !dioHin)timeGut x ;..9;s, a = 2s or 3s.
&&&. Con4iguration and &n4luence
T/2;: measures the time for channel releasing indicator after sending -,
clearing instructions. (t starts after the 8,# sends *D#TO,##& message
to 8T,. (t stops after the 8,# receives the !DHO(N# message sent by 8T,.
Bhen T/2;: expires, the 8,# sends the #HD! !DS7D,T message to
-,#.
&?. Precautions
The sum of T/222 and T/2;: must be greater than !adioHin)TimeGut. (f
T/2;: is over small, the corresponding radio resources are re4allocated before
!adioHin)TimeGut is due 0radio lin) is not released1.
.1(.. T3111
&. 9e4inition
T/222 is a connection release delay timer, used in deactivation of delayed
channel after disconnection of ma6or signaling lin). T/222 aims to spare
some time for repeated disconnections. Bhen 8,# receives the !DHO(N*
message sent by 8T,, T/222 starts. "or time protection, T/222 stops until
expiration and the 8,# sends the !"O#&NO!DH message to 8T,.
&&. For"at
T/222 ranges from ;s to =s.
The recommended value is 3s.
&&&. Con4iguration and &n4luence
fter the disconnection of ma6or signaling lin), T/222 delays the release of
channels. (t allows the base station to retransmit the instruction for releasing
radio channels to -, within delayed time. fter the base station sends a
release request massage, the radio resources remain for T/222 time.
(f the system capacity is small, configure T/222 as short as possible. The
minimum value of T/222 is 3s, over five multiples of the time for resending
-, the instruction for releasing radio channel resources. greater T/222
might be of no help, but affects congestion of ,*##& and T#& easily.
.1(.A Para"eter T3212
&. 9e4inition
(n a +,- networ), the causes to location updating are as follows:
+$e MS attac$.
+$e MS detects t$at its location area c$anes.
+$e net#or? !orces MS to u)date location )eriodically.
+$e net#or? controls $o# !requent t$e MS u)dates location" and
t$e )eriod !or location u)datin is determined by t$e )arameter
+2%1%.
&&. For"at
T/323 ranges from ; to 3==, with unit of < minutes 02$2; hour1. (f T/323 = 2,
it means that T/323 is < minutes. (f T/323 = 3==, it means that T/323 is 3=
hours and /; minutes. (f T/323 = ;, it means that -, is not required for
periodical location updating in the cell. The recommended T/323 is 3.;.
&&&. Con4iguration and &n4luence
s an important means, the periodical location updating enables networ) to
connect to -,s closely. Therefore, the short the period is, the overall service
performance of the networ) is. nyhow frequent periodical location updating
brings two negative aspects:
+$e sinalin !lo# o! t$e net#or? increases s$ar)ly and t$e
utili(ation o! radio resource declines. 9$en t$e )eriod is over
lon" t$e )rocessin ca)ability o! net#or? elements -E;"
includin MSC" BSC" and B+S. is directly a!!ected.
+$e MS must transmit sinals #it$ reater )o#er" so t$e
averae standby time is s$ortened s$ar)ly.
Therefore, configure T/323 according to resource utilization in various
aspects of networ).
T/323 is configured by equipment room operators. (ts value depends on the
flow and processing capability of each ND. #onfigure T/323 as follows:
Con!iure +2%1% to a reater value -suc$ as 13 $ours" %0
$ours" or even %1 $ours. in areas #it$ $eavy tra!!ic and sinalin
!lo#.
Con!iure +2%1% to a smaller value -suc$ as 2 $ours or 3
$ours. in areas #it$ lo# tra!!ic and sinalin !lo#.
Con!iure +2%1% to 0 in areas #it$ tra!!ic overrunnin t$e
system ca)acity.
To configure T/323 properly, you must permanently measure the processing
capability and flow of each 7D in the running networ), such as:
+$e )rocessin ca)ability o! MSC and BSC
A inter!ace" Abis inter!ace" and /m inter!ace
+$e ca)ability o! 'G< and HG<
(f any of the previously listed NDs is overloaded, you can consider increasing
T/323.
&?. Precautions
T/323 cannot be over small. Gtherwise, the signaling flow at each interface
increases sharply and the -, 0especially handset1 consumes increasing
power. (f the T/323 is smaller than /; minutes 0excluding ;1, the networ)
will be fiercely impacted.
#onfiguring T/323 of different cells in the same location area to the same
value is recommended. (n addition, the T/323 must be consistent with related
parameters of switching side 0smaller than the implicit detach timer at
switching side1.
(f the T/323 of different cells in the same location area is the same, in the cell
reselection, the -, continues to time according the T/323 of the original cell.
(f the T/323 of the original and target cell in the same location area is
different, the -, uses the T/323 of the original cell modulo that of the
serving cell.
ccording to the actual tests of -, in the networ), if the T/323 in the same
location area is different, after the -, performs modulo algorithm based on
behaviors of some users, the -, might power on normally. &owever, the -,
fails in originating location updating, so the networ) identifies it as implicit
detach. Now the -, powers on normally, but a u!er $! po"ered off prompt
appears when it is called.
.1(.7 T3122
&. 9e4inition
T/233 defines the period that the -, must wait for before the second trial
calling if the first trial calling fails. (t aims to avoid congestion of ,*##&
due to repeated trial calling by -, and to relieve system load.
&&. For"at
T/233 ranges from ;s to 3==s. The recommended value is 2;s.
&&&. Con4iguration and &n4luence
The value of T/233 is included in the immediate assignment re6ect message.
fter the -, receives the immediate assignment re6ect message 0no channels
for signaling, interface failure, overload of central processing unit, namely,
#J71, it can send new trial calling request after T/233. T/233 aims to relieve
radio signaling and voice channel resources.
T/233 also help avoid systematic overload. Bhen the #J7 is overloaded, the
system multiplies T/233 by a factor 0determined by processorHoad,upconf1
to increase T/233 through overload control. (n pea) load time, you can
manage networ) access by increasing T/233. Namely, you can increase the
interval between two continuous trial callings to relieve networ) load.
.1(.- T312
&. 9e4inition
T/23. is used in occupation process in asynchronous handover. (t is the time
for -, to receive the physical information send by networ) side.
&&. For"at
#onfigure it to <A=ms when the channel type of assigned channel for
&N*GPD! #G--N* message is ,*##& 0I ,##&1. #onfigure it to
/3;ms in other situations.
&&&. Con4iguration and &n4luence
Bhen the -, sends the &N*GPD! ##D,, message on the primary
*##&, T/23. starts. Bhen the -, receives a J&E,(#H (N"G!-T(GN
message, the -, stops T/23., stops sending access burst, activates the J#&
in sending and receiving mode, and connects to the channel if necessary.
(f the assigned channel is a ,*##& 0I ,##&1, you must enable -, to
receive a correct J&E,(#H (N"G!-T(GN message sent by networ) side
in any bloc). (f T/23. expires 0only in asynchronization1 or the low layer lin)
fails in the new channel before sending the &N*GPD! #G-JHDTD
message, the -, proceeds as follows:
1. Deactivate t$e ne# c$annel
%. <estart t$e oriinal c$annel
2. <econnect to +C'
4. +rier to setu) )rimary sinalin lin?
Then the -, sends the &N*GPD! "(H7!D message on the primary
signaling lin) and return normal operation before trial handover. The
parameters for returning the original channel are those before response to the
&N*GPD! #G--N* message 0such as in encryption mode1.
.1(.1( T11
&. 9e4inition
T22 is an assignment request queue timer.
&&. For"at
T22 is determined by equipment room operators. (t indicates the maximum
queuing delay for assignment request.
&&&. Con4iguration and &n4luence
Bhen the 8,# is sending the ,,(+N-DNT !DS7D,T message, no T#&s
are available. The ,,(+N-DNT !DS7D,T message must be put to a
queue and the 8,# sends the S7D7(N+ (N*(#T(GN message to -,#.
-eanwhile, T22 starts timing.
Bhen the 8,# sends the ,,(+N-DNT #G-JHDTD message 0T#& is
successfully assigned1 or the ,,(+N-DNT "(H7!D message 0T#& is
not assigned1 to -,#, T22 stops timing.
(f T22 expires, the corresponding ,,(+N-DNT !DS7D,T message is
removed from queue and the 8,# sends a #HD! !DS7D,T message with
the cause of no rdio re!ource vil%le to -,# to clear calling.
ssignment queuing helps reduce service re6ection times due to congestion,
so enabling it is recommended in a networ). nyhow, T22 cannot be over
great and it must be configured according to customer habits.
.1(.11 T2((
&. 9e4inition
T3;; is important 0both the -, and base station have T3;;1 at 7m interface
in data lin) layer HJ*m. HJ*m has different channels, such as ,*##&,
"##&, and ,##&, and the transmission rate of different channel is
different, so T 3;; must be configured with different values. The type of the
channels corresponding to T3;; is the value of the T3;;.
&&. For"at
*ifferent channels corresponds different values of T3;;. ccording to the
protocol, when ,J( = ; and ,J( = /, the T3;; of corresponding data lin)
is dependently implemented, depending on delay of synchronous processing
mechanism and process in layer 2 and layer 3.
Table 7-1 Halue rane and de!ault o! eac$ ty)e o! +%00
T3;; -inimum
-aximu
m
*efault
T3;;O,*##&O,J(; =; 2;; <;@ $T = <; T = ms T$
T3;;O"##&O"ullO!ate .; 2;; =;@ $T = =; T = ms T$
T3;;O"##&O&alfO!ate .; 2;; =;@ $T = =; T = ms T$
T3;;O,##&OT#& ,J(; 23; 3;; 2=;@ $T = 2=; T 2; ms T$
T3;;O,##&OT#& ,J(/ 23; 3;; 2=;@ $T = 2=; T 2; ms T$
T3;;O,##&O,*##& =; 2;; <;@ $T = <; T 2; ms T$
T3;;O,*##&O,J(/ =; 2;; <;@ $T = <; T = ms T$

&&&. Con4iguration and &n4luence
T3;; avoids deadloc) in sending data in data lin) layer. The data lin) layer
changes the physical lin) in which error occurs easily to data lin) with no
errors. t the two ends of the data lin) communication system, a confirm4to4
resend mechanism is used. Namely, receiving a message by the receiver must
be confirmed by the sender.
(f it is un)nown that the message is lost, both two ends wait for messages, so
the system confronts a deadloc). Therefore, T3;; is used by the sender.
Bhen T3;; expires, the sender 6udges that the receiver fails in receiving the
message, so it resends the message.
Bhen the sender needs to confirm whether the receiver has received the
message, T3;; starts. Bhen the sender receives the response from the
receiver, T3;; stops. Bhen T3;; expires, the resending mechanism starts. (f
the sender receives no response from the receiver after multiple resendings, it
sends D!!G! (N*(#T(GN 0T3;; expiration1 to layer /.
&?. Precautions
T3;; must be properly configured to ensure a predictable behavior at 7m
interface. The rules for configuring T3;; include:
+$e )otentially,e:istin lost !rames in radio lin? must be
detected as )ossible.
Eecessary retransmission o! !rames must start at t$e
earliest )ossible moment.
>! t$e res)onse is delayed due to /; !ailure" t$e +%00
cannot e:)ire be!ore receivin and )rocessin t$e ne:t !rame
!rom t$e o))osite end.
>! +%00 e:)ires and no ot$er !rames are sent by )re!erence"
t$e related !rames must be resent in t$e messae bloc?.
+ %00 starts immediately a!ter ne:t @',<;ADR,+A,S;ED.
.1(.12 !2((
&. 9e4inition
N3;; is the resending times after expiration of T3;;.
&&. For"at
To configure N3;;, follow rules below:
1. 9$en SA@> N 0 or 2" E%00 de)ends on t$e state and t$e
c$annel used.
9$en multi!rame o)eration is set u)" it ensures a common time
value !or layer % lin? !ailure in all c$annels. For layer % lin?
establis$ment and release" con!iure E%00 to 1.
%. >n timer recovery state" con!iure E%00 as belo#*
1 -SACC'.
%2 -SDCC'.
24 -FACC' o! !ull rate.
%9 -FACC' o! $al! rate.
2. 9$en SA@> is unequal to 0 or 2" con!iure E%00 to 1" as
s$o#n in +able 5,% Situations o! SA@> unequal to 0 or 2.
Table 7-2 Situations o! SA@> unequal to 0 or 2
,J( #hannel
Palid response
delay
-inimum
resending
delay
-aximum resending
delay
Tresp Trmin Trmax Note /
; ,*##& -,: 22 =2 =2
8,,: /3
; "##&$"ull rate : 3< /:
; "##&$&alf rate 2; /. ..
/ ,*##& -,: 22 =2 =2 Note 2
8,,: /3
/
,##&0with
T#&1
3=$23: Note 3 /23 .2< Note 3
The TDMA frame is the measurement unit of values in this table equal to
!"#$"%ms &appro'imately ()%!*ms+
Note 2: (t caters for the process without ,J( ; transmission. Gtherwise, it
does not have a upper limit due to the priority of ,J( ; transmission.
Note 3: Eou can configure it to a greater value only when J#& is
unavailable due to ,J( frame transmission if ,J( = /.
Note /: (t caters only for sending monitoring frames that are available and
without " equal to 2.

&&&. Con4iguration and &n4luence
(f the 8,# fails in receiving lay 3 response message after multiple resending,
it sends the D!!G! (N*(#T(GN message 0T3;; expires1 to layer /. The
8,# ta)es statistics of D!!G! (N*(#T(GN message by corresponding
traffic measurement counter. Bhen T3;; or N3;; is configured to an over
small value, call drop occurs probably due to D!!G! (N*(#T(GN.

5 GSM Radio et!or" #$timi%ation
!adio networ) optimization aims to improve networ) performance and
maximize the benefit of the existing networ) resources through parameter
collection, data analysis, parameter ad6ustment, and necessary technical
means.
"rom the perspective of carriers, they hope to configure the system rationally,
utilize networ) resources to the maximum, enhance networ) economic
benefit, and reduce operation costs through networ) optimization. "rom the
perspective of users, they hope to get satisfactory telecommunication services
in terms networ) stability, speech quality, and so on. Therefore, the core tas)
of radio networ) planning and optimization is to see) a balance among
coverage, capacity, and quality based on rational investment and the limited
frequency resources, thus achieving the best rate of investment return.
/.1 !etwor8 3pti"i*ation Procedure
&ereunder details the procedure:
Eet#or? in!ormation acquisition
Eou are required to confirm the actual engineering parameters and networ)
parameters, survey the local radio environment and hot4traffic spots, and understand
customer requirement.
Data collection
Eou are required to collect G-#! traffic statistics data and alarm data@ drive test
data, and the ob6ective reflection of -,.
Data analysis
Eou are required to analyze networ) performance, networ) parameters, and G-#!
traffic statistics using networ) optimization tools.
Eet#or? tunin
Eou are required to tune engineering parameters and networ) functional parameters.
Eet#or? o)timi(ation re)ort
networ) optimization report must include optimization measures, fulfilled networ)
performance indexes, and suggestions for networ) development.
/.2 !etwor8 3pti"i*ation Tools
/.2.1 Test MS
Test -, is a daily must for engineers to perform networ) test. The test -,
can display the service cell of a mobile telecommunication networ) and the
six neighbor cells. (t can also be used to test networ) parameters. The test -,
can be connected to a computer, so it can collect and analyze data with the
help of drive test software.
The functions of a ,+D- test -, are listed below:
Hie# t$e >MS> o! t$e S>M card
Scan BCC'
The ,+D- test -, can scan the 8##&, and it will provides the !xlev and 8,(#
of the scanned 8##& in each cell.
Hie# net#or? )arameters
The ,+D- test -, can display networ) parameters of the service cell of and the
six neighbor cells when it is idle or in conversation state.
Forced cell selection
The test -, can be forced to reselect the designated cell for dialing test and
handover test.
Forced $andover
The test -, can be forced to hand over to the designated cell to analyze whether the
handover is normal during conversation.
Frequency selection
G@<S !unction test -+$e MSs o! SAG;M A+93 su))ort t$is
!unction.
/.2.2 9ri0e Test So4tware
NT, TD-,, and ,"#G are the +,- drive test software in common use.
+enerally, the drive test software consists of two parts: foreground data
collection software and bac)ground data analysis software.
&. Foreground 9ata Collection So4tware
The foreground data collection software is mainly responsible for collecting
the uplin) and downlin) data at the 7m interface of the +,- mobile
networ). (f the 8T,s and the landform map of the test areas are imported, the
trac) of the drive test, the radio parameters of each test point, and the main
information of the service cell, neighbor cell, and the 7m interface can be
observed.
The functions of foreground data collection software are listed below:
+est !unctions
Test functions include call test, scan test, call interference test, double4networ)
synchronization test, dialing record test, dual4band comparison test, and -,
selection test.
+o)ical test
Topical test is responsible for forced handover and frequency loc)ing 0call1 test,
8##& frequency loc)ing 0standby1 test, forced location update test, and band
loc)ing test.
>nter!erence test
(nterference aims to locate the 8##& same4frequency and neighbor frequency
interference and T#& neighbor frequency interference within the same networ) in
time.
@arameter collection
The field strength of the service cell and the neighbor cells, bit error ratio, frame
error rate, and various radio parameters of the cell need to be collected.
Geora)$ic naviation
The foreground data collection software can be used to display the geographic
navigation through combing the digital map and 8T, resources.
SQ> -S)eec$ Quality >nde:. test
,S( test aims to evaluate the speech quality experienced by mobile users.
+ra!!ic statistics
The foreground data collection software can be used for #ST traffic statistics. (t
enables the call setup, call duration, and call release to be recorded respectively. (n
addition, it is also responsible for recording call drop rate and congestion rate.
Data record
The foreground data collection software can record and store the test data of the test
-, synchronously, and can record and store the scanned data of the :;;-&z and
29;;-&z synchronously.
Scannin test
,canning test aims to test and record the field strength of the channels of the +,-
:;;-&z networ) and +,-29;;-&z networ).
Com)etitive test
The comparison between real4time field strength and speech quality is available. (n
addition, real4time chec) of cell parameters is allowed.
&&. +achground 9ata Analysis So4tware
The bac)ground data analysis software can geographically present the radio
networ) test data and reflect the distribution of networ) parameters on the
electronic map visually. (t can locate the problem cell by fully considering the
drive test data, networ) resource data, digital frequency sweep receiver data,
and +,- signaling characteristics, thus guiding engineers to evaluate and
optimize the networ) rationally and effectively.
The functions of bac)ground data analysis software are listed below:
<adio coverae evaluation and analysis
>nter!erence analysis
Eei$bor cell analysis
'andover analysis
Sinalin analysis
S)eec$ record and analysis
/.2.3 Signaling Analy*er
-42; and U23;= are the +,- signaling analyzer in common use. This
section introduces the functions of the signaling analyzer.
-42; signaling analyzer can test the bis interface signaling messages, the
4+ interface signaling messages, and the ,,VA interface signaling messages
of (,7J, T7J, and (NJ when it is on line. (n addition, it can perform J#-
8D!T 0bit error rate test1 and +,- 8D!T. "urthermore, it can enable the
signaling message files saved during on4line monitoring when it is off4line.
The -42; signaling analyzer has five sub4applications. They are listed
below:
MA,10 control ->t is used to test Abis inter!ace sinalin
messaes" A,G inter!ace sinalin messaes" and bit errors #$en
it is on line..
MAE>+A< AB>S o!!line ->t is used to analy(e t$e Abis
inter!ace sinalin messaes #$en it is o!! line..
MAE>+A< MSC o!!line ->t is used to analy(e t$e A,G inter!ace
sinalin messaes #$en it is o!! line..
GSM,B;<+ o!!line ->t is used to analy(e GSM B;<+ #$en it is
o!! line..
@CM,B;<+ o!!line ->t is used to analy(e @C' B;<+ #$en it is
o!! line..
Bith the help of -42; signaling analyzer, networ) optimization engineers
can collect and analyze bis interface data and 4interface data, view the
whole signaling procedure, and obtain the measurement report, and then
compare the information with the downlin) signals obtained from drive test.
These means can help networ) optimization engineers have an overall
understanding of the operation of the networ). (n this case, the causes and
places for the problems, such as call drop, handover failure, and congestion
can be located.
/.2. !etwor8 3pti"i*ation So4tware
+ood networ) optimization software can wor) as a platform for radio
pro6ects and maintenance personnel. N,T!, &uawei networ) planning
and optimization tool, can deeply analyze +,- networ) by integrating
G-#! traffic statistics, configuration data, alarm data, engineering data, and
so on. (n addition, it also provides the interfaces for networ) planning,
performance analysis, alarm analysis, and geography conditions.
The functions of N,T! are listed below:
>m)ort static tra!!ic statistics data -im)ort t$e tra!!ic
statistics !iles o! t$e maintenance console into database.
>m)ort static tra!!ic statistics data accordin to time
>m)ort dynamic 'ua#ei tra!!ic statistics
>m)ort BSC data con!iuration
Dis)lay t$e eora)$ic vie# o! B+Ss
Cancel" add" and move t$e cells based on t$e eora)$ic
vie# o! B+Ss
<ecord s)ectrum utili(ation and analy(e idle c$annel
number
C$ec? same,!requency BS>C
Analy(e nei$bor cells
Measure distance and a(imut$ anle
Dis)lay naviation #indo# and leend
Query )er!ormance analysis tas?s in common use
Ma?e )er!ormance analysis tas?
Query net#or? o)timi(ation tas?s in common use
Ma?e net#or? o)timi(ation tas?s
@er!orm tas?s and alarms automatically
>m)ort and e:)ort tas?s
Dis)lay tas? result and site vie#
Query tra!!ic statistics inversely !rom t$e cell set selected
by site vie#
Dis)lay t$e tra!!ic tas?s in direct vie#s" includin linear
!iure" column" )ie -su))ort t$e dis)lay o! double y,a:is and t$e
simultaneous dis)lay o! multi)le inde:es8 su))ort %D and 2D..
;:)ort electronic table -;:cel. !or query results.
/.3 !etwor8 Per4or"ance =0aluation
8efore putting networ) optimization into practice, you should have an overall
understanding of the networ) performance. The traffic statistics data, *T
0drive test1 data, and #ST 0call quality test1 data are necessary for networ)
performance acquisition.
/.3.1 9T
*T is used to evaluate the connectivity, coverage, call drop, and voice quality
for the main roads and transportation bac)bones in urban areas. The indexes
for the roads include connected ratio, call drop rate, coverage rate, voice
quality, and so on. The indexes for transportation bac)bones include call drop
rate per )ilometer, coverage rate, voice quality, connected rate, and so on.
Connected ratio
#onnected ratio = total connected times$attempted calls T2;;F
Call dro) rate
#all drop rate = call drop times$total connected times T2;;F
Coverae rate
#overage rate = 0W 4:.d8m test road )ilometers1$total test road )ilometersT2;;F
Hoice quality
ccording to bit error rate, voice quality can be divided into 9 classes, from ; to A.
Dach class matches its bit error rate range.
The calculation of voice quality depends on actual conditions. +enerally, the
following method is in common use:
Poice quality = X!xqual 0class ;1F I !xqual 0class 21F I !xqual 0class 31FT2 I
!xqual 0class /1FT I !xqual 0class .1FT;.9 I !xqual 0class =1FT I !xqual 0class
<1FT;.= I !xqual 0class A1FT;.3
Call dro) ratio )er ?ilometer
#all drop ratio per )ilometer = 0W 4:.d8m test road )ilometers1$total call drop times
/.3.2 CDT
#ST is applied to the important spots in urban areas. (t enables you to
experience the networ) quality from the perspective of users. This section
introduces the indexes used to evaluating the #ST.
Coverae rate
#overage rate = 0W 4:.d8m test points1$total calling test points1T2;;F
Connected ratio
#onnected ratio = total connected times$attempted calls T2;;F
Call dro) rate
#all drop rate = call drop times$total connected times T2;;F
Hoice discontinuity=bac?round noise rate
Poice discontinuity$bac)ground noise rate = 0total voice discontinuity occurrence
times I total bac)ground occurrence times1$total connected timesT2;;F
Ane,#ay audio=ec$o=cross,tal?in rate N -total one,#ay
audio times J total ec$o occurrence times J total cross,tal?in
occurrence times.=total connected timesM100L
#ST enables you to use the -G, 0mean opinion score1 to evaluate the voice
quality from the perspective of people5s ob6ective feeling. The -G, can be
divided into five classes, from 2 to =. "or the evaluation standard, see %able
&2 of Cha"ter GS$ Radio Net'or( )lannin* of GS$ Radio Net'or(
)lannin* and +"timi,ation.
/.3.3 !etwor8 3peration &nde'es
Though *T and #ST can detail networ) problems, they are restricted from
test routes and time. Therefore, *T and #ST cannot test the overall networ).
To give an overall evaluation towards the networ), you should collect as
more networ) operation indexes as possible.
The indexes evaluating networ) operation quality are listed below:
Service access ca)acity inde:es
,ervice access capacity indexes include toll networ) connected ratio, short message
connected ratio, 0+J!,1 J*J activation ratio, and (J telephone connected ratio.
Service $old ca)acity inde:es
,ervice hold capacity indexes include call drop rate, worst cell ratio, traffic call drop
ratio, handover success rate, and short message gateway transit success rate.
Bith the expansion of networ) scale, the networ) structure becomes ever
more complicated. (n this case, networ)s with high performance but low cost
are encouraged.
The indexes on networ) utilization can be used to evaluate whether the cost
to run a networ) is low. These indexes include toll circuit utilization rate,
traffic channel availability, busiest and idlest cell ratio, and so on.
&ereunder introduces the methods to calculate the indexes used to evaluate
networ) operation at the radio side.
<adio connected ratio
!adio connected ratio = 02 M ,*##& congestion rate1T02 M T#& congestion rate 0all
busy11T2;;F
Call dro) rate
#all drop rate = T#& call drop$successful T#& seizures 0all busy1
9orst cell ratio
cell with T#& congestion rate higher than =F at busy hours or a cell with T#&
call drop rate higher than /F is defined as a worst cell. The number of the worst
cells varies with networ) scales.
+ra!!ic call dro) ratio
Traffic call drop ratio = total T#& traffic volume T<;F$total T#& call drop times
0all busy1
'andover success rate
&andover success rate = successful handovers$attempted handoversT2;;F
+ra!!ic c$annel availability
Traffic channel availability = the available traffic channels at busy hour$configured
traffic channelsT2;;F
/. Tra44ic Statistics &nde' Analysis
t the networ) optimization stage, the traffic statistics indexes are the basis
for networ) performance optimization. "or networ) optimization, the UJ(s,
such as congestion rate, call drop rate, and handover success rate, are in
common use. These indexes are the external representation of networ)
quality. The radio coverage quality, channel capacity, and cell parameters are
the internal factor to affect the networ) quality. The traffic statistics analysis
aims to loo) into these internal factors through external factors. ,ince the
mobile networ) is a complex system, you should consider related *T
information, signaling messages, and alarm information for the overall
analysis.
/..1 General Analysis Method
Traffic statistics analysis is performed from 8,# overall performance to cell
performance, from primary indexes to secondary indexes.
"irst you should have a rough understanding of the networ) performance
through 8,# performance analysis. &ere the indexes such as T&# traffic
intensity, T#& call drop rate, T#& congestion rate, and inter4cell handover
success rate should be considered. ttention that in addition to chec) the
percentages of the indexes, you should also chec) the absolute numbers of the
indexes, because the percentages may sometimes hide some cell problems.
fter having understood the indexes about the overall networ) performance,
you should analyze the indexes for each cell if finding abnormal indexes.
"irst you should 6udge if the abnormal index is a common phenomenon or it
is really an abnormal one. (f it is a common phenomenon, you should begin
the analysis from the perspective of coverage, capacity, frequency planning,
and cell parameters. (f it is really an abnormal case, you should register the
corresponding traffic sub4items and analyze them in detail. (n addition, you
should also ma)e an overall 6udgment through collecting the information
about alarm, engineers5 operation, and other external causes. (f the traffic
statistics analysis cannot contribute a correct 6udgment, you should employ
*T equipment and signaling analyzer for help.
/..2 )igh Call 9rop $ate Analysis
(f the uplin) and downlin) quality deteriorates to a level that cannot hold
normal conversation, the conversation will be disconnected. This is defined as
call drop. ,ince the user mobility and radio propagation is uncertain, call drop
always exists in a mobile networ). &owever, optimization measures can be
adopted to reduce the call drop rate.
Bhen the call drop rate of the 8,# overall performance is found abnormal,
you can chec) T#& performance to 6udge whether the call drop is 6ust a
common phenomenon or it is an individual phenomenon. fter that, you can
6udge whether the high call drop rate occurs in several cells or in all the
8T,s. (f the call drop is a common phenomenon, you should ma)e an overall
chec) towards the coverage planning, cell parameter planning, and frequency
planning to analyze whether the lin) budget meet the requirements, whether
the configuration of the path failure counter is rational, and whether the
networ) interference is too great. (n addition, you should also chec) the 8,#
hardware, and then perform drive test to chec) the networ) coverage.
(f it the abnormality is caused by the severe call drop in individual cells, you
should confirm whether it is equipment failure that caused the call drop.
+enerally, alarm messages are always come together with equipment failure,
so you can ta)e equipment failure as a reference.
fter the equipment failure is excluded, you can analyze the call drop rate
from the perspective of interference, coverage, and handover.
1. >nter!erence is divided into u)lin? inter!erence and
do#nlin? inter!erence. Rou can analy(e t$e u)lin? inter!erence
accordin to t$e number o! inter!erence bands into #$ic$ t$e
idle +C's dro). >t is normal t$at t$e idle +C's dro) into
inter!erence band 1 and inter!erence band %. For t$e net#or?
#it$ aressive !requency reuse" it is acce)table t$at t$e idle
+C's dro) into inter!erence band 2. 'ere t$e !requency $o))in"
@BG+ $andover and coverae control must be considered. >! t$e
idle +C's dro) into inter!erence band 4 or above" you s$ould
care!ully c$ec? t$e inter!erence. Generally" t$e inter!erence
#it$in t$e net#or? increases #it$ t$e tra!!ic volume. +$e
increase o! t$e <:qual class can be seen t$rou$ t$e <:qual
measurement tas? and <:lev measurement tas?. +$e )oor
$andover ratio can be seen arisin t$rou$ inter,cell $andover
)er!ormance measurement. >n addition" t$e $andover re,
establis$ment !ailures #ill result in more $andover !ailures.
%. >! t$e coverae is inadequate or it is unbalance on t$e
u)lin? and do#nlin?" t$e call dro) #ill also be resulted. Rou can
Fude i! t$e <:lev is adequate t$rou$ t$e mean <:lev o! t$e
)o#er control measurement tas? and t$e )o#er class. >! t$e
<:lev is still lo# #$en t$e transmitter )o#er reac$es t$e
ma:imum" t$ere are areas #it$ )oor coverae. Mean#$ile" you
can ta?e t$e mean <:qual and <:lev durin call dro) as a
re!erence. +$e distribution o! +A -timin advance. values can
$el) you estimate t$e radius o! subscriber distribution. +$rou$
c$ec?in t$e received c$annel strent$ o! t$e nei$bor cells" you
can analy(e t$e cell coverae. Generally" drive test is needed !or
a detailed analysis.
(f the uplin) coverage and downlin) coverage are unbalance, !" component failure
or cable connection problem will occur. The path unbalance can be seen from the
path balance measurement tas), power measurement tas), and call drop
measurement tas). t this time, the alarm information and user complaint also
deserve your attention.
2. 'andover !ailure #ill )revent t$e MS !rom movin to t$e
best cell. >n t$is case" call dro) may be resulted. >n addition"
cross,cell $andover and taret cell conestion may cause call
dro). +o solve t$is )roblem" you can add nei$bor cell
relations$i) and balance t$e tra!!ic #it$in t$e cells.
The high ,*##& call drop rate analysis is similar to high T#& call drop rate
analysis. cting as the point4to4point signaling channel, the ,*##& is more
sensitive to the interference than T#&. (n this case, the common method to reduce
the call drop rate is to ad6ust the access threshold and reduce interference.
/..3 )igh TC) Congestion $ate Analysis
This section discusses T#& congestion, including the congestion caused by
T#& seizure all busy and the congestion caused by T#& seizure failure.
Bhen the congestion rate of the 8,# overall performance is found abnormal,
you can find out the cells with high congestion rate through chec)ing the
T#& performance statistics. (n this case, you can discover the problems
through analyzing each functional sub4item of the T#& performance statistics
of this cell. (n addition, you should chec) whether there is transmission
problem, cloc) problem, or hardware problem through considering the alarm
information.
(t is a must to analyze the load according to the T#& traffic intensity and the
configured T#& capacity.
1. C$ec? i! t$e +'C conestion rate is caused by +C' sei(ure
all busy t$rou$ analy(in t$e +C' )er!ormance measurement o!
t$e cell. >! t$e conestion is caused by $eavy tra!!ic" you s$ould
)redict t$e real tra!!ic o! t$e cell and c$ec? i! ot$er cells can
s$are t$e tra!!ic. >! it is beyond t$e o)timi(ation ca)ability to
enable ot$er cells to s$are t$e tra!!ic" you s$ould consider
e:)andin t$e ca)acity o! t$e net#or?. +$e adFustment
measures !or tra!!ic balance may not be consistent #it$ t$e
)rinci)le o! minimum radio )at$ loss" so t$ey are a))lied to
emerent causes only. >n most cases" you can balance t$e tra!!ic
t$rou$ adFustin coverae sco)e" adFustin access t$res$old"
adFustin C<A and $andover t$res$old" or enablin load
$andover. >! t$e conestion is not caused by +C' sei(ure all
busy" o on #it$ t$e c$ec?.
%. C$ec? i! t$e +<Os o! t$e conestion cell #or? normally. +$e
damae or )er!ormance decline o! t$e u)lin? c$annels may
)revent t$e MS !rom accessin ot$er cells. >n t$is case" many
cells #ill be sei(ed" #$ic$ #ill cause conestion. +$e incomin
cell $andover )er!ormance measurement #ill s$o# t$at many
$andovers to#ards t$is cell are !ailures. >n t$is case" you s$ould
query t$e statue o! eac$ +<O #it$in eac$ cell t$rou$ queryin
t$e <:lev )er!ormance measurement tas? or <:qual )er!ormance
measurement tas?. >n addition" you s$ould !ind out #$ic$ +<O is
related to t$e abnormality t$rou$ queryin t$e u)lin? and
do#nlin? measurement re)orts o! t$e same +<O.
2. C$ec? i! t$e conestion rate is related to inter!erence"
namely" c$ec? i! any abnormality is )resent !rom t$e
inter!erence band 1 to inter!erence band 1 in t$e tra!!ic
statistics. >! t$e inter!erence is )resent in a cell" t$e call dro)
rate o! t$e cell #ill be $i$" and t$e SDCC' conestion rate #ill
increase accordinly. Moreover" t$e <AC' in t$e random access
)er!ormance measurement may be conested" and t$e
immediate assinment success rate #ill decrease.
4. /nder some conditions" t$e conestion o! some cells is a
result o! lare coverae. >n t$is case" you s$ould analy(e t$e
relations$i) bet#een +A value and <:lev t$rou$ queryin t$e
)o#er control mean level" t$e mean level durin call dro)" and
+A. >n addition" you s$ould also use drive test to de!ine t$e
coverae area o! t$e cell. +$rou$ queryin t$e +C' availability
o! t$e nei$bor cell" you can con!irm i! t$e conestion is caused
by nei$bor cell !ailures. +$rou$ queryin )at$ balance
)er!ormance measurement" you can Fude i! t$e reason !or t$e
+C' sei(ure !ailure is t$at t$e do#nlin? )o#er is reater t$an
t$e u)lin? )o#er.
1. Frequent $andovers can also cause +C' conestion. +$rou$
queryin t$e ratio o! t$e $andovers to t$e call sei(ure successes"
you can c$ec? i! t$e ratio is rational. +$rou$ queryin t$e
incomin and outoin ratio" you can c$ec? i! t$e conestion is
caused by irrational $andover.
/.. )igh S9CC) Congestion $ate Analysis
The ,*##& congestion rate is mainly caused by heavy traffic. "irst you
should define if the congestion is a common phenomenon or if it is 6ust an
individual phenomenon. (f it is a common phenomenon, you should analyze
if the location update timer is irrationally set, and then calculate the ,*##&
capacity to see if it meets system requirement. (f it is 6ust an individual
phenomenon, you should analyze it from the perspective of equipment,
location area, and interference.
1. From t$e )ers)ective o! equi)ment" you s$ould !irst c$ec?
t$e +<O sound ratio in t$e BSC overall )er!ormance
measurement and t$e SDCC' availability in t$e SDCC'
)er!ormance measurement" and t$en c$ec? t$e +C' activation
EAC&=+>M;A/+ in t$e +C' )er!ormance measurement. A!ter
t$at" you can de!ine i! t$e conestion is caused by board
)roblem.
%. C$ec? t$e messaes !or SDCC' bearer location u)date.
>rrational location area )lannin #ill cause !requent location
u)date" #$ic$ #ill result in SDCC' conestion. Rou are required
to analy(e o! t$e ede o! t$e location is set at t$e areas #it$ a
reat number o! subscribers by c$ec?in t$e location area
)lannin and actual drive test. >n addition" you are also required
to c$ec? i! t$e location u)date messaes accounts a too larer
)ercentae o! t$e SDCC' sei(ure requests at t$e ede. +$e
met$od is to query t$e ratio o! t$e success!ul SDCC' sei(ures
-location u)date. to t$e total SDCC' sei(ure successes in t$e
SDCC' )er!ormance measurement.
2. >nter!erence also causes SDCC' conestion. ;s)ecially !or
t$e net#or?s in #$ic$ t$e distance bet#een B+Ss is small and
t$e BCC' !requency is aressive" t$e system may receive more
inter!erence random access sinals. +$e net#or? #ill allocate a
SDCC' !or eac$ random access" #$ic$ causes SDCC' conestion.
>n t$is case" t$e immediate assinment success rate #ill
decrease" t$e )ain success rate #ill decrease" and t$e <AC' in
t$e random access )er!ormance measurement may be
overloaded.
/../ Low )ando0er Success $ate Analysis
The analysis for handover success rate is quite complicated, because it
involves capacity, coverage, cloc), signaling, equipment, and even -,.
1. >! t$e $andover success rate o! all cells is lo#" you s$ould
c$ec? t$e )roblem !rom t$e )ers)ective o! $andover
)arameters" A,inter!ace circuit" and BSC cloc?.
%. Filter t$e cells #it$ )oor $andover. >! a net#or? is run by
t$e equi)ments o! di!!erent carriers" you s$ould c$ec? i! it
intero)erability )roblem by com)arin t$e inter,BSC $andover
success rata #it$ t$e intra,BSC $andover success rate in t$e
$andover )er!ormance measurement. Generally" t$e inter,BSC
$andover success rate is a little lo#er t$an t$e intra,BSC
$andover success rate. >n addition" you need to monitor t$e
sinalin messaes and data con!iuration bet#een BSCs and
analy(e t$e radio lin? budet and cloc? o! eac$ carrier.
2. C$ec? i! any )roblem is )resent at t$e /m inter!ace t$rou$
com)arin t$e $andover success rate and radio $andover success
rate. +$e radio $andover success rate is equal to or reater t$an
t$e $andover success rate. >! t$e $andover success rate is !ar
smaller t$an t$e radio $andover success rate" you s$ould analy(e
t$e round lin? and ca)acity. >! t$e di!!erence bet#een t$e radio
$andover success rate and t$e $andover success rate" you need
to consider t$e inter!erence.
4. Analy(e i! it is incomin $andover !ailure or it is outoin
$andover !ailure t$rou$ queryin t$e incomin cell $andover
success rate and outoin $andover success rate in t$e $andover
)er!ormance measurement. A!ter t$at" analy(e t$e outoin cell
$andover )er!ormance measurement and incomin cell $andover
)er!ormance measurement o! t$e )roblem cell so as to !ind out
t$e incomin $andover !ailure cells !rom t$e outoin cell
)er!ormance measurement. Con!irm i! t$e )oor $andover is
caused by taret cell conestion t$rou$ analy(in t$e Yincomin
cell $andover !ailuresY" Y+C' tra!!ic intensityY" and Y+C'
conestion rate -all busy.Y o! all t$e incomin $andover !ailure
cells.
1. C$ec? i! any equi)ment !ails t$rou$ queryin t$e +<O
sound ratio" +C' availability" and +C' activation EAC&=+>M;A/+
o! t$e taret cell. Analy(e i! t$e +<O )er!ormance decreases
t$rou$ queryin t$e <:lev )er!ormance measurement o! t$e
taret cell.
3. C$ec? i! any round lin? equi)ment !ails t$rou$ queryin
t$e A,inter!ace !ailures and t$e round lin? brea?s durin +C'
sei(ure.
Bhen the microwave is used for the transmission or during inter48,# handover, the
cloc) deviation is another cause for poor handover. nd this can be proved by the
intra48,# handover failures. "or the cells where the cloc) synchronization is
unavailable, the 8,(# cannot be decoded, so the handover can never occur. (n this
case, you need to chec) if the cloc) is normal and analyze the call drop rate.
(f these two causes are excluded, you need to ma)e ad6ustment from the perspective
of coverage and interference.
To reduce call drop rate and enhance handover success rate, you can leave a margin
for the !xlev and !xqual during handover. (f the !xlev of a cell is lower than
4:;d8m during handover, you should chec) the mean !xlev and T value of T#&
call drop in the call drop performance measurement and analyze drive test to see if
the coverage distance of the cell is too long and if the signal is not strong enough.
"or the networ)s in which better cell algorithms are enabled, you should chec) the
'attempted handovers 0better cell1'. (t is better that the percentage it accounts <;F of
the handover causes.
The interference will also affect the handover success rate. Bhen the interference is
present, the voice quality will decrease and the call drop rate will increase.
&andover problems are rather complicated. To solve the problems arising in actual
wor), you are supposed to integrate the methods introduce above, the signaling
analyzer, equipment condition, and drive test into consideration.
/./ !etwor8 Pro5le" Solutions
/./.1 Co0erage Pro5le"s
&. Solutuon Procedures
&'ro%lem( t$e cover)e i! %ecomin) !mller fter t$e B*S i! en%led+
fter a 8T, has run for a period of time 0for example, half years1, the
coverage of the 8T, may become smaller or even dead zone may appear due
to various causes. (n this case, the system performance will be affected. The
shrin) of the coverage is not only related to the technical indexes 0such as the
8T, sensitivity and power1, but also related to the engineering quality,
geographic factors, and the electromagnetic environment.
The factor concerning the 8T, problems are as follows:
+ransmitter out)ut )o#er decrease
<eceive sensitivity decrease
Antenna a(imut$ anle c$ane
Antenna tilt c$ane
Antenna ain c$ane
Feeder loss
Cou)ler loss
9or?in band c$ane
@ro)aation environment c$ane
Diversity e!!ect c$ane.
Eou can chec) the problem according to the following procedure:
1. C$ec? t$e conditions around t$e B+S antenna
Eou are required to chec) if there are other antennas 0such as micro antenna1,
decorations, billboard, trees, or glass walls standing around the 8T, antenna. These
barriers may exert a negative effect against the antenna reception and transmission,
thus affecting the coverage of the 8T,. (n this case, you can tune the azimuth angel
of the corresponding antenna or change the antenna height.
%. C$ec? t$e c$ane o! t$e )ro)aation environment
The change of the propagation environment of the electromagnetic wave will
wea)en the signals received by radio terminals. Dspecially for mountains, the
propagation of the electromagnetic waves depends on the reflection of mountains.
"or example, the change in the vegetation of the mountain will reduce the coverage
of the 8T,. (n addition, the climate and other natural factors also have some effect
against the electromagnetic waves. The propagation loss varies with wood intensity,
season, and so on. The maximum loss can reach /; d8. (f new buildings prevent the
propagation of the electromagnetic waves and wea)ens the signals, the areas in the
remote cannot be covered, so the subscribers cannot en6oy the service. Dspecially the
high buildings near the 8T, have a great effect against the propagation of
electromagnetic waves.
2. C$ec? i! t$ere is standin #ave alarm and main diversity
rece)tion alarm at t$e o)eration and maintenance console
This problem can be chec)ed according to the standing wave alarm messages and the
diversity reception alarm messages. (f the alarms of this )ind occur, you should
chec) the corresponding antennas and feeders.
4. C$ec? i! t$e standin #ave ratio is smaller t$an 1.1L
The tolerance of the standing wave alarm threshold of the #*7 or D*7 is great.
Therefore, after chec)ing that the set4top power is normal, you can further chec) if
the standing wave is smaller than 2.=F. (f the standing wave ratio is abnormal, you
need to chec) if the water has penetrated into the antenna or feeder connector, or if it
is lightening protector problem.
1. C$ec? i! t$e to#er am)li!ier #or? normally
#hec) if tower amplifier alarm is present at the operation and maintenance console.
+enerally, the problems are the low noise amplifier was damaged or the water has
penetrated into the amplifier. The amplifier alarm always comes together with the
damage of the low noise amplifier. (f the water has penetrated into the tower
amplifier, no alarm will be generated, but the !" loss is great. (n this case, the
receiver sensitivity will decrease dramatically.
3. C$ec? t$e enineerin )arameters -includin antenna tilt
and a(imut$ anle.
The increase of the antenna tilt or the deviation of the azimuth angle will reduce the
coverage of the 8T,. Therefore, antennas must be firmly fixed so that they can stand
strong wind and storms.
5. C$ec? t$e set,to) out)ut )o#er o! t$e transceiver
"irst you should chec) if the lines are well connected, and then chec) if the set4top
power is normal. (f it not normal, you should replace the problem hardware.
8. C$ec? i! t$e receiver sensitivity is normal
#hec) if the coverage distance is shortened by the low receiver sensitivity. (n
addition, you can monitor the messages at the bis interface and find out the
relationship between level and bit error rate. fter that, you can get the value of the
level when the bit error rate is 3F. This means, however, only applies to the situation
that when the receiver sensitivity drops dramatically.
9. C$ec? i! t$e )arameters a!!ectin t$e coverae are
rationally set
10. C$ec? i! t$e $i$ bac? noise in t$e coverae area is caused
by inter!erence and )oor electromanetic environment.
&Cover)e pro%lem cu!ed %, B*S e-pn!ion+
(f the coverage of the 8T, shrin)s after expansion, in addition to ma)ing the
previous chec)s, you are supposed to chec) the following items.
1. C$ec? i! t$e combiner ?ee)s t$e same be!ore and a!ter
e:)ansion
The loss of different combiners varies greatly Therefore, the combiner configuration
deserves special attention during 8T, expansion. (f different combiners are a must,
you should fully communicate with customers.
%. C$ec? i! t$e antennas are rationally selected
,uitable antennas must be selected for pro6ect installation and networ) planning so
that the best coverage can be achieved. (t must be pointed out that you should use
zero4point filling antenna or the electrical title antenna when the antenna height is
great. (n addition, omni antennas cannot be widely used for the large area coverage.
(n this case, the coverage problem can be solved by directional antennas.
2. C$ec? i! t$e installation o! t$e ne#ly,added antennas are
quali!ied
Eou should first chec) if the design of the antenna height, azimuth angle, and
antenna tilt is qualified. +enerally, the important coverage areas cannot be bared by
tower. -eanwhile, the important coverage areas cannot be perpendicular to the
diversity direction of the antenna. (n this case, the antenna diversity effect can be
excavated to the maximum. To reduce the coverage shadow caused by the tower,
you should pay attention to the distance between the antenna and the tower.
-oreover, the pole of the omni antenna and the !" part of the antenna cannot be
overlapped.
4. C$ec? t$e )osition o! t$e BCC' transmitter antenna
,ince the tower effect is present, the 8##& transmitter antenna must be installed at a
side of the important coverage area. (n this case, the coverage shadow can be
avoided. To prevent the assignment failure caused by the inconsistence of the 8##&
coverage and T#& coverage, you can use the concentric channel allocation
algorithm. (n addition, the important coverage area cannot be perpendicular to the
diversity direction of the antenna.
1. C$ec? i! t$e tilts and t$e a(imut$ anles o! t$e directional
dual transmitter antennas are consistent #it$ eac$ ot$er
(f the tilts and azimuth angles of the directional dual transmitter antennas are
inconsistent, call drop, assignment failure, and handover failure will easily occur. (n
this case, the coverage area of the 8T, will become small. (n addition, since the
tower effect is present, the 8##& transmitter antenna must be installed at a side of
the important coverage area. (n this case, the coverage shadow can be avoided.
-oreover, the important coverage area cannot be perpendicular to the diversity
direction of the antenna.
3. C$ec? t$e set,to) out)ut )o#er o! various +<Os i! t$e
sc$eme !or t$e ma:imum coverae is used.
Bhen the maximum coverage is pursued, the T!%s are required to be combined in
various ways. (n this case, the coverage distance of the 8##& will be longer than
that of the T#&. s a result, the T#& assignment failure will be caused, so the
concentric technology is needed. The channel assignment failure caused by low
transmit level in the inner circle and the channel congestion in the external circle can
be avoided if the T values of the inner circle and the external circle are correctly set
and allocated to the inner circle and external circle according to the right priority.
&Cover)e pro%lem! cu!ed %, B*S !"p or con!truction+
1. C$ec? i! t$e a(imut$ anle and t$e antenna $ei$t are t$e
same be!ore and a!ter t$e B+S s#a)
(f all the antenna and feeder components are newly constructed, the old 8T, can
only be swapped after the new antenna is installed. Therefore, the azimuth angle and
the antenna height may be different from that of the old antenna. (n this case, the
coverage area may decrease. s a result, you should chec) if the azimuth angle and
the antenna height are the same before and after the bas station swap.
%. C$ec? antenna tilt )roblems caused by net#or? s#a)
+enerally, the tilt must )eep the same. (f you need to control the coverage area due
to new 8T,s are added to urban areas, you can consider increasing the tilt.
2. C$ec? i! t$e set,to) )o#er o! t$e s#a) B+S is t$e same as
t$at o! t$e old B+S.
4. C$ec? i! t$e receiver sensitivity o! t$e B+S is normal.
1. C$ec? i! it is t$e inter!erence or t$e )oor electromanetic
environment t$at ma?es t$e bac? noise o! t$e #$ole area too
$i$.
3. C$ec? i! any standin #ave alarm messae or diversity
rece)tion alarm messae is enerated !or antenna and !eeder at
t$e o)eration and maintenance console.
5. C$ec? t$e )arameters t$at #ill a!!ect coverae are
rationally set.
8. C$ec? i! t$e installation o! t$e antennas is quali!ied a!ter
t$e B+S $as been enabled or s#a))ed.
9. C$ec? i! t$e ri$t ty)e o! antenna is selected.
10. C$ec? t$e )osition o! t$e BCC' +<O transmitter o! t$e omni
dual transmitter antenna.
11. C$ec? i! t$e tilts and t$e a(imut$ anles o! t$e t#o
directional antennas ?ee) t$e same a!ter t$e directional dual
transmitter antenna is used.
1%. C$ec? i! t$e antennas and !eeders o! t$e cell are inversely
connected.
12. C$ec? i! t$e to#er am)li!ier #or?s normally.
14. C$ec? t$e set,to) )o#er !or various +<Os #$en t$e
con!iuration sc$eme !or t$e ma:imum coverae is )ursued.
&&. Pro5le"s A44ecting Co0erage and Solutions
&Antenn "ter penetrtion+
(t is quite accidental that the water penetrates into the antenna. Bater
penetration means that the water enters the !" internal channel. (n this case,
the voltage standing wave ratio of the antenna will increase@ the antenna loss
will increase, the coverage area will decrease@ or event the power amplifier
will be disabled.
&Antenn p!!ive intermodultion+
The passive intermodulation of the antenna and various connectors will cause
interference. The exclusive method can be used for the chec). That is, you
can connect the antenna feeders of the neighbor cells where there is no
interference to the test cell. (f any problem is found, you should change the
antenna.
&.mproper ntenn !election+
+enerally, if the antenna height exceeds =;m and if the first zero point under
the main antenna beam is not filled, the 'shadow under tower' may occur.
That is, the area under the tower cannot be covered by signals. (n this case,
you should select the antenna with zero point filling function.
(f three4sector directional antennas are used for vast coverage, the antennas
must have a high gain and their half power angle must be greater than :;
degrees. (f the half power angle is small, the gain of the two neighbor sectors
will be low. (n this case, the coverage radius is small.
(f the antenna tilt is great, the all mechanical tilt antenna is not a suitable
choice. (n this case, you should select the fixed 'electrical tilt I mechanical
tilt' antenna or the 'continuous ad6ustable electrical tilt 0; to 2; degrees1 I
mechanical tilt' antenna.
s the frequency reuse becomes more aggressive, the front4to4bac) ratio of
the antenna may not meet the requirement of a single 8T, or several 8T,s.
Therefore, you should select the antennas with greater front4to4bac) ratio.
&*o"er effect )in!t /mni ntenn rdition+
+$e to#er e!!ect aainst omni antenna radiation deserves enou$ attention. >t
is $ard to estimate t$e damae o! t$e omni antenna directional diaram caused
by t$e to#er. +$e damae varies reatly #it$ t$e distance bet#een t$e to#er
and t$e antenna.
>! t$e antenna is installed on t$e to#er and metal tube" you s$ould )ay s)ecial
attentions to t$e !ollo#in items*
+$e metal tube and t$e e!!ect radiation )art o! t$e antenna
cannot be overla))ed.
+a?e measures to avoid installin t$e #$ole antenna on t$e
metal tube.
>! t$e antenna is installed on t$e to#er" ma?e sure t$at t$e
distance bet#een t$e antenna and t$e nearest end o! t$e to#er
is reater t$an 3 #avelent$s.
+$e omni dual transmitter tec$noloy is not recommended.
+$e antenna must be )er)endicular to 1=8 o! t$e $al! )o#er
beam #idt$ at least.
&Directionl ntenn in!tlltion pro%lem+
Two problems may occur for directional antenna installation:
+$e antenna is inversely or #ronly connected.
+$e a(imut$ anles and t$e tilts o! t$e transmitter antenna
and t$e receiver antenna are inconsistent and or t$e error is
reat.
Dngineering causes are the explanations of the two problems. +enerally, the error
scope of the azimuth angle cannot exceed = degrees, and that of the tilt cannot
exceed ;.= degrees. (f the error is too great, the coverage of the transit antenna and
that of the receiver antenna will be different. (n this case, it is hard to ma)e calls the
coverage edges.
@roblems concernin t$e diversity distance bet#een t$e
transit antenna and t$e receiver antenna or t$e isolation
bet#een t$e antennas and to#er.
The coverage of the antenna will be affected if the diversity distance between the
transmitter antenna and the receiver antenna or the isolation between the antennas
and the tower is not great enough. "or +,- :;;-&z system, the diversity distance
between the transmitter antenna and the receiver antenna is required to be greater
than .m. "or +,- 29;;-&z system, it is required to be greater than 3m. The
antenna mount must be at least 2.=m away from the tower. -eanwhile, the antenna
mount must be installed within the .=4dregree protection areas of the lightening
protector.
+$ere are s$ado#s in coverae areas.
Bhen installing a directional antenna, you should ma)e sure that there is no shadow
within the coverage area. +enerally, if there are huge barrier, such as high buildings
and mountains, around the 8T,, shadows may appear. (f you intend install the 8T,
on the roof of a high building, you should install it at the edges of the building so as
to avoid the shadow. ,ince the environment around the roof is quite complex, the
antenna height must be great enough. (n this case, however, you should consider the
ability of the antenna to stand the wind and storm.
&/mni ntenn in!tlltion pro%lem+
+$e radiator o! t$e omni antenna is barred by antenna )ole.
The coverage will be affected if the radiator of the omni antenna is barred by antenna
pole. +enerally, there is a 6ac)et installed at the bottom of the omni antenna and the
6ac)et is used to connect the omni antenna and the antenna pole. "rom the
perspective of installation, the top of the 6ac)et must be at the same level with or
higher than the top of the pole@ otherwise the radiation will be affected.
+$e )roblems concernin antenna diversity distance and
isolation bet#een antenna and to#er.
(f the antenna diversity distance or the isolation between antenna and tower is not
great enough, the coverage will be poor. (f the antenna diversity distance is too
small, it will reduce diversity gain. (n this case, the receiver sensitivity will reduce.
Though the tower effect against the omni antenna radiation is unavoidable, you can
increase the isolation between the antenna and the tower to reduce the effect.
(t is suggested that the isolation between the omni antenna and the tower is greater
than 3m, the horizontal diversity distance of the :;;-&z omni antenna is greater
than .m, and the horizontal diversity distance of the 29;;-&z antenna is greater
than 3m.
+$e omni antenna is not )er)endicular to t$e $ori(ontal
)lane.
(f the omni antenna is not perpendicular to the horizontal plane, the antenna
directional diagram will be distorted in the coverage area. (n this case, the coverage
of the antenna will be affected.
(t is suggested that installation plane of the antenna mount be perpendicular to the
horizontal plane. (f the mount extends beyond the tower, ma)e sure that the mount is
still in the protection areas of the lightening protector. +enerally, the areas .=4dregee
under the lightening protector top are the protection areas.
&Connection pro%lem! of ntenn nd feeder0 com%iner nd !plitter0 nd
CD1+
(f various connectors of the antenna and feeder system are not connected according
to requirement, the performance of the antenna and feeder system will be affected. (n
this case, the coverage area of the 8T, will also be affected.
9ater )enetration occurs at t$e various connectors o! t$e
antenna and !eeder system.
(f water has penetrated into the connector and feeder, the standing wave ratio will
increase. (n this case, the coverage area will be affected.
Harious connectors are not ti$tened.
(f the connectors for set4top 6umpers, for the cables from T!% boards to combiner
and splitter, and for various !" cables are not tightened, both the reception
performance and the transmit performance of the system will decrease. (n this case,
the coverage area and the conversation quality will be affected.
+$e transmitter antenna and t$e receiver antenna are
inversely connected due to inconsistent con!iuration o! t$e set,
to) Fum)er and data.
+$e connection bet#een t$e Fum)er and !eeder is not ti$t"
#$ic$ results in $i$ loss and standin #ave ratio. >n t$is case"
t$e coverae #ill be a!!ected and inter!erence #ill be caused.
&*o"er mplifier pro%lem+
9ater )enetration #ill increase t$e loss" deteriorate t$e
standin #ave ratio" and decrease t$e receiver sensitivity.
+$e damae o! t$e GEA -it is in t$e to#er am)li!ier. #ill
decrease t$e ain or even decrease t$e ain to a neative value.
+$e in)ut end and t$e out)ut end o! t$e to#er am)li!ier are
inversely connected. >n t$is case" t$e to#er am)li!ier #ill be
s$ort,circuited. >! t$e s$ort circuit lasts !or a lon time" t$e !ront
module #ill be damaed.
&B*S front module pro%lem+
>solator )roblem
Du)le:er and ot$er !ilter damae
Standin #ave ratio error alarm
GEA -lo# noise am)li!ier. damae
Go# +<O or am)li!ier out)ut )o#er
&'rmeter confi)urtion pro%lem+
The parameters affecting coverage are listed below:
+<O )o#er class
+o#er am)li!ier attenuation coe!!icient
MS ma:imum transmit )o#er control )o#er
MS minimum <:lev
<AC' minimum access t$res$old
&&&. Co0erage Cases
C!e 1( 1!e do"n tilt omni ntenn to improve cover)e
&'ro%lem de!cription+
(n a suburban area, the omni antenna with a gain of 22d8i is used for the
8T,. This coverage distance can reach :)m in plain environment. &owever,
the coverage in the area near the 8T, is poor. The !xlev in the small town
9;;42.;;m away from the 8T, is about 4:;d8m.
&'ro%lem nl,!i! nd !olution+
Gn4site survey shows that the antenna height is too great. The height of the
tower on which the antenna is installed =;m. -oreover, the tower is
established on a small mountain, so the town is 23;m below the antenna. The
first 6udgment is that the phenomenon of 'shadow under tower' has been
caused.
"urther analysis of the collected data finds that omni antenna is used for the
8T,. The antenna gain is 22d8i, and the vertical half power angel is A
degrees. (f the valid antenna height is 23;m, the half power points of the
antenna ma6or lobe are scattered in the area about 3;;;m away from the 8T,.
Therefore, this town is not in the coverage area of the 8T,.
Through chec)ing the fluctuation of the !xlev according to the drive test
map, engineers found that this town locates within the radiation area of a zero
power point of the 8T,. &owever, the town is too far away from the
mountains around, so it cannot get the signals reflected by the mountains.
Therefore, the !xlev in this town is quite slow.
A!ter $avin re)laced t$e antenna #it$ an omni antenna #it$ 1 derees o! t$e
do#n tilt anle" enineers retested t$e <:lev and !ound t$at it increased by 11,
%0 dB in t$e areas 2?m #it$in t$e B+S. >n some areas" t$e <:lev is increased by
20 dB. +$ere!ore" t$e coverae $as been im)roved remar?ably.
Case 2, &"proper installation o4 o"ni antenna has e44ect against the
co0erage
&'ro%lem de!cription+
new 8T, has been enabled for a local networ). 7sers complain that the
coverage area become smaller after that. "or the low narrow areas 3)m away
from the 8T,, the !xlev is already lower than 4:;d8m.
&'ro%lem nl,!i! nd !olution+
Through surveying the environment around the 8T,, engineers found that the
ma6or transmitter antenna and the diversity receiver antenna are installed in a
plane parallel to the road. pparently, this )ind of installation does not meet
the criteria.
+$e correct #ay is to install t$e maFor transmitter antenna and t$e diversity
receiver antenna in a )lane )er)endicular to t$e road. >n addition" t$e maFor
transmitter antenna must be located at one side o! t$e road.
C!e 3( .mproper confi)urtion of dt cu!e! poor cover)e
&'ro%lem de!cription+
*uring the optimization for a place, engineers found that the signals at a
section of the road in the suburban area are quite poor. The measured !xlev is
4:=d8m.
&Cu!e nl,!i!+
This section locates in the suburban area and is about /)m away from the
urban area. There is no abrupt change in terms of landform within this
section. Theoretically, the !xlev here should be about 49;d8m, so the
difference between the theoretical !xlev and the measured !xlev is great.
ccording to the frequency sweep test, the strength of the "a signal is about
4:=d8m, and the strength of the "b signal is about 49;d8m. "or this section,
it is covered by three cells of the 8T, and 8T, 8 that are installed in the
urban area 0the 8##& frequencies are "a and "b1. (n addition, a cell of the
8T, # installed at the remote suburban area also covers the section 0the
8##& frequency is "c1.
Through chec)ing data, engineers found that the "b is not included in the
neighbor channel numbers of the 4/ cell in the 82. Bhen the -, moves
from the urban area to the suburban area, it will choose 4/ cell to camp on,
because the "b is not configured in the neighbor channel numbers. (n this
case, the -, cannot reselect the 84/ cell to camp on. (n the cell neighbor
relationship list, the 4/ cell and 84/ cell cannot wor) as the neighbor cell for
each other, and the "b is not configured in the neighbor channel numbers of
the 4/ cell listed in 83. Therefore, in conversation mode, the -, cannot
)eep the conversation in 4/ cell. Bhen it arrives at this section, it cannot
hand over to the 84/ cell. Therefore, the signals are poor, so is the voice
quality.
&Solution+
Dnable the 4/ cell and 84/ cell to wor) as neighbor cell for each other.
C!e 4( .rrtionl B*S !"p ffect! cover)e
&'ro%lem de!cription+
(n an urban area, a 8T, must be swapped for the building on which the 8T,
was installed were to be moved. #onsidering that coverage for the scenic spot
3)m away 0the scenic spot locates behind a hill1 is poor, so engineers
intended to install the 8T, on the top of the hill. Gn the top of the cell, the
whole city and the scenic spot can be seen. &owever, after the 8T, swap,
users complaint that there were no signals in the indoor environment of the
cells near the site where the old 8T, was installed.
&Cu!e nl,!i!+
The buildings of the resident area are densely distributed and the average
height is 9m. 8efore the 8T, swap, the cell used for this area is only 2;;m
away, and the antenna height is 2=m. Therefore, the indoor conversation
quality can be guaranteed. fter the 8T, swap, however, the cell used this
area is 2.9)m away, and the antenna height is /;m. (n this case, the signals
are quite wea) when arriving at the bottom of resident area, though the
signals falling at the top of the building is good. To solve this problem, you
can only increase the output power of the transmitter antenna or increase the
antenna gain. &owever, the coverage is still not to users5 satisfactory even
increase the antenna height to /;m. Therefore, when swapping or
constructing a 8T, at the densely populated area, you should pay attention to
the following items:
>t is suested t$at t$e distance bet#een t$e B+S and t$e
resident area is equal to or smaller t$an 110m" ot$er#ise t$e
coverae !or t$is area #ill become #ea? dramatically.
+$e antenna o! t$e s#a))ed B+S cannot be too reat. >! t$e
B+S is installed amon resident buildins" t$e antenna $ei$t is
suested to be 5,10m. >! t$e B+S is installed beyond t$e
resident buildins and t$e buildins are $i$" you can increase
t$e antenna $ei$t accordinly.
Rou can solve t$e )roblems concernin cross area coverae
can t$rou$ controllin t$e )o#er class o! t$e B+S" tunin t$e
a(imut$ anle o! t$e antenna" or tunin t$e tilt anle o! t$e
antenna.
&Solution+
ccording to on4site survey, engineers found that the indoor signals of this
area are too poor to hold the conversation. This area can be seen clearly from
the tower on which the 8T, is installed. The distance between the 8T, and
the area is only 2.9)m, and between them are vast farmlands. To solve this
problem, you can attempt to tune the azimuth angle and the tilt angle of the
antenna. (f the coverage is not yet improved, you can use the following
methods:
<e)lace t$e common antenna used !or t$is cell -its ain is
11dBi. #it$ t$e $i$,ain antenna used !or t$e scenic s)ot -its
ain is 18dBi.
+$ere are 4 +<Os in t$is cell" all in SC/ mode" re)lace t$e
SC/ mode #it$ t$e dual,CD/ mode.
fter the above methods are done, the antenna gain for this area can be
increased by <d8. fter the antenna replacement, you need to tune the
antenna tilt for the best coverage.
Through retesting the indoor signal level, engineers found that it increased by
<423d8. nd even the common -, can )eep normal conversation.
/./.2 &nter4erence Pro5le"s
(nterference is a )ey factor affecting networ) performance, including
conversation quality, call drop, handover, congestion, and so on.
&. &nter4erence Sources
>n t$e mobile telecommunication system" #$en t$e B+S is receivin t$e sinals
!rom a remote MS" it #ill not only be inter!ered by ot$er telecommunication
equi)ments" but also it #ill be inter!ered by t$e ot$er B+Ss and MSs #it$in t$e
system.
'ereunder introduces t$e inter!erence sources a!!ectin t$e GSM system.
>ntra,net#or? inter!erence
(f the frequencies are improperly planned, or the frequency reuse is too aggressive,
intra4frequency interference or neighbor cell interference will be caused.
<e)eater inter!erence
t the early stage of networ) construction, repeaters are widely used for extending
the coverage distance of the networ). &owever, if the repeaters are improperly
planned, the networ) will be interfered.
(f the repeaters are not installed according to requirement, that is, there is not enough
isolation left between the donor antenna and the subscriber antenna, the 8T, to
which the repeaters attach will interfered.
"or the repeaters enabling broadband non4linear amplifier, the intermodulation
indexes are far greater than that required in the protocols. (n this case, the greater the
power is, the greater the intermodulation will be. Therefore, the 8T, near the
repeaters will be interfered.
>nter!erence !rom ot$er bi,)o#er telecommunication
equi)ments
These equipments include radar, analog 8T,, and other telecommunication
equipments using the same band.
'ard#are )roblems
T!% problem: (f the performance of the T!% decreases, the system may be
interfered.
#*7 problem or splitter problem: ctive amplifier is used in the #*7 splitter and
splitter module. Bhen any problem occurs, the system may also be interfered.
,tray and intermodulation: (f the out4band stray of the power amplifier or the T!%
of the 8T, go beyond requirement, or the isolation of the transmission and the
reception of the #*7 duplexer is too small, the connection channel will be
interfered. -eanwhile, the passive equipments, such as the feeder and the antenna,
will generate intermodulation.
&&. &nter4erence Positioning and =li"ination
&'o!itionin) nd elimintion procedure+
1. Find out t$e inter!erence cell accordin to &@>
(f the call drop rate, handover success rate, traffic volume, congestion rate, and
interference band of a cell deteriorate to a bad level abruptly, it means that
interference may exist in the cell.
(n this case, you can also chec) the historical record of operations made in this cell.
"or example, chec) if the hardware and software of the 8T, has been added or
increased and if the data of the 8T, has been modified. +enerally, the appearance of
interference is related to these operations.
(f these parameters are not ad6usted, the interference may be from the hardware itself
out outside factors. (n this case, you are suggested to chec) if it is hardware problem.
(f it is not, you should chec) outside factors.
%. C$ec? AMC alarm
,ometimes high call drop rate, low handover success rate, and high congestion rate
may be related to equipment problems. (n this case, you can chec) G-# alarm
records. These records are related to the deterioration of these indexes.
2. C$ec? !requency )lannin
(f the interference is doubt in a cell, you can chec) the frequency planning for the
cell and the neighbor cells of the cell. "or this chec), you are required to ma)e clear
the distribution of the antennas, find out the azimuth angle of each cell, draw the
topology, and mar) the 8##&$T#& channel numbers. -eanwhile, you are also
required to compare the planned channel numbers with the configured channel
numbers in the 8,#.
ccording to the accurate frequency planning topology, you can ma)e sure if the
intra4frequency interference or neighbor frequency interference is present in the
networ).
4. C$ec? cell )arameter con!iuration
The cell parameters, such as #!G, threshold, handover duration, neighbor cell
relationship, and so on, may have interference against the system.
(f the #!G is set to a great value, the -, may be guided to an idle cell whose level
is lower than its surrounding cells. Gnce the conversation is started but the #$(
cannot meet the threshold requirement 023d81, interference will be caused.
(f neighbor cells are missing, the -, cannot hand over to a cell with better signal
level and quality. (n this case, the interference will also be generated. (f the handover
threshold and the J$N are too great, the handovers between cells are unavailable. (f
the J$N is too small, however, it will result in frequent handover. (n this case, both
the call drop rate and the system load will be increased.
1. Drive test
*rive test is an effective method to position the interference. There are two drive test
methods: idle mode test and dedicated mode test.
"or idle mode test, the test equipment can test the signal level of both the signal level
and the neighbor cells. (n addition, the test equipment can also perform the
frequency sweep test for the designate channel numbers or bands. (n this case, the
interference caused by cross4cell coverage signals can be discovered.
"or dedicated mode test, the test equipments can test the signal level of the service
cell and neighbor cells, the !xqual, the T, and so on. (f the !xlev is equal to or
greater then 49;d8m and the !xqual is equal to or greater than < in an area, it can be
confirmed that the interference exists in the area. ,ome test equipment can display
the "D! 0frame error rate1. +enerally, if the "D! is greater or equal to 3=F, the
conversation will not be continuous. That is, the interference exists.
3. >nter!erence elimination
Eou can eliminate the interference according to the above chec)ed results, and then
evaluate the elimination through UJ( and drive test.
&2rd"re pro%lem po!itionin) nd elimintion+
Bhen the interference is doubted in a cell, you should first chec) if the 8T,
where the cell locates wor)s normally. (n the remote end, you should chec) if
there is antenna alarm, T!% alarm, or 8T, cloc) alarm generated. (n the near
end, you should chec) if there is antenna problem, water penetration, feeder
06umper1 damage, #J7 problem, T!% problem, wrong 6umper connection or
cloc) problem occurred.
Antenna )er!ormance decline
ntenna a passive component and its damage probability is small. &owever, if the
antenna is damaged or its performance declines, the voice quality will become poor.
Antenna connector )roblem
+,- !" signals are micro wave signals. (f the connections between T!%, #*7,
feeder, and antenna have any problem, both the standing wave ratio and the
intermodulation will increase. (n this case, the interference will be resulted.
>nverse antenna connection
The inverse antenna connection is a commonly seen problem. (f the antenna is
inversely connected, the channel numbers used by the cell and the planned channel
numbers are completely inconsistent. (n this case, intra4frequency interference, inter4
frequency interference, and handover difficulty will be resulted. Dspecially for the
networ)s that have inadequate frequency resource, the inverse antenna connection
has great effect against networ) quality.
Xum)er )roblem
-any 6umpers locate between antennas, so they are often wrongly connected. (n this
case, high call drop rate will be resulted.
+<O )roblem
(f T!% problems occur, the interference will increase, the coverage distance area
will decrease, and the access is difficult.
Cloc? !ailure
(f the cloc) deviation is too great, it is hard for the -, to loc) the frequencies of the
8T,, so the handover failure always occurs, or the -, cannot camp on any cell of
the 8T,. (n addition, if the cloc) deviation is too great, the 8T, cannot understand
the signals of the 8T,, which will result in bit errors. &owever, the cloc) failure will
not really introduce interference, but it is the transmission errors that ma)e the voice
quality decrease.
Conclusion
ny problem concerning the T!%, #*7, feeder, antenna, 6umper, and connector
may cause interference or call drop. Therefore, if interference appears, you should
chec) the hardware of the 8T,. (n addition, 8T, cloc) failure will also cause
interference and call drop.
(t is easy to solve the hardware problems through changing the boards or ad6usting
traffic data. (f there is spectrum analyzer available, you can position the problem
more efficiently. Dspecially when the interference appears without any modification
of networ) data, you should focus on chec)ing the hardware.
&.ntr34et"or5 .nterference+
The intra4networ) interference is mainly from intra4frequency interference
and neighbor cell interference. Bhen #$( is smaller than 23d8 or the #$ is
smaller than 4<d8, the interference is unavoidable. &owever, the aggressive
frequency reuse technology will increase of the occurrence probability of
interference.
Same,!requency and nei$bor !requency inter!erence
(n +,- system, the frequency reuse is unavoidable. Bhen the frequency reuse
distance of two cells using the same frequency is smaller than cell radius, same4
frequency interference will be easily caused. Jast experiences show that the
frequency reuse must be avoided in many cases.
The interference against the uplin) channel numbers can be 6udged by the
interference band data in the traffic statistics.
"or the interference against the downlin) channel numbers, the existing drive
test equipments can be indirectly used to measure if the same4frequency
interference is present. "irst you should loc) the test -, in the service cell
and enable ma)e the -, wor) in conversation mode during drive test. (f you
find that the !xlev in an area is high but the !xqual is low, it is li)ely that the
same4frequency is present in this area.
>nter!erence caused by cross coverae
(n a properly designed networ), each cell covers the areas around the 8T, only and
the -, camps on or holds conversation in the nearest cell. #ross coverage means
that the coverage of a cell is too large and the cell can cover the areas under the
control of other 8T,s. (f cross coverage occurs, irrational traffic absorption,
interference, call drop, congestion, and handover failure may arise.
>nter!erence caused by aressive !requency reuse
#apacity and quality always contradicts to each other. (n urban areas, the aggressive
frequency reuse technology must be used for the number of subscribers in urban
areas are great. (n this case, the networ) quality will surely decrease. (n the areas
where 8T,s are irrationally distributed, the aggressive frequency reuse technology
may cause the collision of same frequency and neighbor frequencies.
>nter!erence caused by re)eater
(t is convenient to use repeater for special coverage. &owever, if a repeater is not
qualified or it is not properly installed, it will cause interference.
>nter!erence caused by outside environment
Gutside environment, such as TP station, big4power radio station, micro wave, radar,
high voltage wire, analog 8T,, and so on, will cause interference.
&&&. &nter4erence Cases
C!e 1( .nterference cu!e %, ntenn performnce decline
&'ro%lem de!cription+
There are = 8T,s in a county. The configuration type is ,.$.$. or ,<$<$<. The
interference band = reaches 2= according to the T#& performance
measurement of the most cells. There is no alarm found at the G-#.
&'ro%lem po!itionin) nd !olution+
1. +$rou$ monitorin and reisterin t$e inter!erence band
tra!!ic statistics !or t$e )roblem cells all day" enineers !ound
t$at t$e inter!erence band 1 mostly a))eared at day time" and it
seldom a))eared at early mornin.
%. +$rou$ sendin t$e idle B/<S+S o! all t$e B+Ss" enineers
!ound t$at t$e inter!erence bands o! t$ese cells a))eared in t$e
early mornin. >! t$e sendin o! t$ese idle B/<S+S sto))ed"
t$ese inter!erence bands disa))eared. +$ere!ore" it can be
)roved t$at t$e inter!erence came !rom t$e net#or?. >t is not
related to ot$er telecommunication equi)ments.
2. +$e !requencies and ot$er data #ere not adFusted be!ore
t$e inter!erence a))eared" so t$e inter!erence is not related to
t$e !requency )lannin.
4. +$rou$ surveyin t$e <OM test inter!ace o! t$e CD/ usin
t$e s)ectrum analy(er durin t$e tra!!ic )ea? at day time"
enineers !ound t$at t$e broadband inter!erence #as stron and
t$e bac? noise #as risin.
1. +$ere #as no inter!erence in one cell" but t$e inter!erence
in anot$er t#o cells #as stron. +$rou$ re)lacin t$e antenna
!eeder o! t$e cell #it$ no inter!erence #it$ t$e antenna !eeders
o! t$e cells #it$ inter!erence and sendin idle B/<S+S" enineers
!ound t$at t$e inter!erence #ent #it$ t$e antenna !eeder.
+$ere!ore" it can be decided t$at t$e )roblem occurred at t$e
antenna and !eeder system.
3. +$rou$ c$anin t$e antenna" enineers !ound t$at t$e
inter!erence #ent #it$ t$e antenna. +$ere!ore" t$e )roblem is
li?ely )resent at t$e antenna.
5. +$rou$ re)lacin t$e antenna #it$ dual )olari(ation
antenna" enineers !ound t$at t$e stron inter!erence
disa))eared immediately. +$rou$ re)lacin t$e old antenna o!
anot$er B+S #it$ a ne# one" enineers !ound t$at t$e
inter!erence also disa))eared.
C!e 2( Cll drop cu!ed %, intr3net"or5 interference
&'ro%lem de!cription+
Customers in a )lace com)laint t$at call dro) $a))en !requently.
FPro5le" analysis and solutionG
1. +$rou$ a care!ul test" enineers !ound t$at t$ere #ere 1%
c$annel numbers at$erin at t$e call dro) s)ot and <:lev
reac$ed ,52dBm. 9$en t$e MS sei(ed c$annel number 11" t$e
inter!erence !rom c$annel number 11% caused t$e call dro).
%. +$rou$ testin t$e CG> o! c$annel number 1% usin test MS"
enineers !ound t$at t$is c$annel number #as one o! t$e BCC'
number o! D2.
2. +$rou$ surveyin B+S D" enineers !ound t$at t$e antenna
o! D2 is installed at t$e to) o! a buildin. >n addition" a $ouse
made o! lass #as !ound 8m a#ay and 4m under t$e antenna.
;nineers tested t$at t$e sinal strent$ near t$e antenna #as
about ,41dBm" and t$e sinal strent$ at near t$e lass #as
,20dBm" #$ic$ #as beyond t$e e:)ectation o! enineers. >n !act"
t$e cause #as t$at t$e sinals re!lected by t$e lass #ere
re!lected to t$e call dro) s)ot.
4. >t is suested to c$ane t$e antenna installation )lace and
c$annel number. Rou s$ould interc$ane t$e c$annel number
111 and c$annel number 114 o! B+S A and increase t$e do#n tilt
anle o! A2 cell. >n addition" to avoid t$e inter!erence caused by
c$annel number 111 a!ter t$e interc$ane" you s$ould adFust t$e
direction o! c$annel number 112 o! C1 cell.
1. +est s$o#s t$at everyt$in is normal a!ter t$e adFustment.
+$e c$annel number 112 o! B+S C $as no e!!ect aainst c$annel
number 114. And t$e call dro) disa))ears.
C!e 3( .nterference cu!ed %, repeter
&'ro%lem de!cription+
7sers in an area complaint that the -, cannot seize a channel to hold
conversation, or the noise is great after channel seizure and the channel and
the -, signal is strong. Two 8T,s are installed in this area. The antenna
azimuth angle of cell2 rightly directs to the north. 8efore user complaint, the
8T, in this area ran normally and the networ) indexes met the requirement.
fter the problem arisen, the traffic volume of the two 8T,s dropped sharply
from the perspective of traffic statistics indexes. (n addition, the traffic
volume of cell2 and cell/ also dropped sharply. Though the signals for the
conversation were strong, the voice quality was quite poor. ccording to
traffic statistics, the interference bands of the four cells were of level /, level
., and level =, and :=F of the channels were interfered. (n addition, other
channels were interfered to some extent. &owever, no alarm messages were
generated at the G-#.
&'ro%lem nl,!i! nd !olution+
1. Accordin to user !eedbac?s" t$e )ossible reasons include
transmission )roblem" antenna !eeder )roblem" $ard#are
)roblem" intra,net#or? inter!erence" and outside inter!erence.
%. +$e u)lin? inter!erence sinals in t$e nort$#est direction
mi$t stron. +$ere!ore" cell1" cell%" and cell2 o! t$e t#o B+Ss
#ere inter!ered" in #$ic$ cell1 and t$e cell2 #ere seriously
inter!ered.
2. +$rou$ on,site dialin test" enineers !ound t$at it #as
$ard to ma?e calls in t$e areas covered by cell1 and cell2. ;ven
i! a call #as )ut t$rou$" t$e voice quality #as quite )oor. >n
addition" t$e voice #as discontinuous and t$e inter!erence #as
stron. +$rou$ usin MS to call a !i:ed )$one" enineers !ound
it #as $ard to $ear t$e voice clearly. An t$e contrary" t$ey could
$ear t$e voice !rom t$e !i:ed )$one clearly. +$is $as )roved t$e
above analysis. +$at is" t$e inter!erence mi$t be !rom t$e
outside" or t$e standin #ave )roblem #as occurrin at t$e
antenna -!rom t$is )ers)ective" it can be Fuded t$at t$e
inter!erence e:isted on t$e u)lin? only..
4. +$rou$ usin antenna !eeder analy(er to )er!orm on,site
test" enineers !ound no )roblem #as e:istin at any B+S. A ne#
re)eater #as !ound in t$is area" and it #as located t#o
?ilometers a#ay !rom t$e B+S in t$e nort$#est direction.
Moreover" t$e inter!erence a))eared Fust #$en t$e re)eater #as
enabled. An,site test !ound t$at t$e B+S became normal state
once t$e re)eater s$ut do#n" and t$e inter!erence bands also
became normal" so did t$e call. >! t$e re)eater #as enabled"
$o#ever" it #as $ard to ma?e calls and t$e inter!erence #as
stron. At last" t$e areement to s$ut do#n t$e re)eater #as
reac$ed. A!ter t$at" t$e conversation became normal.
C!e 4( Micro"ve interference
&'ro%lem de!cription+
*uring networ) maintenance, through analyzing 8,# traffic statistics,
engineers found that the call drop rate of the cell3 and cell/ of a ,3$3$3 8T,
arisen abruptly, and the value even reached 3;F at some time.
&'ro%lem nl,!i! nd !olution+
1. +$rou$ c$ec?in BSC tra!!ic statistics" enineers !ound t$at
t$e number o! idle +C's #as increasin at t$e inter!erence
bands 2,1 around 8*20. Around 10*00" t$e idle +C's #ere !ound
at t$e inter!erence band 4 and inter!erence band 1. Around
%%*00" t$e idle +C's #ere !ound at t$e inter!erence band 1.
+$ere!ore" it could be Fuded t$at t$e inter!erence e:isted.
%. Because t$e B+S ran normally" t$e )roblems cannot be
related to !requency )lannin.
2. Accordin to t$e +<O manaement messaes" enineers
!ound t$at t$e inter!erence e:isted at t$e !our boards o! t$e
cell% and cell2 o! t$e B+S. Because t$e )robability !or t$e !our
boards to be damaed simultaneously is quite small" t$e +<O
)roblem can be e:cluded. 'o#ever" one board #as c$aned in
case o! abnormal conditions" but t$e inter!erence #as not
eliminated.
4. +$rou$ c$ec?in all t$e BSC tra!!ic statistics data"
enineers !ound t$at cells o! t$e B+Ss near t$e B+S and t$e cells
o! t$e cell% and cell2 o! t$e B+S #ere inter!ered to some e:tent.
>n addition" enineers also !ound t$at t$e SDCC's -13 SDCC's in
total. o! t$e seriously,a!!ected cells #ere sei(ed at sometimes.
'o#ever" t$e number o! subscribers determined t$at t$e
)robability !or all t$e SDCC's to be sei(ed simultaneously is
quite small. +$ere!ore" it could be Fuded t$at t$e u)lin? #as
inter!ered by outside !actors. 'o#ever" t$e inter!erence mi$t
be related to direction only.
1. +o !urt$er )osition t$e )roblem" enineers interc$aned t$e
Fum)ers o! cell1 and cell2 at t$e set to). >n t$is case" t$e
inter!erence #as !ound at cell1" but t$e inter!erence #as
disa))earin !rom cell2. +$ere!ore" t$e inter!erence #as not
related to c$annel number.
3. Because t$e inter!erence #as not related to c$annel
number" it mi$t be t$e bi,)o#er sinals t$at caused t$e
inter!erence.
5. +$rou$ usin t$e s)ectrum analy(er to )er!orm t$e
measurement at t$e out)ut inter!ace o! t$e B+S s)litter"
enineers !ound t$at t$e bi,)o#er sinals e:isted at t$e 904M'(
c$annel number -it $as an interval o! 1M bet#een t$e used
c$annel number. For t$e B+S #$ere t$e inter!erence #as stron"
t$e sinal level can reac$ as $i$ as about ,%1dBm. For ot$er
B+Ss" t$e sinal level #as about ,10dBm. +$ere!ore" it could be
Fuded t$at it #as t$is sinal t$at a!!ected t$e B+S.
8. +$rou$ usin t$e s)ectrum analy(er to scan t$e areas near
t$e B+S" enineers !ound t$at t$ere #as a micro#ave antenna
out)uttin bi,)o#er at t$e c$annel number 904.
9. +$e inter!erence disa))eared a!ter t$e micro#ave
equi)ment #as s$ut do#n.
/./.3 Call 9rop Pro5le"s
"or the +,- networ), call drop is users5 ma6or worry and the call drop rate is
an important index evaluating networ) quality.
&. Call 9rop $esasons and Solutions
i# Cll drop due to cover)e re!on!
&6e!on nl,!i!+
Discontinuous coverae -dead (one.
"or a single 8T,, the quality of the signals at the edge of the station is quite poor, so
the -, cannot hand over to another cell. (n this case, the call drop occurs.
(f the landform of the coverage areas is complex or fluctuates greatly, or the radio
transmission environment is complex, the signals will be barred. (n this case, the
coverage is discontinuous and call drop will occur.
@oor indoor coverae
(f the buildings in an area are densely populated, the signal attenuation is great. nd
if the walls of the buildings are thic), the penetration loss is great and the indoor
signal level is low. (n this case, the call drop may easily occur.
>solated island e!!ect
s shown in Drror: !eference source not found, the service cell forms an isolated
island due to various reasons 0for example, the power is too great1. (n this case, the
-, still seizes the signals of the service cell after moving to cell #, but the cell
does not define the neighbor cell #. t this time, if the -, still performs the
handover according to the neighbor cell 8 provided by neighbor cell , it cannot
find a suitable cell. (n this case, the call drop will occur.
Small coverae
(f the coverage is too small, the hardware equipment of a cell may fail. "or example,
the antenna is barred or the T!% failure occurs 0the power amplifier part1.
&7ud)ment met$od!+
"irst you should find out the areas where the coverage is inadequate
according to user complaints, and then you should perform the drive test in a
larger scope to chec) if the signal level and the handover are normal and if
the call drop exists. (n addition, you can employ the traffic statistics recorded
at the G-# to chec) the 8,# overall call drop rate and find out the cell with
great call drop rate. "urthermore, you can still ma)e the analysis and
6udgment by referring to other traffic statistics items. &ereunder lists some
ones:
@o#er control )er!ormance measurement -to c$ec? i! t$e
mean u)lin? and do#nlin? sinal strent$ is too lo#.
<:lev )er!ormance measurement -to c$ec? i! t$e ratio o!
t$e lo# <:lev is too reat.
Cell )er!ormance measurement=inter,cell $andover
)er!ormance measurement -to c$ec? i! t$e level class and t$e
mean <:lev are too lo#.
Call dro) )er!ormance measurement -to c$ec? i! t$e sinal
level is too lo# durin call dro) and i! t$e +A value is normal
be!ore call dro).
De!ined nei$bor cell )er!ormance measurement -to
)osition t$e cell #it$ lo# mean sinal level.
/nde!ined nei$bor cell )er!ormance measurement -to
c$ec? i! t$e unde!ined nei$bor cells #it$ $i$ sinal level e:ist.
@o#er control )er!ormance measurement -to measure t$e
reatest +A value bet#een t$e MS and B+S.
&Solution!+
1. C$ec? t$e areas #$ere t$e coverae is inadequate
Eou can find out the area where the coverage is inadequate through drive test. "or an
isolated 8T, or the 8T,s installed in mountain areas that cannot form seamless
coverage, you can add 8T,s to these areas for seamless coverage. Gr you can
improve the coverage through other means. "or example, you can enhance the
maximum transmit power of the 8T,, change the antenna azimuth angle, change the
antenna tilt, change the antenna height, and so on. (n addition, you should also
analyze if the call drop is caused by landforms. +enerally, call drop can easily occur
at tunnels, big shopping mar)et, underground railway entrance, underground par)ing
lot, and low4lying places. (n this case, you can use the micro cell to solve the
coverage problem.
%. ;nsure indoor call quality
To ensure indoor call quality, you should ma)e sure that the outdoor signals are
strong enough. To strength the outdoor signals, you can increase the maximum 8T,
transmit power, change the antenna azimuth angle, change the antenna tilt angle, and
change the antenna height, and so on. (f the indoor call quality is still not improved
remar)ably, you can consider adding 8T,s. "or improving the indoor coverage of
office buildings and hotels, you can consider using the indoor antenna distribution
system.
2. For t$e cells $avin no nei$bor cells" you can con!iure t$e
nei$bor cells !or t$e cell so as to reduce t$e call dro) rate. +o
eliminate t$e isolation island e!!ect" you can reduce t$e tilt
anle o! t$e B+S.
4. ;liminate $ard#are )roblems
Eou can chec) if there are hardware problems and if the coverage area is too
small through drive test. (f the call drop rate of a cell arises dramatically but
all other indexes are normal, you should chec) if the neighbor cells of this
cell wor) normally. 0+enerally, the downlin) problems may occur. "or
example, T!% problem, diversity unit problem, and antenna problem are
commonly seen. (f the uplin) fails, the outgoing handover failure rate of the
old cell will be high.1
ii# Cll drop due to $ndover re!on!
&6e!on nl,!i!+
>rrational )arameter con!iuration
(f the signal level at the cross4area of two cells is quite low, the level of the handover
candidate cell is too low, and if the handover threshold is too low, some -,s will
hand over to the neighbor cell when the signal level of the neighbor cell is higher
than that of the service cell. (f the signal level of the neighbor cell deteriorates
dramatically 6ust after the handover, the call drop will occur if no suitable cell is
available for the handover.
>ncom)lete nei$bor cell de!inition
(f the neighbor cell definition is incomplete, the -, will hold the conversation in the
existing cells until it moves beyond the edges of the cell but cannot hand over to a
stronger cell. (n this case, the call drop will occur.
Eei$bor cells #it$ same BCC' and same BS>C e:ist.
+ra!!ic conestion
(f the traffic is unbalance, no T#& will be available in the target cell. (n this case, the
handover failure will occur.
B+S cloc? lost sync$roni(ation
(f the 8T, lost synchronization, the frequency offset will go beyond the requirement.
(n this case, the call drop will occur if handover fails.
+2102 e:)iry
The T/2;/ will be started when the networ) sends a handover command.
7pon the reception of the message to complete the handover or the message
to remove the command, the T/2;/ will stop. T/2;/ is used to hold the
channel long enough for the -, to return to the old channel. (f the T/2;/ is
set to a too small value, the -, cannot necessarily return to the old channel.
(n this case, call drop may occur during handover.
&7ud)ment met$od!+
Eou can 6udge if the cells with low handover success rate, frequent re4
establishment failures, and high call drop rate through analyzing traffic
statistics indexes. fter the 6udgment, you can find out what causes the
handover. "or example, the uplin) and downlin) !xlev can cause the
handover@ the uplin) and downlin) !xqual can cause the handover@ power
budget can cause handover@ call direct retry can cause handover@ and also
handover can be initiated by traffic reasons.
To chec) if the 8T, cloc) runs normally, you can chec) if the any alarm is
generated for the 8T, cloc). (f necessary, you must correct the 8T, cloc) to
eliminate cloc) problem. Eou can chec) if there is handover problem through
drive test. (f there is a problem cell, you should perform drive near the cell for
several times. &ereunder lists the indexes concerning call drop:
>nter,cell $andover )er!ormance measurement -!requent
$andover !ailures" !requent re,establis$ment !ailures.
>nter,cell $andover )er!ormance measurement -!requent
$andovers" $i$ re,establis$ment rate.
/nde!ined nei$bor cell )er!ormance measurement -t$e
unde!ined nei$bor cell level and t$e number o! measurement
re)ort o beyond t$e standard.
Autoin cell $andover )er!ormance measurement -!ind out
t$e reasons !or lo# outoin cell $andover !rom t$e $andover
taret cell.
Go# incomin cell $andover success rate8 t$e cell $andover
)arameters are im)ro)erly set8 t$e taret cell is conested.
+C' )er!ormance measurement -t$e $andover times are not
)ro)ortional to t$e +C' call sei(ure successes8 t$e $andover
$a))ens too !requent.
&Solution+
1. C$ec? t$e )arameters a!!ectin t$e $andover. For e:am)le"
you can c$ec? t$e $ierarc$ical and level settin" eac$ $andover
t$res$old" eac$ $andover $ysteresis" $andover time" $andover
duration" t$e minimum access level o! t$e $andover candidate
cell" and so on.
%. >! t$e call dro) is caused by unbalance tra!!ic volume or i!
t$e call dro) occurs due to no $andover c$annel is available at
t$e taret B+S" you can solve t$e )roblem by adFustin t$e
tra!!ic volume. For e:am)le" you can adFust t$e )roFect
)arameters" suc$ as antenna tilt and antenna a(imut$ anle" to
control t$e coverae sco)e o! a cell. +o balance t$e tra!!ic
volume" you can use C<A to uide t$e MS to cam) on ot$er idle
cells" or you can set t$e $ierarc$ical and level )riority to uide
t$e MS to $and over to t$e idle cell. >n addition" you can solve
t$e )roblem by e:)andin t$e +<O directly.
2. Calibrate t$e )roblem B+S cloc? to enable t$e
sync$roni(ation o! t$e cloc?.
iii# Cll drop due to interference re!on!
&6e!on nl,!i!+
(f the -, receives strong same4frequency interference signals or strong
neighbor frequency interference signals in the service cell, the bit error rate
will deteriorate. (n this case, the -, cannot demodulate the 8,(# code of the
neighbor cells accurately, or it cannot receive the measurement report from
the -, correctly. s a result, the conversation will be interfered, the call
quality will become poor, and call drop will occur.
&7ud)ment met$od!+
The interference may be from the networ) itself or the outside networ), or it
may exist in the uplin) signals or downlin) signals. The following methods
can be used to position the interference.
Find out t$e cells mi$t be inter!ered t$rou$ c$ec?in
tra!!ic statistics.
@er!orm t$e call drive test !or t$e areas t$at mi$t be
inter!ered and c$ec? t$e u)lin? and do#nlin? inter!erence
accordin to user com)laint. Rou can !ind out i! t$ere is a )lace
#$ere t$e sinal is stron but t$e call quality is )oor t$rou$
drive test tools. >n addition" you can use a test MS to )er!orm
dialin test to c$ec? i! a c$annel number is inter!ered.
C$ec? t$e !requency )lannin to see i! same,!requency
inter!erence and nei$bor !requency inter!erence occur in t$e
area #$ere t$e !requency is im)ro)erly )lanned.
AdFust t$e c$annel numbers t$at mi$t be inter!ered to see
i! t$e inter!erence can be avoided or reduced.
;:clude t$e inter!erence caused by equi)ment.
>! t$e )revious met$ods !ail to eliminate t$e inter!erence"
you can use t$e s)ectrum analy(er to scan t$e !requencies to
!ind out t$e inter!ered c$annel number and t$e inter!erence
source.
&ereunder lists several traffic statistics indexes used for interference analysis:
>nter!erence band
Eou can chec) the uplin) interference through analyzing the interference band in the
traffic statistics. (f an idle channel appears at the interference bands /4=, the
interference is present. (f it is intra4networ) interference, it will increase as the traffic
volume grows. +enerally, if it is outside interference, it is not related to traffic
volume. (t must be pointed out that the interference bands are reported to the 8,# by
the 8T, T!% channel 0when in idle mode1 through !" resource indication
messages. (f the current channel is busy and cannot report !" resource indication
message, you must consider the traffic volume for the measuring the interference
bands.
<:lev )er!ormance measurement
The !xlev performance measurement provides the matrix relationship between the
signal level and quality. (f the signal level is high but the quality is poor, it means
that the interference 0same4frequency interference, intra4frequency interference, and
outside interference1 is present at the channel numbers of the T!% board.
@oor quality $andover ratio
The cell performance measurement, inter4cell handover performance measurement,
or the outgoing cell handover performance measurement records the outgoing
handover attempt times. (f the frequent handover is caused by poor signal quality, it
means that the interference is present.
<:qual )er!ormance measurement
(t is related to the mean !xlev and !xqual during call drop.
Call dro) )er!ormance measurement
(t records the mean !xlev and !xqual during call drop.
Frequent $andover !ailures and !requent re,establis$ment
!ailures
(t means that the interference may be present in the target cell.
&Solution!+
1. C$ec? t$e inter!ered road and t$e distribution o! sinal
quality t$rou$ dive test. As !ar as t$e actual conditions are
concerned" you can adFust t$e B+S transmit )o#er and antenna
tilt o! t$e related cells or adFust t$e c$annel number )lannin to
avoid t$e inter!erence.
%. /se D+O tec$noloy" !requency $o))in tec$noloy" )o#er
control" and diversity tec$noloy
These technologies can be used to reduce the system noise and enhance anti4
interference capacity of the system. *T% is divided into uplin) *T% and downlin)
*T%. (n this case, the transmit time can be reduced and the interference level of the
system can also be reduced. &owever, you should ad6ust the *T% according to the
actual radio environment and the neighbor cell relationship. Bhen signals received
by the -, are poor, the use of the *T% will result in call drop. (f the downlin) *T%
is enabled, the 8T, will increase its transmit power after the call is established.
*uring the conversation, however, the 8T, will reduce its transmit power. (n this
case, the interference against other 8T,s will be reduced. (f the interference is
present near the 8T,, the downlin) *T% will deteriorate the conversation quality.
Bhen the 8T, reduces its transmit power, the conversation quality will decrease or
the call drop may even occur in the areas where the !xlev is low but the interference
signal is strong.
2. Solve t$e equi)ment )roblems" suc$ as t$e sel!,e:citation
o! +<O boards and t$e antenna demodulation inter!erence.
4. ;:clude t$e outside inter!erence.
iv# Cll drop due to ntenn feeder re!on!
&6e!on nl,!i!+
;nineerin )roblem may be one o! t$e reasons. For
e:am)le" i! t$e transmit antennas bet#een t#o cells are
inversely connected" t$e level o! t$e u)lin? sinal #ill be !ar
)oorer t$an t$at o! t$e do#nlin? sinal. >n t$is case" t$e call
dro)" one,#ay audio" and call di!!iculty #ill be !ound in t$e
areas !ar a#ay !rom t$e base station.
>! )olari(ation antennas are used" a cell $ad t#o sets o!
antennas. >! t$e tilt anles o! t$e t#o antennas are inconsistent
#it$ eac$ ot$er" t$e call dro) #ill occur.
(f a directional cell has a master antenna and a diversity antenna, the 8##& and the
,*##& of the cell may be transmitted through the two antennas respectively. (f the
tilt angles of the two antennas are different, the coverage scope of the two antennas
will be different. (n this case, the -, can receive 8##& signals but cannot seize the
,*##& when starting a call. Thus the call drop is resulted.
>! t$e a(imut$ anles o! t$e t#o antennas are inconsistent
#it$ eac$ ot$er" call dro) may also occur. +$at is" t$e MS can
receive t$e SDCC' sinals" but it may be assined #it$ t$e +C'.
>n t$is case" t$e call dro) #ill occur.
+$e )roblems concernin antenna !eeder #ill also cause call
dro). For e:am)le" i! t$e antenna is damaed" or #ater
)enetrates into t$e antenna" or connector )roblem is )resent"
t$e transmitter )o#er and t$e receiver sensitivity #ill decrease.
>n t$is case" t$e call dro) #ill occur. +o con!irm t$e )roblem"
you can c$ec? t$e standin #ave ratio.
&'ro%lem po!itionin) nd !olution+
1. C$ec? i! any alarm concernin t$e combiner" CD/" to#er
am)li!ier" and standin #ave is enerated and c$ec? i! t$e B+S
boards are normal in t$e AMC.
%. Analy(e i! t$e )at$ balance is reali(ed accordin to tra!!ic
statistics.
2. Furt$er analy(e i! t$e )at$ balance is reali(ed t$rou$
monitorin t$e messaes sent across t$e Abis inter!ace.
4. @er!orm drive test and dialin test. Durin drive test" you
can c$ec? i! t$e BCC' numbers o! t$e service cell are consistent
#it$ t$e )lanned ones" namely" i! t$e transmit antenna o! t$e
antenna is correctly installed.
1. C$ec? and test t$e on,site B+Ss. 'ere t$e installation o! t$e
a(imut$ anle and t$e tilt anle o! t$e antenna must be
c$ec?ed. >n addition" you s$ould also c$ec? i! t$e !eeder and
Fum)er are correctly connected" i! t$ere is connector )roblem"
and i! t$e !eeder is damaed. Furt$ermore" you must still c$ec?
i! t$e standin #ave is normal.
3. Xude i! it is B+S $ard#are t$at causes )at$ unbalance and
call dro). +o solve $ard#are )roblem" you can eit$er c$ane t$e
com)onents t$at may $ave )roblems or disable ot$er +<Os in t$e
cell. +o !ind out t$e )roblem +<O" you can )osition t$e )roblem
t$rou$ dialin test. Ance a )roblem $ard#are com)onent is
!ound" you must re)lace it #it$ a sound one. >! no sound one is
available" you must s$ut do#n t$e )roblem $ard#are com)onent
to )revent it !rom a!!ectin net#or? quality.
&ereunder lists several traffic statistics items for path balance analysis:
@at$ balance measurement -to analy(e i! t$e )at$ balance
is reali(ed.
Call dro) )er!ormance measurement -to analy(e t$e u)lin?
and do#nlin? level and quality durin call dro).
@o#er control )er!ormance measurement -to analy(e mean
<:lev on t$e u)lin? and do#nlin?.
v# Cll drop due to trn!mi!!ion re!on!
(f the transmission quality across the bis interface and 4interface may be
not good and stable, call drop may occur. The following methods can be used
to solve the problem:
1. C$ec? t$e transmission alarm and board alarm and analy(e
i! t$ere is transmission intermittent and board !ailure.
%. C$ec? t$e transmission c$annel" test t$e bit error rate"
c$ec? %M connectors" and c$ec? i! t$e equi)ment roundin is
rational to ensure stable transmission quality and reduce call
dro) rate.
2. C$ec? t$e tra!!ic statistics to see i! t$e !requent call dro) is
caused by transmission )roblems. ;s)ecially you s$ould c$ec?
+C' )er!ormance measurement" because it can indicate i! t$e A,
inter!ace !ailures durin +C' sei(ure is normal" i! t$e +C'
utili(ation is normal" and i! t$e round lin? call dro) times are
reat.
vi# Cll drop due to prmeter re!on!
&ere you should focus on chec)ing if the parameters related to call drop are
irrationally set. (f the following parameters are not irrationally set, the call
drop may be resulted.
<adio lin? !ailure counter
This parameter acts on the downlin). Bhen the -, fails to decode the ,##&, it
will use this parameter to decide when to disconnect the call. (f this parameter is set
to a too small value, the radio lin) failures will easily occur and cause call drop. "or
dead zones or the areas where the call drop frequently happens, you are
recommended to set this parameter to a greater value.
Bhen changing the radio lin) failure counter, you should change the corresponding
T/2;:. The T/2;: must be set to a value great enough for the -, to detect a radio
lin) failure. "or example, if the value of the radio lin) failure counter is 2< 0about 9
seconds1, the value of T/2;: must be greater than 9 seconds 0you can set T/2;: to :
seconds or 2; seconds1.
SACC' multi!rame number
This parameter acts on the uplin). The 8T, uses this parameter to notify the radio
lin) connection failure message to 8,,. The 8,, side 6udges the radio lin) failure
according to the bit error rate on the uplin) ,##&. (f this parameter is set to a too
small value, the radio lin) failure will happen frequently and the call drop rate will
be resulted.
Access control )arameters
The access control parameters include the -inimum !#& !xlev, !#& busy
threshold, and so on. (f the access control parameters are irrationally set, the call
drop will be easily resulted.
+2101" +2105
T/2;2 is started when the 8,# sends a #&NNDH #T(PTD message to the
8T,. (t stops when an D,T8H(,& (N*(#T(GN message is received. This timer
monitors the immediate assignment procedure. (f T/2;2 expires, the allocated
channels will be removed.
T/2;A is started when the 8,# sends an ,,(+N-DNT #G--N* message to
the 8T,. Gnce the 8,# receives the ,,(+N-DNT #G-JHDTD message from the
8T,, this timer will reset. T/2;A is used to hold the channels long enough so that the
-, can return to the old channel. Gr it can also be used by the -, to release a
channel.
(f the two timers are set to a too small value, the system will not have enough time to
send the ,,(+N-DNT #G-JHDTD message to the 8,#. (n this case, the call
drop will occur if the timer expires.
+%008 E%00
T3;; is an important timer used for the HJ*m 0Hin) ccess Jrocedure on the *m
channel1. (t prevents the deadloc) from occurring when the data is transferred across
the data lin) layer. (n +,- system, the messages transferred across radio interfaces
can be divided into two types: the messages needing opposite ac)nowledgement and
the messages not needing opposite ac)nowledgement.
"or the messages needing opposite ac)nowledgement, a T3;; must be started once
the message is sent. (f the opposite ac)nowledgement is not received after a period
of time, the message should be retransmitted. (n this case, the timer must be
restarted. (f the retransmission times exceed the maximum allowed times, the
message will no longer be retransmitted and the lin) will be released. That is, this
call drops. N3;; is the maximum retransmission times allowed. T3;; and N3;; have
different types depending on channel types 0T#& full rate, T#& half rate, and
,*##&1 and service types 0signaling and messages1. The given channel type and
service type matches a pair of T3;; and N3;;.
The call drop rate can be reduced if the message is retransmitted as early as possible
before the opposite ac)nowledgement is received. That is, the value of T3;; must be
set as small as possible and the value of N3;; must be set as great as possible.
&owever, the T3;; cannot be set to a too small value and the N3;; cannot be set to a
too large value. (f the opposite party has ac)nowledged that the lin) had been
removed, the retransmission will ma)e nonsense.
Therefore, to reduce the call drop rate, you can ad6ust the T3;; and N3;; according
to actual radio conditions.
&&. Call 9rop Cases
C!e 1( Cll drop cu!ed %, fre8uenc, $oppin) colli!ion
&'ro%lem nl,!i!+
8T, uses 2 x / !" frequency hopping. fter capacity expansion, the T#&
allocation failure rate is still high due to radio lin) problems. (n addition, the
T#& call drop rate and incoming handover failure rate are high. The ,*##&
call drop rate is normal.
&'ro%lem po!itionin) nd !olution+
8ecause high call drop rate and high incoming handover failure rate come
together with the T#& allocation rate, it can be 6udged that the problem may
arise during T#& assignment or the channel numbers or timeslots seized by
the call are interfered or unstable. 8ecause the ,*##& call drop rate is
normal, it can be 6udged that the probability for the 8##& carriers and
8##& numbers to the interfered are small, but the non48##& carriers and
non48##& numbers may be greatly interfered.
Through chec)ing the hardware, antenna feeder, and transmission, engineers
found no problem. ccording to drive test, engineers found that the signal
level was high but the quality was poor. Through on4site dialing test,
engineers found that the conversation quality was poor. Through chec)ing
engineering parameters, engineers found that the -(G of the new carrier
was the same as that of the old carrier. Therefore, it can be 6udged that the
call drop was caused by the frequency hopping collision. fter modifying the
-(G, engineers found that call drop rate became normal.
C!e 2( Cll drop cu!ed %, i!olted i!lnd effect
&'ro%lem de!cription+
7sers complained that call drop always occurred above the fifth floor of a
building.
&'ro%lem nl,!i!+
1. +$rou$ on,site test" enineers !ound t$at call dro) and
noise e:isted $ere. As !ar as t$e test MS #as concerned" it #as
al#ays in t$e service area o! t$e ot$er B+S -$ereunder called
B+S B. ot$er t$an t$e local B+S -$ereunder called B+S C. be!ore
t$e call dro).
%. >t is estimated t$at t$e service cell belons to B+S B" #$ic$
is 2,4 ?ilometers a#ay !rom t$e buildin. +$ere!ore" it can be
Fuded t$at t$e sinals !rom t$e B+S B are re!lected sinals. As a
result" an area similar to an isolated island is !ormed.
2. +$rou$ c$ec?in data con!iuration" enineers !ound t$at
only t$e cell % o! B+S A $as t$e nei$bor cell relations$i) #it$
B+S B. +$ere!ore" #$en t$e MS is usin t$e sinals in cell % o! t$e
B+S B" i! t$e sinals in cell 2 o! B+S A #ere stron" and i! t$e cell
% o! B+S B $as no nei$bor cell relations$i) #it$ t$e cell 2 o! t$e
B+S A" t$e $andover cannot be )er!ormed.
The signals from the cell 3 of 8T, 8 are reflected many times. Therefore, when
signals 0from 8T, 81 received by the -, became poor dramatically, emergent
handover may be initiated. (n this case, however, either the cell 3 or cell / of the
8T, is not an ideal candidate cell for the cell 3 of 8T, 8. s a result, the -, may
hand over to other 8T, 0hereunder called 8T, #1, but the -, cannot receive the
signals from 8T, #. Therefore, call drop occurs.
&Solution+
Eou are recommended to change the data in the 82 08##&1 list, 83
0,##&1 list, and neighbor cell relationship list. "or example, you can
configure the cell / of 8T, as the neighbor cell of cell 3 of 8T, 8. To
eliminate the isolated island effect, you should also further optimize the
engineering parameters. fter that, the call drop problem can be solved.
C!e 3( 6educe cll drop rte t$rou)$ optimi9in) $ndover prmeter!
&'ro%lem de!cription+
The drive test in an area found that the call drop rate at a cave near the 8T,
high because the handover cannot be performed in due time.
&'ro%lem nl,!i! nd !olution+
The cave is near the 8T,. The signal level of the target cell is about 49;d8m
in the cave, but the signal level of the old cell drops below 42;;d8m. The
downlin) power of the two cells outside the cave is good, so the handover
cannot be initiated. &owever, the signal level deteriorates dramatically in the
cave, so the call drop occurs before the measurement time is arriving.
To reduce the call drop rate, you can optimize and ad6ust the handover
parameters:
1. >! no )in,)on $andover is )resent and t$e conversation is
continuous" you can ma?e t$e @BG+ $andover $a))en as easily as
)ossible.
%. Set t$e t$res$old to trier t$e emerent $andover
rationally so t$at t$e emerent $andover can be triered be!ore
call dro).
"or the parameter modification, see
'andover )arameter o)timi(ation
Jarameter Name 8efore -odification fter -odification
J8+T handover measurement time = /
J8+T handover duration . 3
J8+T handover threshold A3 <9
7plin) quality threshold for emergent
handover
A; <;
-inimum downlin) power for handover
candidate cell
2; 2=

C!e 4( Cll drop cu!ed %, cloc5 pro%lem
&'ro%lem de!cription+
The cell of an 29;;-&z networ) has been cutover. fter the establishment
of a cell at site 8, the calls made in the cell handing over to the +,-:;;
-&z cell that shares the same 8T, site drops in the +,-:;; -&z cell. nd
the call drop rate is great.
&'ro%lem nl,!i! nd !olution+
Dngineers find that the cloc) of the +,-:;; -&z 8T, and that of the
+,-29;; -&z 8T, are asynchronous. Bhen the calls established on the
+,-29;; -&z cell intend to hand over to the +,-:;; -&z cell, the drive
test data shows that the "D! arises dramatically first, and then gradually
disappears to none. (f the handover is from a +,-:;; -&z cell to a
+,-29;; -&z cell, this phenomenon is also present. Through monitoring
signaling, engineers find that the conversation held several seconds before the
call drop is 6ust process for call re4establishment. &owever, the test -,
shows that the call has been handed over to the +,-:;; -&z cell.
Therefore, it can be 6udged that the cloc)s are seriously asynchronous. To
solve this problem, the carrier concerned and the +,-:;; equipment
provider cooperate with each other on cloc) calibration. fter that, the
abnormal call drop disappears. Therefore, for dual4band handover, the cloc)
of the +,-:;; -&z 8T, and that of the +,-29;; -&z 8T, must be
synchronous.
/./. )ando0er Pro5le"s
The -, is always moving during conversation. To ensure channel quality,
the -, must measure the quality of the channels of the surrounding channels
without stop, and then send the measurement report to the 8,# through the
service cell. The 8,# will perform radio lin) control according to the signal
level and quality contained in the measurement report. (f the -, moves to
another cell, the new cell will replace the old cell to ensure the continuity of
the service. The handover enables each cell to form a seamless networ).
&. )ando0er Pro5le" Positioning Steps
1. Find out i! t$e )roblem occurs at an individual cell or all
cells and !ind out t$e c$aracteristics o! t$e )roblem cells. For
e:am)le" i! t$e cells are t$e nei$bors cell o! a cell" or i! t$ey
are co,BSC cells" or i! t$ey are co,MSC cells.
(f the handover between two cells fails, you should focus on chec)ing if the data of
the two cells is correctly configured. (n addition, you should also chec) the hardware
of the two cells.
(f the problem is found in all the neighbor cells of a cell, you should focus on
chec)ing of the data of this cell is correctly configured. (n addition, you should also
chec) the hardware of the cell.
(f the problem is found in all the cells under the same 8,#, you should focus on
chec)ing the data configuration between the 8,# and -,#.
(f the problem is found in all the cells under the same -,#, the cooperation between
the local exchange and the opposite exchange may fail. "or example, the signaling is
incompatible and the timer setting is irrational.
%. C$ec? i! t$e data $as been modi!ied be!ore $andover
)roblems occur.
(f the problem is found in an individual cell, you should focus on chec)ing if the data
configuration for this cell has been modified.
(f the problem is found in all the cells under the same 8,#, you should focus on
chec)ing the data configuration for the local 8,# and the opposite -,# has been
modified.
(f the problem is found in the cells under the same -,#, you should chec) if the
data configuration for the opposite -,# has been modified.
2. C$ec? i! it is t$e $ard#are !ailure t$at causes t$e $andover
)roblem.
4. <eister t$e related tra!!ic statistics items" suc$ as t$e
$andover )er!ormance measurement and +C' )er!ormance
measurement.
C$ec? i! t$e +C' sei(ure o! t$e )roblem cell is normal. For
e:am)le" i! t$e call dro) rate is $i$.
C$ec? i! t$e outoin $andover success rate is normal.
Find out t$e causes !or t$e $andover !ailure.
C$ec? i! t$e radio $andover success rate is normal.
1. @er!orm drive test !or t$e )roblem cell and analy(e t$e
drive test sinalin.
C$ec? i! t$e u)lin? and do#nlin? o! t$e )roblem cells are
balanced" because unbalanced )at$ may cause $andover
)roblem -B+S )roblem may cause t$e unbalance..
C$ec? i! t$e measurement re)ort !or t$e )roblem cell
contains correct nei$bor cell list.
C$ec? i! a call can $and over !rom a )roblem cell to a
nei$bor cell correctly and c$ec? i! it can $and over !rom a
nei$bor cell to t$e )roblem cell.
Analy(e i! t$e sinalin )rocedure !or t$e $andover is
normal.
&&. )ando0er Pro5le" Analysis Methods
i# 2ndover cnnot %e initited
(f the -, is in a cell where the signal is poor, it cannot hand over to another
cell. (n this case, you should consider of the handover conditions are met and
if there is an outgoing cell available.
&ereunder details the possible reasons:
+$e $andover t$res$old is set to a lo# value
"or edge handover, the handover triggering condition is that the !xlev must be
smaller than the handover threshold. (f the edge handover threshold is set to a too
small value, the signal level of the neighbor cells will be far higher than that of the
service cell. (n this case, the handover cannot be initiated. s a result, the
conversation quality will be affected, or even the call drop will be resulted. The
setting of the handover threshold depends on the coverage scope of the cell. Eou can
change the size of the service area of the cell through changing the handover
threshold.
Eei$bor cell relations$i) is not set
Though the signal level in the neighbor cells of the service cell is high, the -, will
not report the neighbor cells if the neighbor cell relationship is not set. (n this case,
the -, cannot hand over to a neighbor cell. Through performing cell reselection or
dialing test, you can chec) the neighbor cell list reported by the -,. (f the -, has
moved to the ma6or lobe of a cell but this cell is not found in the neighbor cell list,
you should chec) if the correct neighbor cell relationship has been correctly set.
*uring the drive test, you can use another -, to scan the 8##& numbers to chec) if
the strong 8##& numbers are in the service cell or in the neighbor cell list.
'andover $ysteresis is irrationally set
(f the difference between the signal level of the handover candidate cell and that of
the service cell is greater than handover hysteresis, the cell can wor) as a target cell.
(f the hysteresis is set to a too great value, the handover is hard to be initiated.
+$e best measurement time YEY and Y@Y are irrationally set
*uring normal handover, the -, uses N4J rules to list the handover candidate cells
in a certain order. (f a candidate cell is the best cell within J seconds out of N
seconds, it will be treated as the best cell.
Bhen there are two cells become the best cell alternately, the -, may find it hard to
select a best cell through N4J rule, which ma)es the handover difficult. (n this case,
you can ad6ust the values of N and J and reduce the measurement time to ma)e the
handover decision more sensitive to level change.
(f the landform and the ground ob6ects of the service cell are quite complicated, the
signals received by the moving -, will fluctuate greatly. (n this case, the handover
candidate cell cannot meet N4J rule, which will ma)e the handover difficult.
ii# 2ndover pro%lem cu!ed %, $rd"re filure
(f the data configuration for the problem cell and the neighbor cells has not been
modified recently but the handover problems occur abruptly, you should first
consider if the problems are caused by 8T, hardware equipment.
(f the cells sharing the same base station with the cell have similar problem, you
should consider if the problem is caused by the common hardware of the cells.
(f the problem is found in only one cell under the base station, you should consider if
it is the hardware of this cell that causes the problem. "or example, if some of the
carriers are damaged. To test the problems of this )ind, you can disable some of the
carriers. (f the handover success rate returns to normal state after a carrier is
disabled, you can chec) if the problem is present at this carrier or if the #*7 and
antenna feeder part related to this carrier fails. (f signals of a cell on the uplin) and
downlin) are seriously unbalanced, frequent handover will be caused and the
handover success rate will decrease.
To chec) if the signaling flow of the cell is normal and if the uplin) !xqual and
downlin) !xqual are good, you can monitor the messages sent across the bis
interface. (f the !xqual is poor, it means that the hardware equipment of the fails or
serious interference is present in the cell. (n this case, the signaling exchange is
unavailable and the handover problem will occur.
iii# 2ndover pro%lem cu!ed %, irrtionl dt confi)urtion
For stand,alone net#or?in mode" i! t$e outoin MSC or
incomin MSC $andover is abnormal" you s$ould c$ec? i! t$e
sinalin coo)eration o! t$e t#o MSCs is correct. >n addition" you
s$ould also c$ec? i! t$e data con!iuration !or t$e o))osite MSC
and t$e local MSC $as been modi!ied recently.
For co,MSC net#or?in mode" i! t$e $andover is )er!ormed
#it$in t$e BSCs o! di!!erent )roviders and t$e inter,BSC
$andover is abnormal" you s$ould !irst c$ec? i! t$e sinalin
coo)eration bet#een t$e BSCs is normal" and t$en c$ec? i! t$e
data con!iuration !or t$e BSCs $as been modi!ied.
>! t$e abnormal $andover is !ound at a cell only" you need
to analy(e t$e abnormal $andover accordin to actual
conditions.
(f the incoming handover of a cell is abnormal, you need first chec) if the incoming
handovers to this cell is abnormal. +enerally, when the handover is abnormal, the
handover success rate is low, or even the handover does not occur.
(f all the incoming handovers to this cell is abnormal, you should chec) if the data
configuration for this cell is correct. &ere the data configuration includes both the
data configured for this cell and the data configured for other cells but is related to
this cell. "or example, the #+( of this cell may be correctly configured, but it may be
wrongly configured in other cells.
(f there is only one incoming handover to a cell is abnormal but other incoming
handovers to this cell are normal, in addition to chec)ing if the data configuration for
this cell is correct, you should also chec) if the data configuration for the neighbor
cells is correct. "urthermore, you should also chec) if the hardware equipment of the
cell is normal.
The methods to analyze the abnormal outgoing handovers are similar to the methods
to analyze the abnormal incoming handovers.
C$ec? t$e timers -suc$ as +2101" Ey1" +2102" and +214%.
related to t$e $andover.
T/2;= indicates the interval for continuous J&E,(#H (N"G!-T(GN to
be sent to the -,. The networ) will start T/2;= for the sending of the
J&E,(#H (N"G!-T(GN. (f the timer fails before receiving any correct
frame from the -,, the networ) will resend the J&E,(#H
(N"G!-T(GN and restart the timer. piece of J&E,(#H
(N"G!-T(GN can be sent Ny2 times to the maximum. &ere the product of
Ny2 and T/2;= must be greater than the sum of T/23. and delta 0'delta'
indicates the interval between the expiry of T/23. and the reception of the
&andover "ailure message of the old 8,#1, otherwise the -, cannot perform
successful handover.
T/23. is a timer waiting for the J&E,(#H (N"G!-T(GN from the
networ) side during asynchronous handover. Bhen sending the
&N*GPD! ##D,, message on the *##& for the first time, the -, will
start T/23.. 7pon receiving a piece of J&E,(#H (N"G!-T(GN, the -,
will stop T/23.. (f the channel type allocated in the &N*GPD!
#G--N* message is ,*##& 0I,##&1, the T/23. is set to <A=ms. "or
other cases, the T/23. is set to /3;ms.
&&&. )ando0er Cases
C!e 1( 4o $ndover cndidte cell i! vil%le due to C:. error
&'ro%lem de!cription+
The handover in an area is abnormal. Bhen the -, moves from cell to cell
8, the signals in cell 8 are stronger than that of cell , but the handover does
not happen. fter the -, moves from cell 8 to cell #, the -, hands over
from cell to cell #.
&Cu!e nl,!i!+
(f a cell can wor) as a service cell and can hand over to other cells, but the
incoming handover is unavailable, you can chec) if the #+(, 8,(#, 8##&
number of the cell are correct.
&'ro%lem !olution+
1. /se t$e test MS to loc? t$e BCC' numbers o! cell B. +$e call
is normal. +$e MS can $and over to any ot$er cell by !orce.
%. Ma?e a call a!ter loc?in t$e BCC' umber o! any nei$bor
cell o! cell B" and t$en !orce t$e MS to $and over to cell B" but
t$e $andover does not $a))en" because no $andover command is
seen in t$e drive test so!t#are.
2. +$e $andover )rocedure requires t$e MS detectin t$e
nei$bor cell sinals and re)ortin t$e detected sinals to BSC
#it$ a measurement re)ort. /)on receivin t$e measurement
re)ort" t$e BSC must ma?e t$e $andover decision. >! t$e
$andover conditions are met" t$e BSC s$ould activate t$e +C' o!
t$e service cell" and t$en send a $andover command to t$e MS.
4. >! t$e sinals o! cell B are !ar stroner t$an t$at o! cell A
and t$e $andover conditions are met -t$e @BG+ $andover
t$res$old is 50." but no $andover command is sent" it means t$at
errors occur durin t$e activation o! t$e taret cell +C'.
1. >! t$e cell B #or?s as t$e taret cell but t$e +C' cannot be
activated" t$e data may be incorrectly con!iured !or t$e cell. >n
t$is case" t$e BSC t$at contains t$e cell cannot !ind t$e taret
cell" so t$e +C' cannot be activated and no $andover command
can be !ound in t$e service cell.
3. +$e CG> error is !ound in cell B t$rou$ data c$ec?in. +$e
$andover is normal a!ter t$e CG> is c$aned to correct value.
C!e 2( 1n%lnced pt$ cu!e! lo" $ndover !ucce!! rte
&'ro%lem de!cription+
The incoming 8,# handover success rate is quite low for the two cells under
a 8T,, ranging from 2;F to /;F.
&Cu!e nl,!i!+
+enerally, if the data problems, such as #+( error or intra4frequency
interference, exists and if there is dead zones in heavy4traffic areas, or if it is
hard for the -, to access the cell due to poor uplin) signals, the incoming
8,# handover success rate is low.
&Solution+
1. +$e cell data is !ound normal.
%. +$rou$ c$ec?in tra!!ic statistics items" enineers !ound
t$at all incomin cell $andover success rates #ere lo#.
2. +$rou$ drive test" enineers !ound t$at !requent $andover
attem)ts #ere made in t$e area %?m a#ay !rom t$e B+S" but t$e
$andover al#ays !ailed. ;ven i! a success!ul $andover #as made"
call dro) occurred immediately. Durin t$e $andover" enineers
!ound t$at t$e do#nlin? level #as about ,81dBm. ;nineers
made 10 dialin tests #it$ !requency loc?ed" all t$e oriinatin
calls !ailed. For t$e ans#erin calls" t$ey can be connected but
cannot be called out.
4. >t is estimated t$at t$e CD/ u)lin? c$annel loss is reat" or
t$e Fum)ers are incorrectly connected at t$e B+S to). >n t$is
case" t$e u)lin? sinals #ill be )oor" #$ic$ causes t$e )roblem.
1. A!ter c$anin t$e CD/" enineers !ound t$at t$e incomin
$andover success rate increased to 91L.
C!e 3( .mproper ntenn plnnin) cu!e! lo" $ndover !ucce!! rte
&'ro%lem nl,!i!+
The handover success rate among the three cells under a 8T, is quite low
according to traffic statistics. "or the handover from cell2 to cell/ and the
handover from cell3 to cell/, the success rate is lower than /;F.
&Cu!e nl,!i!+
+enerally, low handover success rate is caused by board failure, handover
data error, or improper antenna planning.
&Solution+
1. +$e B+S $ard#are is normal and no alarm concernin
$andover )arameters is enerated" so t$e $ard#are )roblem and
)arameter settin )roblem can be e:cluded.
%. +$e B+S locates at t$e eastern side o! a sout$,nort$ road
and is 500m a#ay !rom t$e road. +$e a(imut$ anles o! t$e
t$ree cells are 0`" 80`and 130`. +$ey t$ree cells direct to t$e
t#o directions and t$e o)en resident areas lyin under a $ill in
t$e east res)ectively. Amon t$e t$ree cells" t$e do#n tilts o!
t#o cells are 5`. +o ma?e t$e coverae as s)eci!ied as )ossible"
enineers concentrated t$e antenna a(imut$ anles o! t$e t$ree
cells in desin. >n t$is case" $o#ever" t$e cells o! t$e B+S are
seriously overla))ed in t$e east. For t$e areas Fust in t$e #est"
t$e coverae is )rovided by t$e side lobes and bac? lobes o! t$e
t$ree cells. +$ere!ore" #$en t$e MS is movin on t$is road" !irst
it is covered by cell1. 9$en it moves to t$e #est" t$e sinals o!
t$e t$ree cells are )oor and !luctuatin reatly. >n addition"
since t$e $andover measurement time and t$e $andover
duration is set to a small value" t$e $andover is rat$er sensitive"
and t$at]s #$y t$e !requent $andover !ailure occurs.
2. A!ter settin t$e a(imut$ anles o! t$e t$ree cells to 30`"
180`" and 210`" enineers !ound t$at t$e $andover success rate
o! amon t$e t$ree cells increased to 91L.
C!e 4( 'ro%lem! concernin) t$e coopertion of different crrier!;
e8uipment cu!e lo" out)oin) BSC $ndover !ucce!! rte
&'ro%lem de!cription+
There is a dual4band networ) in which the +,-:;; -&z networ) and the
+,-29;; -&z networ) are stand4alone. fter the two networ)s completed
cell reselection and handover parameter setting, engineers found that the
dual4band handover success rate was low@ especially the success rate of the
handover from the +,-29;; -&z networ) to the +,-:;; -&z networ)
was low, ranging from <;F to 9;F. &owever, the success rate of the
handover from the +,-:;; -&z networ) to the +,-29;; -&z networ)
was higher than :3F.
&Cu!e nl,!i!+
"or a dual4band networ), if the problems concerning the cooperation of
different carriers5 equipment are found, you must )now the data configuration
of the equipment. "or example, if the equipment supports Jhase 3I and D"!.
&Solution+
1. +$rou$ usin sinalin analy(er to analy(e t$e messae
!lo#in across t$e A,inter!ace and ;,inter!ace" enineers !ound
t$at t$e MSC o! t$e GSM1800 M'( net#or? #ould send bac? a
'andover <eFect messae to t$e BSC o! t$e GSM1800 M'(
net#or? #$en t$e BSC sent a 'andover <equired messae to t$e
MSC.
%. +$e MSC o! t$e GSM1800 M'( net#or? sent a @re)are
'andover messae to t$e MSC o! t$e GSM900 M'( net#or?. /)on
receivin t$e messae" t$e MSC o! t$e GSM900 M'( net#or? sent
bac? an Abort messae.
2. Because t$e success rate o! t$e $andover !rom GSM900 M'(
net#or? to t$e GSM1800 M'( net#or? #as $i$" enineers !ound
t$at t$e voice version carried in t$e @re)are 'andover messae
-!rom t$e GSM900 MSC to GSM1800 MSC. is $al! rate version 1"
but t$e voice versions carried in t$e @re)are 'andover messae
-!rom GSM1800 MSC to GSM900 MSC. are !ull rate version 1" !ull
rate version %" and $al! rate version 1" #$ic$ belon to @'AS;
%J. 'o#ever" MSC o! )rovider A does not su))ort t$e @'AS; %J"
so t$e $andover !ailure is caused.
4. +$rou$ modi!yin t$e MSC data o! t$e circuit MSC data at
t$e A,inter!ace and selectin t$e !ull rate version 1 only"
enineers !ound t$at t$e voice versions carried in t$e @re)are
'andover messae -!rom GSM1800 MSC to GSM900 MSC. are !ull
rate 1 and %. A!ter t$at" t$e dual,band $andover success rate
#as reatly increased.
/././ Congestion Pro5le"s
This section introduces the methods to handle ,*##& congestion and T#&
congestion, in which T#& congestion indicates ,*##& seizure all busy. The
T#& congestion has two cases. Gne is T#& seizure all busy. "or this case,
the real channels cannot be allocated to the -,, so the -, will fail to request
the channels. The other one is that the T#& assignment fails after an
assignment is sent due to various reasons.
&. Congestion Pro5le" Solutions
Conestion caused by $eavy tra!!ic
Eou can chec) if the ,*##& traffic and T#& traffic are normal through viewing
traffic statistics. (f the congestion is caused by heavy traffic, the most efficient
method to solve the problem is to expand the capacity of the networ). (n addition,
you can adopt traffic sharing technologies to ease the congestion. "or example, you
can modify the #!G, enable direct retry or load handover function.
SDCC' conestion caused by burst tra!!ic
(f the ,*##& congestion rate is high and the traffic is heavy but the T#& traffic is
normal, the ,*##& congestion may be caused by burst traffic. The ,*##&
congestion always occurs at 8T,s along railways and tunnel exits, because the 8T,s
are installed in remote places and the capacity of a 8T, is small. s a result, when
the train moves fast along the railways or stops at a railway station, most of -,s
failing to capture a networ) will perform location update, which will result in
,*##& congestion. (n addition, when short messages are sent at a concentrated
time, the ,*##& congestion will also occur easily. ,*##& congestion cannot be
completely avoided, but some measures can be ta)en to ease the congestion. "or
example, you can add the number of ,*##&s, or enable the dynamic conversion
between ,*##& and T#&.
Conestion caused by +<O )roblems
Bhen a carrier configured in a multi4T!% cell cannot provide services, the channel
congestion will also occur. To solve the problem, you should replace the problem
T!% with a sound one. (f the T!% problem cannot be positioned, you should chec)
if the antenna feeder connection is correct and if the antenna standing wave is
normal. (f yes, rechec) the T!% where the problem may be generated.
Conestion caused by inter!erence
The interference present across the radio interfaces will also cause congestion. (n this
case, you should solve the interference problem first.
C$annel assinment !ailure caused by inconsistent coverae
(f the concentric technology is not used, the transmit power of the T!%s within the
same cell will be inconsistent, which will result in inconsistent coverage. (n this case,
the channel assignment failure will easily occur. To position this problem, you can
chec) if the connection between the splitter and connector and if the connection
between #*7 and ,#7 are correct.
Bhen a cell uses multiple transmitter antennas, inconsistent coverage will be
inconsistent, which will result in channel assignment failure. To solve this problem,
you should ma)e the coverage of each transmitter antenna as consistent as possible
through engineering ad6ustment.
(n addition, if the transmitter antenna and the receiver antenna of a cell is not in the
same plane or the antenna tilt angles are inconsistent, the channel assignment failure
will also be caused. (n this case, you can calibrate the antennas to solve the problem.
Conestion caused im)ro)er data con!iuration
(f the congestion is caused by improper location area planning, you can reduce the
,*##& congestion rate by planning the location area properly.
(f the congestion is caused by the problems concerning ,*##& dynamic allocation,
you can reduce the ,*##& congestion rate by enabling the ,*##& dynamic
allocation function.
"or dual4band networ), you can properly set the parameters 0such as #!G, #8,
and cell reselection hysteresis1 to reduce the ,*##& congestion rate.
(f the timers, such as T/2;2, T/2;/, T/2;A, T/233, T/323, and T/222, are not
properly set, ,*##& congestion will also be caused.
&ereunder are the solutions to the previous problems.
Eou can ease the congestion caused by ,*##& dual allocation through reducing the
T/2;2 to a smaller value. (f the T/2;2 is set to a large value, the radio resources will
be seized for a long period of time. To fully use the radio resources, therefore, you
can reduce the T/2;2 value.
Eou can save the T#& resources through reducing the T/2;/ and T/2;A to a rational
value. +enerally, T/2;/ and T/2;A are set to about = seconds.
The T/233 must be stopped once the -, receives an (--D*(TD ,,(+N
!DYD#T message. Gnly after the T/233 expires, the -, can send a new channel
request message. (f the -, sends channel request messages frequently, the !#&
load and ###& load will increase. To solve this problem, you can increase the
T/233 to a larger value.
T/323 stands for the time limit value for periodical location update. Eou can ease
the ,*##& load by increasing the T/323 to a larger value.
T/222 is related to release latency. (t is used for the deactivation of the latency
channels after the ma6or signaling lin) brea)s. T/222 can be initiated during either
T#& release or ,*##& release. The value of T/222 must be consistent with that of
the T/22; at the -, side. +enerally, it should be 3 seconds. (f the T/222 is set to a
large value, great ,*##& congestion rate may be caused.
&&. Congestion Cases
C!e 1( SDCC2 con)e!tion cu!ed %, "ron) <AC confi)urtion
&De!cription+
8T, is configured as ,2$2$2. (t is found that the ,*##& congestion rate for 3 cells
reaches as high as 9F.
&'ro%lem nl,!i! nd !olution+
1. +$rou$ c$ec?in t$e measurement inde:es !or +C' and
SDCC'" enineers !ound t$at t$e +C' tra!!ic #as not $eavy. +$e
tra!!ic volume !or eac$ cell durin busy $ours is lo#er t$an
%.%;rl. 'o#ever" t$e requests !or SDCC' sei(ure are reat"
reac$in 202% times durin busy $ours. +$e SDCC' tra!!ic
reac$es 1.83;rl" and t$e conestion rate reac$es 8L.
%. +$e main reasons !or SDCC' sei(ure include t$e messaes
sent be!ore call setu)" t$e messaes sent durin $andover" t$e
location u)date messaes sent under t$e idle mode" and ot$er
s$ort messaes.
2. +$e +C' tra!!ic is normal" t$e requests !or +C' sei(ure
-includin $andover. are normal -218 times." and t$e $andover
requests are also normal -143 times.. +$ere!ore" t$e SDCC'
conestion may be caused by a lare number o! location u)date
messaes or s$ort messaes.
4. +$e GAC o! t$e B+S is 0100" and t$e GACs o! ot$er cells o!
t$e surroundin cells are 01%0. A!ter c$anin t$e GAC o! t$e B+S
to 01%0" enineers !ound t$at t$e requests !or SDCC' sei(ure
durin busy $ours #ere %98" t$e SDCC' tra!!ic #as 0.%5;rl" and
t$e conestion rate reduced to 0.
C!e 2( SDCC2 con)e!tion cu!ed %, %ur!t loction updte!
&'ro%lem de!cription+
The radio connected ratio of a local networ) is lower than average level.
ccording to traffic statistics analysis, it is found that the ,*##& congestion
happened at several 8T,s.
&'ro%lem nl,!i! nd !olution+
1. +$rou$ analy(in tra!!ic statistics" enineers !ound t$at
t$e SDCC's o! t$e conested cells #ere sei(ed !or 200 to 400
times durin busy $ours. 'ere t$e B+S #as con!iured as S1=1=1
and eac$ cell #as con!iured #it$ 8 SDCC's. +$ere!ore" t$e
SDCC's can be sei(ed by 400 times" but t$e SDCC's #ere
conested !or tens o! times durin busy $ours.
%. As !ar as t$e reistered tra!!ic statistics items #ere
concerned" most o! t$e SDCC' sei(ures #ere caused by location
u)date. +a?in t$e B+S location into consideration" enineers
!ound t$at most o! t$e B+Ss #ere installed at t$e intersections o!
t#o rail#ays. +$ere!ore" it mi$t be t$e burst location u)date
t$at caused SDCC' conestion.
2. +o veri!y i! it #as t$e burst location u)date t$at caused t$e
conestion" enineers reistered t$e tra!!ic statistics items in 1
minutes and !ound t$at most o! t$e location u)date $a))ened
#it$in t$e !ive seconds. +$rou$ queryin t$e train time table"
enineers !ound t$at t$ere #ere 4 to 1 trains )assin by #it$in
t$e !ive seconds. 9$en t$e trains )assed t$e intersections" a
lare number o! location u)dates #ere enerated in a s$ort
time. >n t$is case" t$e conestion #as caused.
Therefore, if the 8T,s are installed at the railway intersections, you are
suggested to enable the ,*##& dynamic allocation function and configure a
suitable margin for the ,*##&.
C!e 3( :ret *C2 con)e!tion rte cu!ed %, t$e incon!i!tent tilt n)le!
of t"o ntenn! under t$e !me cell
&'ro%lem de!cription+
(t is found that the T#& congestion rate of a cell is great 0greater than =F1
according to traffic statistics.
&'ro%lem nl,!i! nd !olution+
1. +$rou$ c$ec?in BSC tra!!ic statistics" enineers !ound t$at
it #as t$e +C' sei(ure !ailure t$at directly caused t$e reat +C'
conestion rate.
%. Generally" +C' sei(ure !ailure is caused by +C' assinment
!ailure. +$rou$ monitorin Abis inter!ace" enineers !ound t$at
most o! t$e +C' assinment !ailure occurred at t$e Eo.4 +<O and
Eo.1 +<O" and t$e )robability !or t$e assinment !ailure rate !or
t$e Eo.4 +<O #as near t$at !or t$e Eo.1 +<O.
2. +$rou$ c$ec?in t$e antenna !eeder )art" enineers !ound
t$at t$e tile anle corres)ondin to t$e transmitter antennas o!
t$e t#o +<Os #ere too reat -it is 10 derees t$an t$at o! t$e
antenna !or BCC'" because t$e antenna nuts #ere !ound loosen.
+$ere!ore" i! t$e MS is !ar a#ay !rom t$e B+S" it can receive t$e
BCC' sinals but cannot receive t$e +C' sinals. >! t$e +C' is
assined to t$e MS #$en t$e MS starts a call" t$e +C' sei(ure
!ailure #ill occur.
4. +o solve t$e )roblem" you can enable t$e tilt anel o! t$e
+C' antenna and t$at o! t$e BCC' antenna to be consistent #it$
eac$ ot$er. >n t$is case" t$e +C' conestion rate can be reduced
to %L or lo#er.
C!e 4( 2i)$ *C2 con)e!tion rte cu!ed %, do"nlin5 interference
&'ro%lem de!cription+
cell of 8T, is responsible for covering a large area of sea surface along the
coast. ccording to the registered traffic statistics items, engineers found that
when the traffic volume was lower than 2Drl, the T#& congestion of the cell
reached 2;F at some time. &owever, no alarm was generated. ll the
interference bands fell within the interference band2 and the hardware and !"
connections were normal.
&'ro%lem nl,!i! nd !olution+
8ecause all the interference bands fell within the interference band2, the
uplin) interference is impossible. #onsidering that the coverage distance
reaches <; to A; )ilometers and multiple normal cells are present along the
coast, the probability for the channel numbers of the downlin) areas and the
cell to be interfered is high. Through modifying the channel numbers of the
cell, engineers found that the T#& congestion rate was improved. Through
further optimizing the channel numbers, engineers found that the T#&
congestion rate was lower than 2F. Therefore, it can be 6udged that the 2;F
of congestion rate is caused by the downlin) interference of some areas.
/./.. 3ther Pro5le"s
&. Su5scri5er &s !ot in Ser0ice Area
Bhen a subscriber is not in service area, the -, wor)s as the called party
when the signals received by the -, are good and the calling party hears a
voice saying that the subscriber is not in the service area. (f the coverage,
operation, and data configuration of a networ) are good, the occurrence
probability for the problem must be lower than 2F, otherwise it is other
causes that result in the problem.
+enerally, if a subscriber is not in the service area, the following causes may
be present:
Coverae )roblem
(f the subscriber complaint happens at cell edges, the problem may be related
to coverage. (n this case, the probability for the calling party failure is equal
to the probability for the 'subscriber is not in service area'.
@arameter settin
(f the subscriber is not in service area, the parameters concerning the paging,
access, and immediate assignment may be not properly set. (n this case, you
can chec) if the messages concerning !#& overload, J#& overload, and
,*##& overload are generated through querying traffic statistics and alarms.
(f yes, the subscriber may not in the service area.
System ca)acity limit or overload
(f system capacity limit or overload occur 0for example, &*8 overload, #J7
overload, or capacity overload during busy hours1, the system may fail to read
the subscriber information, or subscribers cannot access the radio networ). (n
this case, the subscriber may not in the service area. To solve this problem,
you can ta)e measures to expand the system capacity.
+ransmission )roblem
(f the lin)s between systems 0such as the HJ* lin) of the bis interface and
the lin)s of the each networ) entity1 and the lin)s within a system 0such as
the lin) among the modules of 8#,$-,#1 are not stable, the messages sent
through these lin)s may be missing. (n this case, subscribers may not in the
service area. To position the problem, you can chec) the alarms.
;qui)ment causes
(f the designs concerning -,# and 8,# are incomplete, the probability of
'subscriber is not in service area' will increase.
MS causes
(f the !" parts or the software parts of the -, have problems, for example,
the reception capability of the -, is poor@ the frequency deviation goes
beyond the requirements defined in the protocols@ and the dual4band
performance of the -, is poor, this problem may also occur.
&&. Signal Fluctuation
,ignal fluctuation indicates that change of the -, signal strength. The
following factors may cause signal fluctuation.
<adio #ave )ro)aation
The strength of the signals received by the -, is the amplitude of the sum of
the vectors of various propagation paths. 8ecause the propagation
environment is ever4changing, the attenuation of radio channels is ever4
changing. Therefore, even if the -, does not move, the strength of the
received signals will change.
Cell reselection and $andover
Bhen the -, moves from the old cell to the new cell through handover or
cell reselection, the signals of the -, will fluctuate because the signal
strength of the old cell and that of the new cell are inconsistent.
Antenna s$a?e
Bhen the antenna sha)es, the antenna gain will change, so the signals will
also fluctuate.
Gocation u)date or c$annel assinment occurs at non,BCC'
+<O
(f the ,*##& is assigned to a non48##& T!% during location update and
power control is available on ,*##&, the signal strength may fluctuate.
&&&. ?oice 9iscontinuity
Poice discontinuity stands for pauses or words loss occurs in conversation. (f
the voice continuity is remar)able, the conversation quality will be affected.
The following factors may cause voice discontinuity
Frequent $andovers
Gnly hard handover is available in +,- system. Therefore, when the -,
hands over from the source channel to the target channel, the downlin)
frames may loss at the bis interface. s a result, the voice continuity is
unavoidable during handover. +enerally, frequent handovers may occur at
cell edges or during cross coverage. (n this case, the voice discontinuity will
become a headache of subscribers. To avoid the frequent handovers, you can
ad6ust the antenna tilt and height and configure the data, such as uplin) and
downlin) quality threshold and restriction properly.
<adio lin? inter!erence
!adio lin) interference will increase the bit error rate, which will cause voice
discontinuity. (n addition, the signals always fluctuate greatly at cell edges, so
the conversation quality at the cell edges is bad.
@oor net#or? coverae
(f the networ) coverage is poor, the !xlev and !xqual will become poor, thus
the conversation quality will be affected.
B+S transmission )roblem
8T, transmission problems will affect conversation quality. "or connectors,
you should chec) of the connection among the connector is good. "or optical
transmission, you should chec) if the optical headers are clean and if the
transmission errors are great. "or micro wave transmission, it may be affected
by weather. (f there is large amount of dust in the equipment, the
conversation quality may also be affected. (f both micro wave transmission
and optical transmission are used, you should pay attention to the cooperation
of the transmission impedance at the equipment interfaces.
+<O board !ailure
&ardware problem will result in poor conversation quality. (n this case, you
should replace the problem hardware with the sound one.