s Network Design and Consulting Planning & Optimisation of Frequency Hopping in GSM Networks Authors: U. Rehfuess, ICM N MR Dr. K. Dietrich, ICM N MR A. Volke, ICM N MR B. Kronmueller, ICM N ST
Oct 30, 2014
s
Network Design and Consulting
Planning & Optimisation of
Frequency Hopping in
GSM Networks
Authors:
U. Rehfuess, ICM N MRDr. K. Dietrich, ICM N MRA. Volke, ICM N MRB. Kronmueller, ICM N ST
s
Network Design and Consulting
Outline
Planning & Optimisation of Frequency Hopping in GSM Networks
Implementation Aspects
Frequency Assignment in FH Networks
BSS - Database Parameters
Optimisation Aspects
Summary
s
Network Design and Consulting
Implementation Aspects
Planning & Optimisation of Frequency Hopping in GSM Networks
s
Network Design and Consulting
Frame N° 0Frame N° 1Frame N° 2Frame N° 3
Baseband FH
RF1
RF2
RF3
RF4
BB1
BB2
BB3
BB4
Logical Channel
1
2
3
4
Synthesizer FH
• Mobiles use Synthesizer Hopping only• BS implementation: power down, synthesizer re-
tuning and power up again within guard period 2 Synthesizers are implemented
RF1..n
RF1..n
RF1..n
BB1
BB2
BB3
BB4
RF1..n
Logical Channel
1
2
3
4
Implementation AspectsKey Differences Between Baseband and Synthesizer FH
s
Network Design and Consulting
Baseband FH Synthesizer FH
Implementation Aspects Combining Equipment in Baseband and Synthesizer FH
RF1
RF2
RF3
RF4
BB1
BB2
BB3
BB4
1
2
3
4
Filt
er
Com
bin
ing
TX Antenna
RF1..n
RF1..n
RF1..n
BB1
BB2
BB3
BB4
RF1..n
1
2
3
4
Hyb
rid C
om
bin
g
TX Antenna
• Narrow Band• Low insertion loss (3-4 dB)
• Wide band• Higher insertion losses (~3 dB/stage)• On-air combining possible (DUCOM)
s
Network Design and Consulting
Implementation Aspects Combiner, Link Budget of Synthesized Frequency Hopping
Power Output Power
Amplifier W
PA 25 W40 W
HPA 60 W
Combiner Attenuation
Type dB
DUCOM 2:1 2.5
DUCOM 4:1 5.7
HYCOM 1:1 2.0
HYCOM 2:1 3.7
HYCOM 4:1 6.5
FICOM 2:1 2.4
FICOM 4:1 3.0
FICOM 6:1 3.3
DUCIT 2.8
Combiner Losses and Output Power
Example: GSM 900
• Determine configuration and hardware status before SFHimplementation
• Determine necessary upgrades(TPU, PA, Combiner)
• Actualize and check Link budgets
• Introduce further HPA wherenecessary
• Max. number of TRXs per celldepends on hardware configuration
s
Network Design and Consulting
BFH SFH
TPU X
TPU 2 X X
PA X X X*
HPA X X
* not all types of first generation power amplifiers are suitable for for SFH
Implementation Aspects TPU, PA for Synthesized Frequency Hopping
s
Network Design and Consulting
• Software Release: BR 3.7 or higher
• Cell Synchronization: up to 2/2/2 BS 6x/2xup to 8/8/8 BS 24x
• No. of Hopping Frequencies:max. 16 per cell (BFH incl. BCCH)max. 15 per cell (SFH w/o BCCH)max. 64 per cell with BR 6.0
• BS11: SFH only (BR 4.0)
Implementation Aspects Hardware and Software for Synthesized Frequency Hopping
s
Network Design and Consulting
• Baseband hopping Narrowband RFcombining sufficient
BCCH TRX except for TS0 may hop
• Synthesizer hopping Wideband RF combining required
One TRX per hopping frequency required!
• No. of RF = No. of TRX
• No. of RF > No. of TRX BCCH TRX must not hop
More hopping frequencies than TRXs feasible
Implementation Aspects Key Differences Between Baseband and Synthesizer FH
s
Network Design and Consulting
Implementation AspectsHardware Requirements: Repeater
Wideband Repeaters:
• Usable for SFH and BFH
• Careful implementation (amplification of signals
in the whole frequency band)
Channel selective Repeaters:
• Usable for BFH
• Number of frequencies is limited
• Usually not usable in tight reuse scenarios
s
Network Design and Consulting
Frequency Assignment in Hopping Networks
Planning & Optimisation of Frequency Hopping in GSM Networks
s
Network Design and Consulting
Frequency PlanningProcess Frequency Assignment
Split of BandBCCH - TCH
DedicatedCommon
Multiple Reuse
Planning of BoundariesHopping – Non Hopping
Available Spectrumfor Hopping
Frequencyassignment with
fixed reuse schemes
Reuse 1x3 Reuse 1x1 other
Tools Interference Table Separation Settings
Tool optimizedfrequencyassignment
MAIO and HSNPlanning
Cyclic HoppingRandom Hopping
DatabaseGeneration
Planning of AnchorFrequencies in SFH
• Guideline for RF-planners
• Focus on SFH planning and hopping TCH - carriers
• BCCH - carrier assignment: planning with tool is always recommended
• Planning must be adjusted to each individual network
s
Network Design and Consulting
Frequency Planning Common Band - Dedicated Band - Multiple Re-Use Patterns
5 hopping frequenciesPC on, DTX on
[%]
90%@FER2%
Dedicated Band 71.8%
Dedicated Band
15 BCCH carriers 28 TCH carriersCommon Band
59.7%
Common Bandtotal operator bandwidth 8.6 MHz = 43 carriers
43 carriers for both BCCH and TCH
MRP 54.3%
15 BCCH carriers 12 TCH + 9 TCH + 7 TCH carriers
Multiple Re-use Patterns (MRP)
Achievable System Load
s
Network Design and Consulting
Fixed reuse scheme to all hopping cells
possible reuses:3/9, 2/6, 1/3, 1/1
Tool supported frequency assignment based on interference matrix considering FH gains
Frequency PlanningFrequency groups - Tool supported planning
TCH 1 TCH 3
TCH 2
TCH 1 TCH 3
TCH 2
TCH 1 TCH 3
TCH 2
TCH 1 TCH 3
TCH 2
s
Network Design and Consulting
Frequency Planning Planning of BCCH
BCCH 3 BCCH 2
BCCH 1
BCCH 9 BCCH 8
BCCH 7
BCCH 6 BCCH 5
BCCH 4
BCCH 13 BCCH 11
BCCH 10
BCCHe.g. 4 x 12 Reuse
• Reliability
• Neighbor Measurements
• BSIC Decoding
• BCCH Frequency active at all timeslots in the downlink-> no interference averaging
s
Network Design and Consulting
Cluster 1/3
Channel
1, 4, 7, 10, ...
2, 5, 8, 11, ...
3, 6, 9, 12, ...
With a deliberately MAIO - assignment and identical HSN assignment to sectors you can avoid adjacent-channel interference between the sectors within one site
Frequency Planning Examples for frequency groups (I)
Co-channel interference is avoided by the frequency groups MAIO TRX1 TRX2 TRX3 ...
Sector 1 0 2 4 ...
Sector 2 1 3 5 ...
Sector 3 0 2 4 ...
Min # RF 6 12 18
TCH C TCH B
TCH A
TCH C TCH B
TCH A
TCH C TCH B
TCH A
TCH C TCH B
TCH A
TCH C TCH B
TCH A
TCH C TCH B
TCH ATCH A
TCH B
TCH C
s
Network Design and Consulting
Frequency Planning Examples for frequency groups (II)
• Each sector within a site uses a different Frequency Group
• No co-channel collisions between sectors of a site
• Synchronisation between the sectors and MAIO management avoid adjacent channel collisions
• Homogeneous network:no co-channel collisions between serving cell and all nearest neighbour cells
TCH uses each frequency onlypart of the time (e.g. 50%)
50% fractional load
TCH 1 TCH 3
TCH 2
TCH 1 TCH 3
TCH 2
TCH 3
TCH 2
TCH 1 TCH 3
TCH 2
TCH 1
s
Network Design and Consulting
Cluster 1/1
• All sectors same frequency group
• Identical HSN to sectors of one site
• MAIO assignment to avoid co- and adjacent channel interference
Frequency PlanningExamples for frequency groups (III)
MAIO TRX1 TRX2 TRX3 TRX4 ...
Sector 1 0 6 12 18 ...
Sector 2 2 8 14 20 ...
Sector 3 4 10 16 22 ...
Min # RF 6 12 18 24
s
Network Design and Consulting
Frequency PlanningExamples for frequency groups (IV)
• Each sector within a site uses the same frequency group
• Synchronisation between the sectors and MAIO management required to avoid co-channel collisions
• Homogeneous network:Co-channel collisions between serving cell and nearest neighbor
TCH TCH
TCH
TCH TCH
TCH
TCH TCH
TCH
TCH TCH
TCH
s
Network Design and Consulting
Frequency PlanningHopping Sequence Generation (I)
MAI = (FN + MAIO) modulo N if HSN = 0 (cyclic hopping)
GSM 05.02.
MAI ... Mobile Allocation Index (integer 1...N-1)
FN ... TDMA Frame Number (0... 26*51*2048-1 = 2 715 647)
MAIO ... Mobile Allocation Index Offset (0 ... N -1)
N ... Number of allocated frequencies
For example: (MAIO=0)
MA = 1,4,7,10,13,16,19,21,24,27,30,33,36,39,41
set of ARFCN numbers to be
used in the hopping sequence
N=15
1. burst FN = 0: MAI = (0 + 0) mod 15 = 0 ARFCN = 12. burst FN = 1: MAI = (1 + 0) mod 15 = 1 ARFCN = 4
14. burst FN = 14: MAI = (14 + 0) mod 15 = 14 ARFCN = 4115. burst FN = 15: MAI = (15 + 0) mod 15 = 0 ARFCN = 1 16. burst FN = 16: MAI = (16 + 0) mod 15 = 1 ARFCN = 4etc...
s
Network Design and Consulting
Frequency PlanningHopping Sequence Generation (II)
MAI = (S + MAIO) modulo N if HSN 0 (random hopping)
with:
S = M’ if M’ < N
S = (M’ + T’) modulo N else
M’ = M modulo [2^Integer(log2(N)+1)]
T’ = T3 modulo [2^Integer(log2(N)+1)]
M = T2 + RNTABLE((HSN xor T1R)+T3)
T1R, T2, T3 ... Different Time ParameterRNTABLE ... Table of 114 Integer numbers
s
Network Design and Consulting
Frequency PlanningExample for MAIO - Management (I)
Frequency group 1x1 reuse / Random Hopping (1, 2, 10, 7, . . . )
10
7
Time (TDMA - frame)
Time (TDMA - frame)
1
35
911
13
1517
2
46
8
TRX0
TRX1
TRX2
TRX3
BCCH
MAIO = 2
MAIO = 8
MAIO = 14
TRX0
TRX1
TRX2
BCCH
MAIO = 4
MAIO = 10
MAIO = 16 TRX3
TRX0
TRX1
TRX2
TRX3
BCCH
MAIO = 0
MAIO = 6
MAIO = 12
1012
14
1618
1 2 3 4 5 6 1817161514131211987 10
1214
16
182
4
68
7
9
. . .
. . .
. . .
. . .
. . .
. . .
. . .
. . .
. . .
Time (TDMA - frame)
s
Network Design and Consulting
Frequency PlanningExample for MAIO - Management (II)
Frequency group 1x1 reuse / Random Hopping (1, 2, 10, 7, . . . )
1 2 3 4 5 6 1817161514131211987 10
Avoid Co - channel collision:
min # RF = number of hopping TRX (example 9 frequencies)
Avoid Adjacent - channel collision:
only odd or even RF numbers on air at same time
Minimum total number of frequencies for hopping system with MAIO - Management = 2* number of hopping TRX of site (18 frequencies in example)
TRX0
TRX1
TRX2
TRX3
BCCH
MAIO = 2
MAIO = 8
MAIO = 14
TRX0
TRX1
TRX2
BCCH
MAIO = 4
MAIO = 10
MAIO = 16 TRX3
TRX0
TRX1
TRX2
TRX3
BCCH
MAIO = 0
MAIO = 6
MAIO = 12
s
Network Design and Consulting
Frequency PlanningExamples for frequency groups and MAIO - Assignment
TRX0
TRX1
TRX2
TRX0
TRX1
TRX2
TRX0
TRX1
TRX2
TRX0
TRX1
TRX2
TRX0
TRX1
TRX2
TRX0
TRX1
TRX2
TRX0
TRX1
TRX2
HSN = 1
HSN = 2
HSN = 3
TRX0
TRX1
TRX2
TRX3
f B
f B
f B
f B
f B
f B
f B
f C
f C
f C
f C
f C
f C
f C TRX3
TRX0
TRX1
TRX2
TRX3
BCCH
BCCHBCCH
BCCH
BCCHBCCH
BCCH BCCH
BCCH
MAIO = 0
MAIO = 1MAIO = 0
MAIO = 0
MAIO = 1MAIO = 0
MAIO = 0 MAIO = 1
MAIO = 0
MAIO = 2
MAIO = 3MAIO = 2
MAIO = 2
MAIO = 3MAIO = 2
MAIO = 2 MAIO = 3
MAIO = 2
f A
f A
f A
f A
f A
f A
f A
MAIO = 5MAIO = 4
MAIO = 4
Frequency group: A: 1 4 7 10 13 16B: 2 5 8 11 14 17C: 3 6 9 12 15 18
s
Network Design and Consulting
Frequency Planning ToolsThe Automatic Frequency Planning Process
Input datafrom radio networkplanning tool
•Automatized Planning Routines•Variety of Planning Algorithms
•Setting of planning constraints•Common / Dedicated Band Planning
•Global / Local Parameter Settings
Frequency Assignment
Live Network Data
Minimisation of
interference
•Consideration of FH, PC, DTX
Evaluation of the assignments•C/I and FER plots•C/I and FER analysis on per carrier basis
Evaluation of the assignments•C/I and FER plots•C/I and FER analysis on per carrier basis
s
Network Design and Consulting
Frequency Planning ToolsThe SIEMENS Advanced Automatic Frequency Planning Tool
Efficient algorithms for different optimization targets: Minimizing global interference
Minimizing worst interfering cell relations
...
Features for advanced network planning strategies Frequency hopping
Power Control
Discontinuous transmission
Graphical evaluation of frequency assignments based on C/I
FER
Very good results in European research program COST 259 benchmarks in quality of result at short execution times (typically seconds to minutes)
High performance proved in live networks with different customers
s
Network Design and Consulting
Frequency Planning Tools Consideration of Radio Link Control Options
Automatic consideration of hopping gains and
interference reduction due to PC and DTX on cell basis
during
• interference matrix calculation
• optimum assignment of frequencies by using highly
efficient optimisation algorithms
Graphical evaluation of the assignment results based
on FER
s
Network Design and Consulting
EIRP - Path
loss
EIRP - Path
loss
Potentially serving signal SC
Potentially interfering signal SI
Ai
Ai
Ai 50%
Frequency Planning ToolsGeneration of the interference matrix
s
Network Design and Consulting
Frequency Planning Tools
Required C/I in FH-GSM (TU3), Cyclic Hopping
FH Gains as determined via Real Network Simulations
Shift: 6.5 dB 13.5 dBGain: up to 7 dB
NH2 Ch3 Ch4 Ch5 Ch8 Ch
50%
s
Network Design and Consulting
Frequency AssignmentFrequency Reuse & C/I values (Non Hopping)
Required no. of frequencies Cluster size / Reuse distance: q = SQRT(3*N) C/I rule of thump: C/I abs 1,5 * N2
N Anzahl f q C/I [dB]2 6 2,45 7,783 9 3,00 11,304 12 3,46 13,805 15 3,87 15,746 18 4,24 17,327 21 4,58 18,668 24 4,90 19,829 27 5,20 20,8510 30 5,48 21,7612 36 6,00 23,3415 45 6,71 25,2818 54 7,35 26,8720 60 7,75 27,78
s
Network Design and Consulting
Frequency Planning Tools
Required C/I in FH-GSM for different environments
Typical frequency hopping gainThe following table shows the typical gain from frequency hopping in a GSM 900 network (example of the signal-to-noise ratio required to obtain 0.2% residual BER for class 1b bits):
Frequency hopping TU3 TU50 HT100
None 11.5 7.5 6.82 frequency 10.0 6.5 6.74 frequency 8.25 6.0 6.68 frequency 7.5 6.0 6.616 frequency 6.75 6.0 6.6
Source: SIEMENS TED-BSS
s
Network Design and Consulting
Frequency Planning Tools Analyses of FER
Graphical FER analysis of an SFH network
1x3 reuse, 0,3 fractional load 1x3 reuse, 0,6 fractional load
< 1%
< 2 %
< 3 %
3 %
FER in %
< 1%
< 2 %
< 3 %
3 %
FER in %
s
Network Design and Consulting
Frequency AssignmentExample for Tool-supported Planned Reuse (I)
3
33
4
443
22
3
43
3
33
4
22
4
44
4
443
33
2
44
2
44Network Example:
• 11 Sites
• 33 Cells• 6 cells 2 TRX• 12 cells 3 TRX• 15 cells 4 TRX• 33 TRX BCCH• 75 TRX TCH
No. of TRX
s
Network Design and Consulting
Frequency AssignmentExample for Tool-supported Planned Reuse (II)
Given Spectrum: 42 channels
12 frequencies for BCCH - TRX
30 frequencies for TCH - TRX (hopping)
Reuse of: 4 30/4 = 7.5 frequencies per cell in average
5 30/5 = 6 frequencies per cell in average
6 30/6 = 5 frequencies per cell in average
7 30/7 = 4.2 frequencies per cell in average
Network Example:• 11 Sites• 33 Cells• 6 cells 2 TRX• 12 cells 3 TRX• 15 cells 4 TRX
s
Network Design and Consulting
Frequency AssignmentExample for Tool-supported Planned Reuse (III)
Planning Rule: (example)
1 Hopping TRX 3 frequencies2 Hopping TRX 4 frequencies3 Hopping TRX 6 frequencies
No. of assigned
frequencies for FH
Frequency Reuse Factor:
156 / 33 = 4.7 frequ. / cell in average
30 frequ. / 4.7 frequ. per cell = 6.3
4 / 6
4 / 66 / 4
3 / 4
3 / 44 / 3
2 / 3
4 / 66 / 4
No. of TRX
Network Example:• 11 Sites• 33 Cells• 6 cells 2 TRX• 12 cells 3 TRX• 15 cells 4 TRX
s
Network Design and Consulting
Frequency AssignmentExample for Tool-supported Planned Reuse (IV)
Separations for hopping TCH:• Intra cell separation: 3
• Intra site separation: 1
• Neighbour separation: 1
Interference Matrix for hopping TCH:• co-channel: C/I curve 7 dB (50% probability)• adjacent channel: C/I curve -6 dB (50% probability)
MAIO and HSN:• HSN = 0 for all cells (cyclic hopping)• MAIO = 0 for TRX1 (TRX0 = BCCH)
• MAIO = 1 for TRX2• MAIO = 2 for TRX3 etc.
s
Network Design and Consulting
Frequency AssignmentExample for Tool-supported Planned Reuse (IV)
etc....
Example for a site list:Site Id Sector TRX BCCH f1 f2 f3 f4 f5 f6 MAIO HSN0001 1 0 2 - -
1 1 19 36 41 0 02 0 4 - -2 1 13 18 21 25 30 42 0 02 2 13 18 21 25 30 42 1 02 3 13 18 21 25 30 42 2 03 0 9 - -3 1 15 23 27 32 0 03 2 15 23 27 32 1 0
0002 1 0 12 - -1 1 22 26 34 40 0 01 2 22 26 34 40 1 02 0 8 - -2 1 16 19 28 0 0
s
Network Design and Consulting
Frequency PlanningStrategies Cyclic Hopping - Random Hopping
Cyclic hopping sequence {... f4, f0, f1, f2, f3, f4, f0, f1, f2, f3 ...}, MAIO 0Cyclic hopping sequence {... f1, f2, f3, f4, f0, f1, f2, f3, f4, f5 ...}, MAIO 2
F
r e
q
u
e n
c
y
TDMA frame
f0
f1
f2
f3
f4
Principle of Cyclic Hopping
• Optimum frequency Diversity• Sufficient Interference diversity by avoiding frequency groups• No Interference diversity using frequency groups
Random hopping sequence {... f1, f4, f2, f0, f0, f3, f0, f1, f2, f4, ...}, MAIO 0Random hopping sequence {... f3, f1, f4, f2, f2, f1, f2, f3, f4, f1, ...}, MAIO 2
F
r e
q
u
e n
c
y
TDMA frame
f0
f1
f2
f3
f4
Principle of Random Hopping
• Optimum interference diversity
• Less frequency diversity
s
Network Design and Consulting
Frequency PlanningPlanning of Anchor Frequencies
Each TRX must be assigned with a fixed frequency which belongs to the hopping frequency group of this TRX
In case of disabling FH the TRX transmit the anchor frequency
Tool Supported Planning of anchor frequencies
2 - 5 - 8 - 11 - 14 - 17
Example Frequency group 1x3 reuse: A: 1 4 7 10 13 16B: 2 5 8 11 14 17C: 3 6 9 12 15 18
TRX0
TRX1
BCCH
TRXFREQ = 183 - 6 - 9 - 12 - 15 - 18
TRX0
TRX1
TRX2
TRX3
BCCH
TRXFREQ = 7
TRXFREQ = 13
TRXFREQ = 4
MOBALLOC = 1- 4 -7- 10 - 13 - 16 MOBALLOC = 1 - 4 - 7 - 10 - 13 - 16 MOBALLOC = 1 - 4 - 7 - 10 - 13 - 16
TRX0
TRX1
BCCH
TRXFREQ = 2......
......
s
Network Design and Consulting
BSS - Database Parameters for Radio Control Features
Planning & Optimisation of Frequency Hopping in GSM Networks
s
Network Design and Consulting
SpecificationName
Object/Package
DB Name Range Meaning
CA BTS/PKGBTSB
CALL 0...1023&...&0...1023
Cell Allocation:list of all frequencies used in this cell except theBCCH-frequency
HOPP_MODE BTS/PKGBTSO
HOPMODE BBHOPSYNHOP
Flag indicates whether baseband or synthesizerhopping is used
FH_SYS_NUMBER
FHSY FHSY 1...10 Number to identify a frequency hopping system
HSN FHSY HSN 0...63 Hopping sequence number0: cyclic hopping1...63: random hopping
MA FHSY MOBALLOC 0...1023& ...&0...1023
Mobile allocation:list of frequencies within the FH system(maximum number of frequencies = 16).
FH_SYSTEM_ID CHAN FHSYID 0...10 Specifies the frequency hopping system (givenby FHSYN) to which a channel shall belong. (0:no hopping)
MAIO CHAN MAIO 0...63 Mobile allocation index offset: defining thestarting frequency (number in the MA frequencylist) for a hopping sequence at a certain framenumber FN, i.e. different channels using thesame FH system shall have different MAIO.
BTS_ISHOPPING BTS/PKGBTSO
HOPP TRUEFALSE
Flag to enable/disable FH within the BTS
Source: M. Doss, ICN CA CV D 22
Database ParametersAdministration of Database Parameters for FH (I)
s
Network Design and Consulting
Database ParametersAdministration of Database Parameters for FH(I)
Maximum No. Of Hopping FrequenciesTheoretical Limit: 16 Frequencies (BFH)Practical Limit: 15 Frequencies (SFH)
In case of SFH all Time Slots on BCCH TRX (BCCH TS, SDCCH TS as well as TCH TS) must not hop
In case of BFH the BCCH TS must not hop (FHSYID = 0)
s
Network Design and Consulting
L_RXQUAL_XX_P
RXLEV63
U_RXQUAL_XX_P
0
7
RXQUAL
POW_RED_STEP_SIZE
L_RXLEV_XX_P U_RXLEV_XX_P
Power Decrease(good quality)
Power Decrease(good level)
Power Increase(bad level)
Power Increase(bad quality)
Database ParametersAdministration of Database Parameters for Power Control (I)
s
Network Design and Consulting
Specification Name DB Name/Object Range Meaning
MS_TXPWR_MAX MSTXPMAX /BTSB
2...150...15
maximum TXPWR an MS may use in the serving cell
BS_TXPWR_RED PWRRED/ TRX
0...6 Static reduction of the TRX output power:BS_TXPWR_MAX = PBTS - 2 * PWRRED
POW_INCR_STEP_SIZE
PWRINCSS/ PWRC
DB2,DB4,DB6
Step size for power increase in dB
POW_RED_STEP_SIZE
PWREDSS/ PWRC
DB2,DB4 Step size for power reduction in dB
P_CONFIRM PWRCONF/PWRC
1...31 Maximum interval for waiting for a confirmation of the new transmitpower level. unit: 2 TSACCH
L_RXLEV_DL_PL_RXLEV_UL_P
LOWTLEVDLOWTLEVU
0...63 RXLEV threshold on downlink/uplink for power increase
U_RXLEV_DL_PU_RXLEV_UL_P
UPTLEVDUPTLEVU
0...63 RXLEV threshold on downlink/uplink for power decrease
L_RXQUAL_DL_PL_RXQUAL_UL_P
LOWTQUADLOWTQUAU
0...7 RXQUAL threshold on downlink/uplink for power increase
U_RXQUAL_DL_PU_RXQUAL_UL_P
UPTQUADUPTQUAU
0...7 RXQUAL threshold on downlink/uplink for power decrease
P_CON_INTERVAL PCONINT / PWRC 0...31 Minimum interval between changes of the RF transmit power level
Database ParametersAdministration of Database Parameters for Power Control (II)
s
Network Design and Consulting
Specification NameObject/Package
DB Name Meaning
DTX indicator uplink BTS/BTSO DTXUL 0: MS may use DTX (if possible)1: MS shall use DTX2: MS shall not use DTX
DTX indicatordownlink
BTS/BTSO DTXDL FALSE: downlink DTX disabled at BTSTRUE: downlink DTX enabled at BTS
Database ParametersAdministration of Discontinuous Transmission (DTX)
s
Network Design and Consulting
No Frequency Hopping
0
1
2
3
4
5
6
7
0,1 1 10 100
FER@90% [%]
RX
QU
AL
@9
0%
2% FER
Cyclic FH 2 Frequencies
0
1
2
3
4
5
6
7
0,1 1 10 100
FER@90% [%]
RX
QU
AL
@9
0%
2% FER
Cyclic FH 4 Frequencies
0
1
2
3
4
5
6
7
0,1 1 10 100
FER@90% [%]
RX
QU
AL
@9
0%
2% FER
Cyclic FH 8 Frequencies
0
1
2
3
4
5
6
7
0,1 1 10 100
FER@90% [%]
RX
QU
AL
@9
0%
2% FER
Database ParametersAdministration of Database Parameters for FH - RxQual
s
Network Design and Consulting
Database ParametersParameter Settings for Control Loops using RxQual Measurements
Handover: Default Default Default (No Hopping) SFH (> 8 HF) BFH
HOLTHQUDL: 4 (5*) 5 ... 6 4 ... 5HOLTHQUUL: 4 (5*) 5 ... 6 4 ... 5HOAVQUAL: 8-2 8-2 8-2
Power Control:LOWTQUAD: 3 (4*) 4 ... 5 3 ... 4LOWTQUAU: 3 (4*) 4 ... 5 3 ... 4UPTQUAD: 1 1 1 UPTQUAU: 1 1 1PAVRQUAL 4-1 4-1 4-1
*) as recommended by database planning
s
Network Design and Consulting
Database ParametersFlexible Configuration of different Hopping Modes
Flexible configuration of the Hopping Mode for each cell of a BSC via BSS parameter:
• No Hopping
• Baseband Hopping
• Synthesizer Hopping
Configuration of the first SDCCH on the BCCH TRX mandatory
s
Network Design and Consulting
Optimisation Aspects
Planning & Optimisation of Frequency Hopping in GSM Networks
s
Network Design and Consulting
Points of Examination
Impact of RLC on Network Quality
Impact of RLC on Network Capacity
Comparison BFH - SFH
Comparison SFH Tight Reuse Implementation Cases
Optimizing Radio Parameters
(Thresholds for Handover and Power Control)
Optimisation AspectsField Trials
s
Network Design and Consulting
Optimisation AspectsNetwork Measurements Performance Measurements:
• Performance and Quality Indicators- Dropped Call Rate- TCH Drop Rate (Loss of Connection, Loss during Handover)
• Handover Statistics (Inter, Intra, Causes, Failures)
• Uplink Interference Measurements on Idle TCH
• SDCCH Performance
Test Mobile Measurements:
• RxLev, RxQual, FER, SQI (Speech Quality)
Tracer Measurements: Abis Protocol Analyses
s
Network Design and Consulting
0.0%
1.0%
2.0%
3.0%
4.0%
5.0%
No Hopping 4.55% 1.23% 1.90%
SFH 1x1 4.39% 1.08% 1.59%
SFH 1x3 4.40% 1.04% 1.52%
Dropped Call Rate TCH Drop Rate Call Drop Rate BS
Optimisation Aspects Performance Measurements: Quality Indicators
Comparison of Drop Rates: Non Hopping - SFH 1x3 - SHF 1x1
~ 4 % Improvement
~ 14 % Improvement
~ 18 % Improvement
s
Network Design and Consulting
0.0%
20.0%
40.0%
60.0%
80.0%
100.0%
No Hopping S1 / SFH 1x3 SFH 1x3
Uplink Quality (Per) Downlink Quality (Per) Uplink Strengh (Per) Downlink Strength (Per)
Distance (Per) Better Cell (Per) Direct Retry (Per)
Optimisation AspectsHandover Performance (I)
Handover Performance: Inter cell HO - Causes
Percentageof
Quality HO increases
Parameter Optimisation
S1/SFH 1x3 SFH 1x3HAND:HOAVQUAL 5-2 8-2PWRC: LOWTQUAU 3 4
s
Network Design and Consulting
Optimisation Aspects Handover Performance (II)
0,0%
10,0%
20,0%
30,0%
40,0%
No Hopping 34,0% 1,1%
S1 / SFH 1x3 32,9% 3,1%
SFH 1x3 30,4% 0,7%
InterCellHO/TCHAssignm IntraCellHO/TCHAssignm
S1/SFH 1x3 SFH 1x3HAND:HOAVQUAL 5-2 8-2PWRC: LOWTQUAU 3 4
Optimisation of Database Parameters Percentage of Intra cell HO decreased again
UL-PC enabled from the beginning (incl. No hopping)
s
Network Design and Consulting
Optimisation Aspects Handover Performance (III)
0%
20%
40%
60%
80%
100%
120%
Intra Downlink Quality 54% 62% 74%
Intra Uplink Quality 46% 38% 26%
No Hopping S1 / SFH 1x3 SFH 1x3
S1/SFH 1x3 SFH 1x3HAND: HOAVQUAL 5-2 8-2PWRC: LOWTQUAU 3 4
RXQUAL_DL > 5RXLEV_DL > 35
RXQUAL_UL > 5RXLEV_UL > 31
s
Network Design and Consulting
Rate of idle traffic channels per interference band
96.3%
98.7%
98.9%
98.9%
99.3%
1.9%
1.2%
1.1%
1.1%
0.7%
0.9%
0.0%
0.0%
0.0%
0.0%
0.9%
0.0%
0.0%
0.0%
0.0%
94.0% 95.0% 96.0% 97.0% 98.0% 99.0% 100.0% 101.0%
Non Hopping
BFH
SFH
SFH + PC
SFH + PC + DTX
Interference Band 1 Interference Band 2 Interference Band 3 Interference Band 4 Interference Band 5
Optimisations AspectsInfluence of RLC Features on Idle Traffic Channel Measurements
Less percentage of measurements in higher bands Reduction of Interference
s
Network Design and Consulting
Implementation of PC & DTX in SFH 1x1
Optimisation AspectsHandover Performance - Influence of PC & DTX
0%
10%
20%
30%
40%
50%
60%
Uplink Quality (Per) 14% 23%
Downlink Quality (Per) 20% 16%
Uplink Strengh (Per) 8% 7%
Downlink Strength (Per) 7% 6%
Distance (Per) 0% 0%
Better Cell (Per) 49% 47%
Direct Retry (Per) 2% 1%
SFH 1x1 PC, DTX SFH 1x1
Increase of no. of uplink quality handovers
Decrease of no. of downlink quality handovers
Power Control was enabled in Downlink additionally
Reduction of Downlink Quality HO
s
Network Design and Consulting
Optimisation Aspects Influence of Cell Synchronisation
0,00%
0,20%
0,40%
0,60%
0,80%
1,00%
1,20%
Synchronized cells 1,08 %
Change of HSN 1,07 %
TCH Drop Rate SFH Reuse 1x3
1,00 %
1,09 %
TCH Drop Rate SFH Reuse 1x1
Results of Changing HSN - 1x3 and 1x1 in Reference Cells
Different HSN within one site
No Synchronisation between the cells
Synchronization
No Synchr. No Synchr.
HSN = a HSN = a
HSN = b
s
Network Design and Consulting
SFH 1x1: Reduction of frequencies
0,0%
1,0%
2,0%
3,0%
4,0%
5,0%
PC, DTX SFH 1x1 4,7% 1,22%
Carr_Red 3Fr SFH 1x1 4,6% 1,16%
Carr_Red 6Fr SFH 1x1 4,6% 1,21%
Dropped Call Rate TCH Drop Rate
Reduction of 3 Frequencies
Reduction of 6 Frequencies
Optimisation AspectsReduction of Frequencies
s
Network Design and Consulting
Optimisation Aspects Drive Test Data - FER Evaluation (I)
FER (full) Samples Percent. Samples
Perc. Sampl Cumul.
Total FER Rate
0% 22883 97,08% 97,08% 0,00%4% 407 1,73% 98,81% 0,07%8% 87 0,37% 99,18% 0,10%
12% 26 0,11% 99,29% 0,11%16% 34 0,14% 99,43% 0,14%20% 46 0,20% 99,63% 0,17%24% 16 0,07% 99,69% 0,19%28% 13 0,06% 99,75% 0,21%32% 10 0,04% 99,79% 0,22%36% 6 0,03% 99,82% 0,23%40% 19 0,08% 99,90% 0,26%44% 6 0,03% 99,92% 0,27%48% 6 0,03% 99,95% 0,28%60% 6 0,03% 99,97% 0,30%88% 6 0,03% 100,00% 0,32%
23571 100% 0,32%
T
T
T
T
T
T
T
T
T
T
T
T
S T
T
T
T
T
T
T
T
T
T
-
T
T
T S
Traffic Channel (TCH)
Slow Associated Control Channel (SACCH)
26 frames = 120 ms
Measurement Mobile reports FER values each SACCH frame
FER can only evaluated in steps of 4 %
s
Network Design and Consulting
Optimisation Aspects Drive Test Data - FER Evaluation (II)
Detailed Evaluation of Non Hopping BCCH vs. Hopping TCH :
Hopping Gain can be seen in RxQual- and FER Distribution
• RxQual values of Hopping samples are worse but FER values are better.
• Better FER samples on the hopping carriers.
FER Distribution (RxLev > 10 & RxQual >4)
Non Hopping BCCH vs. Hopping TCH
0,0%
10,0%
20,0%
30,0%
40,0%
50,0%
60,0%
0 4 8 12 16 20 > 20
FER [%]
No Hopping BCCH Hopping TCH 1x3 Hopping TCH 1x1
RxQual Distribution (RxLev >10 & RxQaul >4)
Non Hopping BCCH vs. Hopping TCH
0,0%
10,0%
20,0%
30,0%
40,0%
50,0%
60,0%
70,0%
5 6 7
RxQual
No Hopping BCCH Hopping TCH 1x3 Hopping TCH1x1
s
Network Design and Consulting
Optimisation AspectsInfluence of Power Control on RxLev , RxQual Distribution
s
Network Design and Consulting
Optimisation AspectsMeasuring FH improvements in the Field
Call drop rates cannot show full FH gains, since SACCH performance is not strongly related to FH
RXQUAL statistics for both uplink and downlink get worse with FH and need to be interpreted -> required RXQUAL
Currently no vendor supports speech quality related FER measurements in the BSS - for downlink, no MS reporting is standardised - for uplink, BS vendor specific implementations are feasible
TEMS drive/walk test can show FH improvement on downlink speech quality
BR6.0 will have measured FER statistics for the uplink and estimated FER statistics for the downlink
s
Network Design and Consulting
Summary
Planning & Optimisation of Frequency Hopping in GSM Networks
s
Network Design and Consulting
Summary SFH Planning & Implementation for the Expansion (I)
The benefits of Frequency Hopping have been successfully verified in Field Trials by using different implementation alternatives (BFH, SFH, loose reuse, tight reuse)
Measurements showed improvements of Quality and Capacity (reduction of frequencies).
Quality Feature Capacity Feature
Using FH in real networks provides measures to enhance the reuse (overall reuse including BCCH frequencies of better than 9 maintaining speech quality at the same time)
s
Network Design and Consulting
SummarySFH Planning & Implementation for the Expansion (II)
It is recommended to implement BFH as a basic feature.• BFH allows for narrow band combining (e.g. filter combiners) with low insertion loss
useful in noise limited scenarios
It is recommended to use SFH in mature high capacity networks• SFH requires wide band combining (e.g. hybrid combiners)
recommended for interference limited scenarios
Implementation of Features does not replace quality and capacity improvements to be achieved via maintenance and network optimisation activities !
s
Network Design and Consulting
SummarySiemens Reference Networks: Synthesiser Frequency Hopping
China
Croatia
Czech Rep.
Germany
Kuwait
RSA
Syria
Taiwan
Thailand
USA
Siemens SFH Networks in
High capacity configurations:
• Network with site configurations up to 4/6/4: 4/4/4, 4/5/4, 4/6/4
• Cells are significantly loaded with traffic
• Call Drop Rate less than 2 % TCH Drop Rate better than 2% (in selected cases better than 1%)
Achievable quality in the networks depends on • coverage situation• available spectrum• Traffic load and traffic distribution• homogeneity of the network and topography of the landscape
s
Network Design and Consulting
SummaryAdditional Information
SIEMENS Technical Descriptions Base Station System (TED-BSS BR.xx)
PLMN SBS Radio Network Parameters (SIEMENS ICN Training Institute)
PLMN SBS Performance Measurements (SIEMENS ICN Training Institute)
ETSI GSM Recommendation GSM 05.05, 05.08 and 04.08