Umts Outdoor RF Design Guidelines
Umts Outdoor RF Design Guidelines
Umts Outdoor RF Design Guidelines
Author T. ASZTALOS
Owner T. ASZTALOS
Organization NCS\Digital Network Design\RAN Expertise
Approver P.SABATIER
Document ID ULITPALIL5R3-1390467857-1964
Version Status Date Author Owner Reviewed by Reviewed Approver Approval Description of changes
date date
1.0 Approved 28-02-2003 P. SABATIER P. SABATIER ARFD Team 10-03-2003 M. LIEM 21-03-2003
2.0 Approved 07-03-2004 P. SABATIER P. SABATIER ARFD Team 10-03-2004 M. LIEM 22-03-2004
3.0 Approved 18-03-2005 P. SABATIER P. SABATIER ARFD Team 28-03-2005 M. LIEM 13-04-2005
3.1 Approved 18-11-2005 P. SABATIER P. SABATIER ARFD Team 28-11-2005 M. LIEM 10-12-2005
4.0 Approved 26-01-2007 P. SABATIER P. SABATIER ARFD Team 08-02-2007 M. LIEM 19-02-2007
5.0 Approved 04-01-2009 P. SABATIER P. SABATIER ARFD Team 16-01-2009 P. ARNAUD 27-01-2009
6.0 Approved 15-11-2010 P. SABATIER P. SABATIER ARFD Team 25-11-2010 P. ARNAUD 13-12-2010
7.0 Approved 21-07-2011 P. SABATIER P. SABATIER ARFD Team 31-07-2011 P. ARNAUD 17-08-2011
8.0 Approved 16-04-2012 P. SABATIER P. SABATIER ARFD Team 27-04-2012 P.SABATIER 12-05-2012 Update UA8x, UA9x
9.0 Approved 12-05-2013 T. ASZTALOS T. ASZTALOS ARFD Team 23-05-2013 P.SABATIER 07-06-2013
10.0 Approved 01-09-2014 T. ASZTALOS T. ASZTALOS ARFD Team 12-09-2014 P.SABATIER 26-09-2014
11.0 Approved 15-10-2015 T. ASZTALOS T. ASZTALOS ARFD Team 25-10-2015 P.SABATIER 10-11-2015
12.0 Approved 06-06-2016 T. ASZTALOS T. ASZTALOS ARFD Team 17-06-2016 P.SABATIER 28-06-2016
13.0 Approved 30-06-2017 T. ASZTALOS T. ASZTALOS ARFD Team 07-07-2017 P.SABATIER 21-07-2017
14.0 Approved 19-05-2020 T. ASZTALOS T. ASZTALOS RAN Expertise 26-05-2020 P. SABATIER 06-06-2020
Team
2.2.2.2 Downlink................................................................................................................................. 16
3 BS PRODUCTS ....................................................................................................................... 17
3.1.5 Channel element capacity for Macro-cell NodeB & RRH ................................................ 31
3.3.2 RF repeater............................................................................................................................ 35
4 HSxPA ...................................................................................................................................... 37
5.1 UMTS co-existence with other technologies in different or same frequency bands. 47
5.2 FREQUENCY OFFSET BETWEEN UMTS & GSM IN THE SAME BAND ................................. 49
5.3 FREQUENCY OFFSET BETWEEN UMTS & CDMA IN THE SAME BAND .............................. 52
5.4 FREQUENCY OFFSET BETWEEN LTE & UMTS IN THE SAME BAND .................................. 52
5.5 FREQUENCY OFFSET BETWEEN UMTS & UMTS IN THE SAME BAND .............................. 53
6.2.2 Comparison between 2100MHz and 900MHz for TMA recommendations ................. 61
6.4.2.1 Assumptions.......................................................................................................................... 64
8.1 OVERVIEW............................................................................................................................... 80
8.5.3 UL/DL effective service area & user rejection ................................................................. 114
8.6.2 PRE-OPTIMIZATION PHASE WITH RNP & ACP TOOLS ...................................................... 117
9.11 SCANNER or CW DATAS IMPORT & COMPARISON WITH STUDIES .................................. 163
10.4.6 ACP with traffic map and load balancing .......................................................................... 196
This document describes the process of dimensioning and designing the radio system of a UMTS
network. It provides a detailed presentation of the different steps to perform.
This document synthesizes several studies done by Core engineering team, on the impact or
gain expected by new features implementation.
The studies have been detailed in documents or presentations which are mentioned in this
document and referenced at its end.
All the parameters settings for 9955 radio planning tool V7.3.2 build 11009 version and
Forsk ASP/ACP tools associated
NOTE: In the case of an existing operator, reusing sites (for economical and administrative
reasons) has a great impact on the cell planning process and can lead to a large increase on the
total number of sites.
Depending on the target quality of service, the service availability, the number of carriers
available and the traffic forecast, the number of sites can dramatically be increased by a factor 2
or even more.
These assumptions should thus be specified in close relation with the customer needs, keeping
in mind all the possible implications in terms of cost, design constraints and global quality.
The traffic assumptions for a W-CDMA network are necessary to define the coverage objectives
together with the offered capacity.
Both are closely linked, and the size of a cell will depend on the services provided, but also on
the total number of captured users.
The traffic assumptions should be based on the busy hour, that is, the average traffic during the
busiest hour of the day.
It will describe, for each service (Speech, CS/ PS, data rate) the offered traffic density (in Erlang
per /km² or duration of communication per hour per km² for circuit switched data and speech,
and kbps/km² for packet switched data).
Services availability:
The first requirement is to define the services which should be supported by the network.
The UMTS specification proposes several services such as voice service, different data rates in
circuit switched mode (Long Constraint Delay mode), and in packet mode (Unconstraint Delay
Data).
Each of these services requires different radio quality in terms of Eb/No, and will have different
impact on the design.
In most of the cases the services required are: speech, PS64, CS64, PS128, PS384, HSDPA and
HSUPA.
Analyze areas
User characteristics (speed, call profile.) should be defined for each area
It will help to determine if the design assumptions are not too constraining in terms of number
of sites, and if the traffic assumptions will be met.
It also helps to evaluate the limitation of the network, and gives baselines for the design
strategy. As in GSM, a W-CDMA network can be coverage or capacity limited.
Coverage limitation means that the total number of cells is determined by the target quality of
coverage.
Capacity limitation means that the final cell count is determined by the traffic assumptions.
When coverage limited, the noise rise assumption given (UL Interference margin) in the link
budget can be relaxed.
• The design should then be done in order to meet exactly the coverage
requirements, in order to minimize the number of sites.
When capacity limited, different solutions may be chosen, such as increasing the noise rise level
(load) in the link budget, resulting in reducing the size of the cells, adding a new carrier, or
implementing a second layer (small cells).
These solutions will depend on the target offered capacity. They will be treated in next parts.
The parameters given in this chapter are based on the Link Budget guidelines (see [R1] & [R2]
linked in part 11 of this document)
The link budget helps to determine the UL available path loss in a cell for a given service at a
required quality of coverage, in a given environment, for a given capacity.
Link budget is essentially used in pre-sales, in order to calculate the Node B and site number for
the analyzed area.
The LB allows:
For each service, calculate the design threshold, based on engineering margin
Cell radius calculation for each environment type, using a propagation model
Several parameters like shadowing margin, Eb/No, orthogonality factor, UL load ….are re-used in
post-sales phase.
The main parameters are discussed hereafter in the paragraph 8.3.3 (for a more details see [R1]
& [R2] linked in part 11 of this document).
The default propagation model is Okumura Hata model @ 900MHz, or 850MHz and Cost-
Hata@1800MHz to 2100MHz. The cell size of the dimensioning service, for a given environment,
will then be used for the cell count estimation.
In the link budget is given the estimated uplink capacity per cell, given the noise rise. It is based
on the N-pole capacity equation:
2.2.2.1 Uplink
The number of simultaneous connected users, for a given service, that can be served per UMTS
cell carrier, in the uplink, is given by the N-pole capacity equation:
1 1
N pole = 1+
(1 + f ). (C / I )
X UL 1
N= 1+
(1 + f ). (C / I )
Where
It should be noted here that the number of radio links available in a cell is dependent on the
loading of the cell, which is determined by the RF designer in the link budget. Based on this
loading the interference margin is automatically processed in the tool, in order to determine the
total margin to be applied in the cell.
2.2.2.2 Downlink
A downlink capacity equation is also available. It characterizes the fact that the total power
available in the base station PA is shared between the power used for common channels, the
power reserved for SHO, and all the users.
The following formula gives a simplified approach to derive the DL capacity per service type.
Eb
−1
N (i)
OF + 0 Target
(PA − CCCH Pwr − SharedPwr )
PG (i ) G SHO
1
DL Sector Capacity (i) =
AF (i ) ( DL MeanF + OF ) PA + UE Th. Noise DL Mean Total PL
With
AF (i ) : Activity factor
DL MeanF : mean ratio between interference extra-cell and intra-cell typically equal to 0.6 - 0.8
PG (i ) : processing gain
GSHO
: is the average gain obtained on the Eb/No due to uplink reselection diversity in soft
handoff.
Such simple UL and DL monoservice capacity formulas are implemented in an Excel sheet [R5]
linked in part 11 of this document.
All the dimensioning approach done in pre-sales is detailed in [R6] linked in part 11 of this
document.
3 BS PRODUCTS
All the technical values specified are typical ones
For more detailed information about fALU WCDMA macro BS (Node-B or RRH) products see
product engineering guides [R14], [R15], [R16] and [R25] of linked in part 11 of this document.
In this section, the RF front-end performances are summarized (except Eb/No values which are
detailed in section 8.5).
3.1.1.1 OneBTS
o 1900/AWS
o 850MHz
• TMA supported
• ACLR: 61dB
• ACS: 55dB
• RF output power @antenna port: 40W nominal RF power (TRDU40-21): 40W x 1 carrier -
20W x 2 carriers
• TMA supported: RETA HW ready (AISG1.1 and AISG 2.0 RET and TMA support with
additional AISG primary controller)
• ACLR: 61dB
• ACS: 55dB
• Bands of operation:
• LTE ready
• TMA supported
• ACLR: 61dB
• RF features:
Note: fALU TRDUs with double PA (TRDU2x WCDMA products) includes 2 RF transmitters
(MCPAs), allowing it to support of VAM (Virtual Antenna Mapping, used to optimize WCDMA power
pooling across PAs) and 2x2 MIMO (allowing increased W-CDMA throughputs). It is LTE-capable,
and ready to supporting mixed mode (Multi-Standard radio, or MSR) WCDMA-LTE.
• Bands of operation:
• discontinuous spectrum across the two PA (support for disaggregated spectrum usage)
• LTE ready
• TMA supported
• ACLR: 61dB
• ACS: 55dB
• RF features:
• discontinuous spectrum across the two PA (support for disaggregated spectrum usage)
• LTE ready
• TMA supported
• ACLR: 61dB
• ACS: 55dB
• RF features:
3.1.1.6 MC-TRX-09
The 9100 Base Station product allows implementation of the UMTS and GSM standard defined
by ETSI in one single cabinet. The MC-TRX-09 module is usable in GSM as well as in WCDMA 900
MHz. Herein we are looking on some aspects of this later quality.
• RX diversity: 2-way
• Bands of operation:
• RF power configuration:
• LTE compliant
• RX diversity: 2-way
• Noise Figure (NF): 3.0 dB typical in the lower 850MHz band (typical sensitivity is –124,5
dBm) & 2.7 dB in the upper PCS &AWS bands (-124.8 dBm sensitivity)
• TMA supported: RETA HW ready (AISG 1.1 - class 1 RET); AISG 2.0
• Bands of operation:
• Nominal operation band: 15MHz & 20MHz; RRH 60-09A provides compatibility with
RRH60-09, but with operational bandwidth increased to 19.6MHz; 20MHz BW in the
upper bands
• RX diversity: 2-way
• LTE compliant
• Noise Figure (NF): 2.0 dB typical in the lower 900MHz band (typical sensitivity is –125,5
dBm) & 1.8 dB in the upper 2100 MHz band (-125.7 dBm sensitivity)
• TMA supported: RETA HW ready (AISG 1.1 - class 1 RET); AISG 2.0
Note: fALU RRHs with double PA (RRH2x WCDMA products) includes 2 RF transmitters (MCPAs),
allowing it to support of VAM (Virtual Antenna Mapping, used to optimize WCDMA power pooling
across PAs) and 2x2 MIMO (allowing increased W-CDMA throughputs). It is LTE-capable, and
ready to supporting mixed mode (Multi-Standard radio, or MSR) WCDMA-LTE.
• Bands of operation:
• Number of carriers :
• discontinuous spectrum across the two PA (support for disaggregated spectrum usage)
• LTE ready
• TMA supported
• ACLR: 61dB
• ACS: 55dB
• RF features:
o 2100MHz
Combination between 45W and 60W PA can be done on a same Node-B with STSRx+y
configuration
For more information about STSRx see documents [R14], [R15], [R16] and [R25] linked in part
11 of this document
• Up to 6 TRDU supported
• PA size :
TRDU2x80-21 may be used as an TRDU1x, with only one active PA, supporting configurations
including STSR1, STSR2 and STSR3, without VAM or MIMO.
By using one TRDU2x80-21 per sector, the following STSRx+y configurations are supported:
• STSRx+y configurations with x≠y, like STSR3+1, are also supported (x≤3 and y≤3)
STSRx+y configurations using TRDU2x80-21 are perfectly suited to support UTRAN sharing.
TRDU2x80-21 is designed to support VAM and MIMO features without the need for a second
TRDU per sector. This allows an efficient deployment of MIMO and VAM with only 3 TRDU per
Node B.
When MIMO or VAM are used, STSR1-VAM (80W+80W per carrier), STSR2-VAM
(40W+40W/carrier), STSR3-VAM (26W+26/carrier) configurations are supported.
6 RX
iCEM D sector
iCCM xTRM MCPA ‘beta’
3 TX D
iCEM
D sector
MCPA ‘gamma’
iCEM
D
Iub
D sector
MCPA ‘alpha’
D
6 RX
D sector
xTRM MCPA ‘beta’
3 TX D
D sector
MCPA ‘gamma’
D
Based on 2G Evolium BTS with 4 GSM TRX. GSM1800/ GSM900 combination is possible d2U unit
implementation which supports 3 CEM cards
40W TRDU supports 2 carriers per PA
The table below is not exhaustive, but gives some configurations that would make sense on the
field when introducing one or several WCDMA carriers. The MC-TRX module has the full flexibility
to support other mixed configuration in the limit of the Power Amplifier capacity.
GSM
GSM
AN AN 3G AN
3G
The Remote Radio Head is a product used to relocate the full radio part from the base station
cabinet to a more adapted location.
For the same capacity, RRH introduces more flexibility on deployment and reduces its cost. On
the other hand, it can eliminate the main losses inside antenna’s chain.
RRH60-xx (or RRH1x60-xx) with only one active PA, may support configurations including STSR1,
STSR2 and STSR3, without VAM or MIMO.
With two RRH60-xx on a sector you may use up to 2 transmitters and 4 receivers in various
configurations including Tx diversity, VAM and MIMO & various STSRx+y modes.
By using two RRH60-xx per sector, the following STSRx+y configurations are supported:
• STSR2+2, i.e. 4 carriers configuration providing 30W per carrier (or 4 carriers providing
STSRx+y configurations using RRH60-xx are perfectly suited to support UTRAN sharing.
STSRx+y configurations each pair of RRHs must be the same type of RRH.
RRH2x60-xx may be used as an RRH1x60-xx, with only one active PA, supporting configurations
including STSR1, STSR2, STSR3 and STSR4, without VAM or MIMO.
By using one RRH2x60-xx per sector, the following STSRx+y configurations are supported:
• STSRx+y configurations with x≠y, like STSR3+1, are also supported (x≤4 and y≤4)
STSRx+y configurations using RRH2x60-xx are perfectly suited to support UTRAN sharing.
RRH2x60-xx is designed to support VAM and MIMO features without the need for a second RRH
per sector. This allows an efficient deployment of MIMO and VAM with only 3 RRH per Node B.
When MIMO or VAM are used, STSR1-VAM (60W+60W per carrier), STSR2-VAM
(30W+30W/carrier), STSR3-VAM (20W+20/carrier) configurations are supported.
All the information concerning Node-B and RRH iCEM, xCEM & eCEM capabilities are available
with the link detailed in [R28].
Note: Nokia WCDMA Flexi Multiradio products including RF Modules and RRH Equipments are
described in details by document [R58] linked in part 11. These, with few exceptions, are with
Software programmable technologies, i.e. are LTE-capable, ready to support mixed mode (Multi-
radio) GSM-WCDMA-LTE.
• Products: Flexi RFM 3-pipe 850 (FXCA), Flexi RF Module 3TX 850 (FXCB), Flexi RF Module 3TX
900 (FXDA, FXDJ), Flexi RFM 3-pipe 900 (FXJB), Flexi RF Module 3TX 1800 (FXEA, FXEB), Flexi
RF Module 3TX 1900 (FXFB, FXFC), Flexi RF Module 3TX 2100 (FRGP, FRGT), Flexi RF Module
3TX 1700/2100 (FRIE)
• Bands of operation:
• MIMO ready
• LTE ready
• MIMO ready
• LTE ready
• RF output power (max total power): 1x60W nominal RF power distributed amongst
supported number of carriers/sector;
• RX diversity: 2-way
• MIMO: no
• Products: Flexi RRH 2-pipe 900 80W (FHDA), Flexi Remote Radio Head 2TX 900 (FHDB), Flexi
Remote Radio Head 2TX 2100 (FRGQ)
• Bands of operation:
• RX diversity: 2-way
• MIMO ready
• LTE ready
• Rx diversity
• 64QAM capable
• A-GPS
• PoE+
• NF: 13dB
• Cube based
• Max power: 1W
• Rx Diversity
• NF: 9dB
• GbE backhaul
3.3 REPEATER
RF link repeater
Advantages:
Drawbacks:
High Cost
Propagation delay in the fiber limits the distance between the remote unit and the
NodeB
3.3.2 RF repeater
Donor Antenna between the repeater and the Node-B: Highly directive antenna
It amplifies the signal both in the UL and in the DL independently or equally depending on the
repeater type.
Advantages:
Drawbacks:
RF noise constraint
The gain setting is very important for a repeater; extra noise is created by the repeater at the
Node-B for the uplink:
Noisewith _ repeater = Noisew / o _ repeater + ( Noiserepeater * GUL _ repeater ) / Coupling_ lossrepeater _ NodeB
With
GUL _ repeater
: repeater gain set in UL
For more details concerning repeater types see [R19] & [R4] linked in part 11 of this document.
4.1 HSDPA
The general approach is to assign a lower priority to HSDPA service than to the R’99 ones.
HSDPA introduces new common channels HS-DPCCH in UL and HS-SCCH in DL. This part
analyzes the impact of these channels and, also presents the throughput calculation method.
The introduction of a new UL common channel induces a new spread signal value weight βhs for
HS-DPCCH added to existing R’99 spread signal value weight βc for DPCCH, and βd for DPDCH.
This has a direct impact on the PDCH UL power calculation, and on the UL Eb/No values.
2 + 2
PDCH (dB ) = MaxUETxPow er (dB ) + 10 * log 2 c 2 d
2
c + d + hs
2 + 2
LOSS = 10 log 2 d 2 c 2
d + c + hs
The Eb/No loss values calculated for each UL R’99 service are the following ones. These losses
must be added UL Eb/No values for only HSDPA users in the cell and not all the cell users.
PS64 1.9
PS128 1
PS384 0.4
PS64 0.9
PS128 0.55
PS384 0.3
Figure 14: HS-DPCCH impact on UL xCEM Eb/No
4.1.2.1 HS-SCCH
HS-SCCH is power controlled. It is calculated, based on an iterative process using the table
below.
1–7 0
8–9 -3
10 – 12 -5
13 – 30 -8
In 9955, the HS-SCCH power control has not been implemented like this. The HS-SCCH power
control is based on Ec/No calculation, and the user had to specify an Ec/No target value. The
Ec/No formula is the following one.
PHS − SCCH
Ec / No =
Ie 1 1
( . Rxdiv _ gain / 10 + ).PBS − .PHS − SCCH + Pn .L. Rxdiv _ gain / 10
I i 10 10
With
After many simulations, based on the same path loss matrix, with our internal RF simulator, and
comparison with 9955 results, the Ec/No target value which provides realistic results in term of
cell throughput, throughput distribution, and area where HSDPA is supported is the following
one: Ec/No= -13dB
The aim is to have the same equivalence path loss point in UL and DL between two cells. This
point has been analyzed in part 2.2 of this document.
In order to calculate cell throughput in 9955, MUG tables have been extrapolated from Alcatel-
Lucent RF simulator considering 55% of PedB profile and 45% of PedA profile.
MUG tables have been provided by R&D and are set in 9955 RNP tool
4.2 HSUPA
The general approach is to assign a lower priority to HSUPA service than to the R’99 ones.
All the HSxPA parameters settings are detailed in [R7] linked in part 11 of this document.
E-DPCCH in UL,
E-AGCH (Absolute Grant Channel), E-HICH (HARQ Indicator Channel) and E-RGCH (Relative
Grant Channel) in DL
This part analyzes the impact of these channels and the throughput calculation method
4.2.1.1 Description
E-DCH traffic is assigned the unused UL load up to the max. A R99 call cannot be dropped due
to an UL load increase caused by HSUPA.However, an increase in E-DCH RoT is comparable to an
increase of R’99 RoT (since R’99 UEs must transmit at higher level to be received correctly by
NodeB), which –if UL iRM Scheduling is activated– could cause the downgrade of high speed UL
PS calls (e.g. PS384 downgraded in PS128).
The Node-B noise figure is required in order to estimate correctly the UL load. It is highly
important to have a good reference value for the Node-B noise as it is the main input for UL
load computation
In R99/R5 networks, the design is usually done assuming 3dB max UL load (50% UL load).
In order to support high E-DCH throughput, this value should be increased, drawback is:
Coverage reduction for R99 traffic (mix carrier) considering the same R’99 dimensioning
service without downgrading
Higher interference
In order to limit the amount of interference, correct neighboring declaration is needed => UL
RSSI cleaning strategy required
The UL load recommended for HSUPA is 75%, so 6dB, in order to have good throughputs, which
can induce a reduction on R’99 UL coverage for mixed carrier configuration.
% UL load 50% 65% 75% 50% 65% 75% 50% 65% 75%
(delta between supported service area supported @ 50% UL load & with other %of UL load)
UMTS 2100 93% 88% 83% 91% 86% 76% 91% 84% 70%
(-5%) (-5%) (-15%) (-6%) (-21%)
(-10%)
In such cases a user which was able to establish a call at cell edge with 50% UL load; will have a
high risk of CAC failure at the same position with 75% UL load.
4.2.2 DL IMPACT
E-AGCH is enough for early deployment. In case there are two users, 2 TTI will be necessary to
grant both users.
E-RGCH is enough (up 15 signatures). The E-RGCH power is negligible, it carries one bit per user
signature
The activity factor of the E-AGCH and E-HICH should be low in early E-DCH deployment.
E-AGCH is not transmitted all the time (as for the HS-SCCH). Once the user is granted, E-AGCH
is not transmitted again.
A fix power is reserve at the RNC level for DL E-DCH channels. This power is preempted from
HSDPA max power and is taken into account in the R99 RNC CAC.
-2.5 dB
E-AGCH
per user signature
-8.0 dB
E-HICH
per user signature
The R99 CAC will reject R99 calls earlier than before in case of highly loaded cell.
2100MHz Node-B 45W PA
Speech CS 64 PS 64 PS 128 PS 384
30 m cable + 0.4dB jumper
For each cell, the remained UL noise rise available is evaluated, considering R’99 has the highest
priority.
Based on this, the path loss prediction, the UE power acceptable is evaluated in order to respect
the max UL load in the cell.
Then with the intra cell and extra cell interferences calculation, the max Eb/No value is calculated
per HSUPA mobile.
The throughput is obtained with the correspondence table between Eb/No value and
Throughput.
The carrier strategy allocation for HSxPA deployment based on simulations is detailed in [R7]and
[R37] linked in part 11 of this document
fALU HSxPA implementation recommendations on existing R’99 outdoor networks are the
following ones:
• 2 carriers available
o 2 shared carrier R’99 & HSxPA
• 3 carriers available
o Option 1
▪ 1 carrier R’99
▪ 2 shared carriers R’99 & HSxPA
or
o Option 2:
▪ 3 shared carriers R’99 & HSxPA
The HSPA+ benefits mobile operators and end users in the following aspects:
64QAM: DL 64QAM allows the use of 64QAM in HSDPA to increase the number of bits per
symbol and thus to obtain higher transmission rates. The peak rate at the MAC layer can reach
21 Mbit/s.
2x2MIMO: MIMO increases transmission rates through space multiplexing and improves channel
qualities through space diversity. The network side can dynamically select single- or dual-stream
transmission according to channel conditions. The peak rate at the MAC layer can reach 28
Mbit/s.
DL Enhanced Layer 2: This feature allows Uu L2 to use flexible PDU size on RLC layer and
segmentation on MAC layer. The feature prevents the L2 from becoming the bottleneck of
higher Uu rate increased by MIMO and 64QAM.
CPC: CPC allows the uplink and downlink transmissions to take place at periodic intervals. This
feature reduces the transmitted power (and thus increases the UE battery life) because the UE
does not have to monitor and transmit overhead channels every TTl. This reduction in the
transmitted power also helps to increase the uplink capacity by decreasing the total
interference. This improvement is especially significant when there are users who transmit data
infrequently as VoIP users. CPC feature consists of DL-DRX, UL DTX and HS-SCCH Less
Operation.
Enhanced CELL_FACH: Enhanced CELL_FACH operation allows the use of HSDPA technologies
for the UEs in the CELL_FACH, CELL_PCH, and URA_PCH state. The purpose is to increase the
peak rates in these states and reduce the signaling transmission delay during service setup or
state transition with the result improving the user experience.
The 64QAM and MIMO are the most important and commercialized features of the HSPA+
technology.
From the aspects of mobile operators and terminal users, the HSPA+ has the following
advantages:
- Expand the downlink capacity of the network and obtain higher spectral efficiency in the
case of higher SNR.
Using the space diversity method, the MIMO technology adopts the multi-antenna
technology on the transmitting side and receiving side to improve the transmission
capacity of the radio communication system by several times without increasing transmit
power and bandwidth in the high SNR environment (the transmission capacity is in
proportion to the number of antennas). Theoretically, if the HSPA+ MIMO technology is
adopted, the peak rate is 27.952 Mbit/s (Formula: TB_Size/TTI = 27952 / 2ms x 2 (double
data stream) = 27.952 Mbit/s).
For operators, the use of the HSPA+ can reduce the unit cost for transmitting every
mega bytes of data stream, increase average system capability, enhance downlink data
service performance of a UE, and improves the cell throughput.
The HSPA+ offers a higher data transmission rate, short service response time, and
reliable service performance for a UE. Therefore, it improves the service experience of
the UE.
Mobile operators concern about the expenses for building an HSPA+ network. This
depends on the equipment price and the service strategies of a single operator. As a
high-rate data service enhancement technology in the WCDMA R5, the HSPA+ is
compatible with the HSPA of earlier versions and R99. The operator can upgrade NodeBs
in the existing WCDMA R99 network to introduce the HSPA+ with little impact on the
existing architecture. This helps shorten the network construction period and protect the
investments of the operator.
UE requirement
In R7, six categories of UEs (Category 13–Category 18) are increased to support the HSPA+;
category 13 and category 14 support only the 64QAM; category 15 and category 16 support
only the MIMO; category 17 and category 18 support the 64QAM and MIMO, but cannot use the
two technologies at the same time.
In R6, the HSDPA increases a transmission channel (HS-DSCH) and three physical channels, that
is, HS-SCCH, HS-PDSCH, and HS-DPCCH. The HSPA+ still uses the preceding channels of the
The basic method of implementing the adaptive modulation and coding (AMC) technology is to
measure the quality of downlink channels and adjust the coding and modulation solution in an
adaptive manner based on the measurement result (expressed by CQI) to select the appropriate
modulation and coding rate to maximize the data transmission rate. The 1/3 Turbo code is the
basic code. The modulation modes include QPSK, 16QAM, and 64QAM. The 64QAM high-order
modulation which is a new function of the HSPA+ is used to improve the downlink peak rate.
4.4.2.2. HARQ
The Hybrid Automatic Repeat Request (HARQ) is an error correction technology. The HARQ
combines the forward error code (FEC) and automatic Repeat Request (ARQ) technologies. R99
adopts the traditional ARQ method that is implemented on the RLC layer. Similar to the HSDPA,
the HSPA+ adopts the SAW HARQ protocol. If the SAW HARQ protocol is adopted, for each
channel, the next data packet is transmitted only when the correct acknowledgement
information of the previous data packet is received. The protocol is simple, but the channel
utilization is low. The SAW HARQ protocol can solve the problem of low channel utilization.
4.4.2.3 Schedule
Similar to the HSDPA, the HSPA+ provides downlink HS-DSCHs for all the users to transmit data.
Resources are shared through code multiplexing and time multiplexing.
Before the Layer 2 Enhancement technology is introduced in the R7 protocol, RLC PDU Size is
set to a fixed value. Due to the great change of the transmission of the Uu interface, RLC PDU
Size is usually set to 320 bits or 640 bits. By default, RLC Window Size is set to 2048. According
to the preceding typical configuration, the RTT delay between the data sending to the receiving
of the acknowledge message is 100 ms, and the supported highest transmission rate is
13.1Mbps (Formula: 640bits x 2048/0.1s = 13.1Mbps). This cannot meet the high-speed
requirement of the HSPA+. Therefore, the Layer 2 Enhancement technology is introduced. The
basic principle of the Layer 2 Enhancement technology is to introduce a variable length PDU in
the RLC layer. According to the protocol, the RLC layer supports up to 1500-byte PDU. In
addition, to support the Layer 2 Enhancement technology, MAC-ehs is introduced in the MAC,
and the related HS-DSCH FP changes.
The 64QAM technology adopts higher order modulation to provide data traffic higher than the
HSDPA by quickly adjusting downlink modulation and coding mode in better radio environment.
Theoretically, if the 64QAM technology is adopted, the peak rate is 21.096 Mbit/s (Formula:
TB_Size/TTI = 42192 / 2ms = 21.096 Mbit/s)
4.4.2.6 2x2MIMO
Using the space diversity method, the MIMO technology adopts the multi-antenna technology
on the transmitting end and receiving end to improve the transmission capacity of the radio
communication system by several times without increasing transmit power and bandwidth in the
high SNR environment (the transmission capacity is in proportion to the number of antennas).
Currently, the typical scenario is 2x2MIMO, that is, dual-fed and dual-receiving mode.
Theoretically, if the MIMO technology is adopted, the peak rate is 27.952 Mbit/s (Formula:
TB_Size/TTI = 27952 / 2ms x 2 (double data stream) = 27.952 Mbit/s) Note that the 64QAM and
MIMO cannot be configured in the R7 protocol at the same time.
Below a table where all the isolation required between UMTS and other technologies (except
LTE) is summarized. More details are available in guideline referenced in [R30] in part 11.
5.2 FREQUENCY OFFSET BETWEEN UMTS & GSM IN THE SAME BAND
3GPP recommends blocking -47dBm GMSK signal with a 2.8MHz offset in UL, and -56dBm in DL.
For both case analyzed below the degradation target is to have a sensitivity degradation less
than 0,5dB or a capacity loss lower than 5%.
Recommendations below have been done considering hopping TCH as GSM adjacent channels
of UMTS900 band.
The GSM channels adjacent of the UMTS band must support TCH with frequency hopping.
Below two cases are analyzed case where GSM and UMTS are co-localized on a same site and
case where they are not and so the worst configuration then is when GSM cell edge is close to
UMTS site
• Standard filter whose attenuation starts at 2.2MHz offset from center frequency
which provides an attenuation higher than 50dB after 2.4MHz frequency offset
from center frequency
• Reduced filter whose attenuation starts at 2.0MHz offset from frequency band
center and provides more than 50dB attenuation after 2.2MHz frequency offset
from frequency band center. The impact reduced filter in DL is negligible
Standard filter
This implies that 5MHz must be free in upper or lower edge of the operator GSM
frequency band.
Reduced filter
This implies that 4.6MHz must be free in upper or lower edge of the operator GSM
frequency band.
Co-located case:
For the surrounded area of UMTS900/GSM900 cluster where only GSM900 is deployed, a
2.2MHz frequency offset is sufficient; this induces to free 4.2MHz.
This implies that 4.2MHz must be free in upper or lower edge of the operator GSM frequency
band.
ALU recommendation is to have 2,2MHz frequency offset without any frequency hopping
constraint
This implies that 4.2MHz must be free in the GSM frequency band on the area where UMTS is
deployed.
In order to avoid interferences between GSM900 area where all operator 900MHz band is used
and UMTS900 cluster, it is necessary to have a buffer zone around UMTS900 cluster where only
channels which respect the previous recommendations are used.
This buffer zone is not homogeneous because generally environment is not homogeneous, as
sites can be positioned on a small mountain, so they have more coverage impact than those in
city center.
For this reason the rule to respect to define sites where all 900MHz band of an operator can be
used without impact UMTS900 cluster is the following one:
Let’s consider cell A as a UMTS900 cell of a UMTS 900MHz cluster, and cell B a GSM900 cell to
re-use channels inside UMTS900 band. BCCH cell B must respect
5.3 FREQUENCY OFFSET BETWEEN UMTS & CDMA IN THE SAME BAND
3.385MHz frequency spacing should be respected between UMTS and CDMA frequency band
3.385MHz
270 kHz
Figure 25: Frequency spacing rule between an UMTS carrier and a CDMA carrier
All the analysis is detailed in the document [R31] linked in part 11 of this document.
5.4 FREQUENCY OFFSET BETWEEN LTE & UMTS IN THE SAME BAND
The requirement is that there is less than 0.5dB degradation on thermal noise floor on each
technology
No frequency gap required between LTE and UMTS. Studies for LTE& UMTS collocated and non-
collocated cases case have been done in guidelines referenced as [R34],[R35]and [R36] in part
11. For both cases no guard band is required to ensure less than 5% capacity degradation which
corresponds to less than 0.5dB sensitivity degradation.
5.5 FREQUENCY OFFSET BETWEEN UMTS & UMTS IN THE SAME BAND
5MHz frequency spacing should be respected between two UMTS frequency bands.
The following curves have been established considering the Tx filter and Rx filter of ALU product.
They show the capacity loss vs the frequency offset between two adjacent UMTS frequency
bands.
Figure 28: UMTS capacity vs adjacent UMTS interferer spacing (MHz) for co-located case (doted
line) and not co-located case
Several solutions are proposed for having 4,2MHz (co-located case) or 5MHz free (non co-
located case)
Increase the GSM1800 capacity by modifying the frequency plan to reduce the traffic on
GSM900. This can be possible when inter-site GSM 900 is in the same order than the
GSM1800, so generally in dense urban/urban.
Re-optimize the GSM900 frequency plan, in order to reduce the band used.
6 900 MHz
900 MHZ has been created in order to support UMTS technology in rural areas as 2100MHz
necessitates too many sites compared to 900MHz.
Anyway reusing GSM frequency band for UMTS imposes some frequency spacing rules.
We also see in this part what we can expect with 900 MHZ in rural environments and in dense
urban/urban environments where the limitation is essentially due to interferences.
All the solutions specified here for antenna sharing between UMTS900 and GSM900 are also
available for other frequencies like UMTS850/GSM850 or UMTS1900/GSM1900.
6.1.1.1 Description
GSM and UMTS emissions must be separated to avoid inter-modulation. This solution is a “Full
band” solution which means there is no need of frequency planning.
GSM BTS receives the RX signals of the both networks (high linearity LNA)
Attenuators allow reducing the UMTS signals, to be compatible with UMTS Node B.
UMTS Node B is in mode mix TMA. DC Blocks are used on the two 3G ways.
DL impacts:
Nothing in GSM (if GSM is transmitted on one antenna port before UMTS900
implementation)
UL impacts:
Drawbacks
▪ Increase PA power
• GSM BTS and UMTS Node-B must be at the same distance from the antenna
Alcatel-Lucent defines a system requirement, that translates into 2G BTS requirements and
Dual-duplexer requirements.
Without TMA
With TMA
NF < 3.5 dB
The combiner must be placed at BTS level and not at antenna level in order to avoid doubling
the cables along the pylon
▪ Solution 1:
▪ UL impacts:
If GSM transmissions done on the two antenna ports before UMTS900 implementation
then new Hybrid duplexer risk to be required: 3 dB impact on the GSM DL link budget
Some GSM BTS cannot manage TMA; an attenuator must be added when combiner is used in
active mode with 8dB gain
Currently this type of antenna exists only in 900MHz band, but RFS, and power wave are
studying to combine 900MHz, and 1800-2100MHz band, in order to have a solution for bi-band
or tri-band antennas.
Advantages:
Tilt can be changed independently per sector without any impact on the existing sites
No impact on GSM even if GSM900 used two antenna ports before UMTS900
implementation which is not the case for combiner solutions
Drawbacks:
• Antenna width 1.5 larger which induces a highest wind loading than a single
antenna. The pylon must support it
MC-RRH & MC-TRX are solutions to deploy UMTS900 & GSM900 sharing Xpol antenna.
6.2.1 Description
The table below summarizes the different parameters change between 900MHz and 2100 MHz.
45W MCPA output power is used at 2100MHz and 55W SCPA is used at 900MHz.
Dense urban
Rural Suburban
/Urban
UE antenna gain 0 dB 0 dB 0 dB
+6,5 dB/+4,5dB in
+ 7,5dB in UL +10,5dB in UL
2100 MHz with TMA and UL
In rural environments, UMTS 900 should be positioned on GSM 900 sites, and the aim is to have
the same coverage. As GSM900 do not have any TMA, it seems not necessary to use one for
UMTS.
If 900 MHz should replace a 2100MHz site in order to extend coverage, TMA using is not
necessary.
Based on the previous table, the comparison between 900MHz and 2100MHz with TMA, for 40m
cable length in rural area, shows 3 dB UL gain with 900MHz configuration. Such difference
represents more than 70m 7/8” cable length. So 900MHz configuration does not require TMA
since 7/8” cable length do not exceed 110m.
For urban areas TMA are used at 2100MHz to increase chances to cover deep indoor areas, but
with 900MHz, the TMA using is not necessary as the indoor coverage is ensure due to the lower
frequency.
The both link budget have been compared for rural environment considering the following
assumptions:
GSM900
GSM900
140,5
UE output power: 0.8W
UL load: 30%
UL load: 50%
Based on the previous analysis, TMA is required in order to ensure same coverage as GSM900
in case where UL CS64 is the dimensioning service.
For an UL load higher than 30%, a TMA is required for UMTS900 to reach the same coverage
radius than with GSM900.
The impact of 900 MHz on the number of sites compared to 2100 MHz has been simulated
with 9955. The design approach is exactly the same than for 2100MHz, but the UL
dimensioning service could be PS64 in rural, and PS128 in urban.
Based on this study, a 40% reduction on the number of sites can be expected with 900 MHz
compared to UMTS 2100, in rural environment.
900 MHz allows deploying sites at the same positions than GSM 900 ones. 900 MHZ antennas
can re-use GSM900 antennas with dual duplexer using without impacting significantly GSM900
cell radius.
With this solution, the azimuths and tilts cannot be changed without impacting GSM900
network.
Results:
With 47% less 900MHz sites the coverage CDF is quite indentical than with 2100MHz network
6.4.2.1 Assumptions
A study based on simulations has been done in order to show the impact and possible benefits
of UMTS 900 compared to UMTS 2100, for R’ 99, HSDPA and HSUPA.
Several assumptions have been done in order to guarantee a correct performance comparison
between UMTS2100 and UMTS900.
This study has been done based on the same R’99 call profile, on urban environment
Case 1: Same number of site between UMTS 900 & UMTS 2100. This case characterise a
dense UMTS900 deployment in urban area
Case 2: Number of UMTS900 sites deployed must be sufficient to support the same R’99
traffic density over the same area. This case characterise a mature UMTS900
deployment in urban area
Case 3: Number of UMTS900 sites deployed must be sufficient in order to guarantee the
same CPICH QoC than with UMTS2100.
The UMTS 900 sites are positioned only where there are UMTS2100 existing sites. Only electrical
tilt can be changed.
Case1
R99 capacity gain with UMTS900 (Dense urban/ suburban) + 20%/ +40%
HSDPA mean cell throughput gain with UMTS900 (Dense urban/ suburban) + 21%/28%
E-DCH mean cell throughput gain with UMTS900 (Dense urban/ suburban) + 5%
Case2
Site reduction with UMTS900 for a same capacity network (Dense urban/
29% / 44%
suburban)
HSDPA mean cell throughput gain with UMTS900(Dense urban/ suburban) 16% / 21%
E-DCH mean cell throughput gain with UMTS900 (Dense urban/ suburban) + 5%
Case3
Site reduction for a same pilot QoC (Dense urban/ suburban) 58%/ 48%
HSDPA mean cell throughput gain with UMTS900 (Dense urban/ suburban) 16% /21%
E-DCH mean cell throughput gain with UMTS900 ( Dense urban/ suburban) + 1%
The improvement from case2 to case1 doesn’t seem significant and required 30% more sites.
Hot spot
• Improve capacity
UMTS900 sites deployed to ensure better deep indoor RSCP and complete 2100MHz
layer (case3)
This ssolution allows having a homogenous layer with around 50% less
sites than 2100MHz layer.
This solution is appropriate for an earlier phase which requires only deep
indoor coverage improvement but not necessarily a capacity
improvement, as the traffic density supported is 40% to 58% less than
2100MHz layer.
UMTS900 deployed to ensure a better deep indoor RSCP and same capacity than
2100MHz layer
6.4.3.2 Suburban/Rural
• Best compromise as the traffic density supported is between 20% less than
2100MHz layer
• Suburban
Deploy the number of sites required to reach -95dBm RSCP value over
95% of the area. Re-use a maximum of existing GSM sites
• Rural
This part details basic solutions to improve coverage, essentially for rural.
The coverage gain expected is between 15% (from 30m to 40m) to 40% (from 30m to 60m).
The drawback of this solution is the cost, and the difficulty to negotiate some sites over 40
meters.
It cancels the slant polar loss which provides 1.5 dB gains in urban and 2.5 dB gain in rural, but
requires at least a 10 wavelength spacing between antennas. So it can only applied in rural due
to the lack of place and the difficulty of adding an antenna on the roof tops in urban areas.
4 Vpolar antennas per sector which is not realistic even in rural or 2 Xpolar antennas.
But in that case, 4 way receivers just cancel slant polar loss, and it is as efficient as space
diversity.
The only way to have an improvement compare to space diversity solution is to use 21 dBi H65°
antenna, which is possible only in rural, and at 2100MHz.
These antennas are only available in Xpolar. The combination with such antennas and 4 way
receiver gives a better improvement than space diversity.
This is a good solution to eliminate cables losses for any rural site.
It can be easily applied in chain configuration along railways, motorways as optical fiber is
available.
For coverage comparison two cases have been analyzed, considering the UL coverage limitation
The UL path loss difference between Node-B configuration and RRH configuration
is equal to the cable losses between Node-B cabinet and 7/8” top mast
connector.
RRH cell radius configuration > Node-B cell radius configuration is except if
RRH - antenna connector cable losses= Node-B- antenna connector cable losses
UL cable loss with TMA is 0.8 dB (0.4dB due to jumper between TMA an antenna,
0.4 dB between Node-B and TMA) for any cable length.
As 0.4 dB jumper loss can not be reduce, the cable losses which can be add
between RRH and antenna should be lower than 0.4 dB.
Based on this the max cable length between RRH and jumper supported in order
to have RRH 20W cell radius configuration = Node-B cell radius configuration, is
6m.
In case of the cable is higher than 6m which can happen if the pylon cannot
support RRH + antennas then RRH can be positioned at the same area than Node-
B, with a TMA; so the cell radius is the same than with the Node-B configuration
7.1.5 REPEATER
Repeater can be used to extend coverage in all environment types, but the approaches are
different.
In rural, repeaters are used to resolve hole of coverage problems, due to local obstacles.
Generally RF repeaters are used particularly in hilly environment, as optical fibers are not
available.
RF repeater UL and DL gains should be set taking into account the following aspects:
A repeater placed far from the Node-B needs to have a high UL gain in order to be able
to transmit and receive signal.
The aim is to find the best compromise in order to extend coverage enough without polluting
Node-B.
The repeater should be placed not at cell border but before: leads to better signal for users at
cell border (TX Power saved) and reduces desensitization effect
The gains are set to ensure the coverage expected, and not necessarily at maximum output
power (to limit noise rise)
Gt = G BS − L + G Donor + GUL
This rule is also written with another equation: (in dB)
With
GBS
: Node-B antenna gain
GDonor
: Donor antenna gain
GUL
: UL repeater gain set
The % of capacity improvement is directly linked to the % of traffic in the area covered by the
repeater compared to the traffic supported by the serving cell.
The software feature extended and ultra extended cell allows the extension of the
synchronization window in order to support time delay which corresponds to cell radius up to
150km.
The Cell size and Tcell parameters setting are described in document [R18] linked in part 11 of
this document.
Even if the time delay can be managed in order support such cell radius, the radio propagation
conditions must be ideal to reach such distance which means that they must be near the free
space conditions.
Such conditions can only be found over sea, big lakes, or other 200km flat area; no obstacle is
allowed to support such distance.
The second aspect is the earth rotundity; which start to have an impact on a radio link when
distance exceeds 30km, even if the environment is flat.
To avoid the earth interference on the radio link the transmitter and receiver must be in visibility
(LOS).
Considering a 200km flat environment and a receiver height between 1.5m and 2m, the
transmitter height must at least have the following values in order to ensure the different cell
radius.
Transmitter
antenna height 90m 140m 200m 250m 400m 900m
@2100MHz
Transmitter
antenna height 70m 100m 140m 200m 350m 800m
@900MHz
Figure 43: Transmitter height required in flat environment to support long cell radius
The UL load increase has a direct impact on the cell radius. The aim of this part is to propose
solutions in order to keep the same cell radius even with UL load increase. The UL load increase
appears only in dense urban and urban areas, so only these environments area considered here.
This is the same solution, than the one used to increase cell radius in rural environment.
The hardware constraints are the same than those detailed in part 7.1.4.
Advantages Drawbacks
One solution to compensate coverage reduction due to UL load increase is to increase antenna
gain.
The drawback is the height increase compared to 18dBi antenna, which can be hard to
negotiate.
As antenna height is limited in urban areas in major European countries, using 30°
antennas instead of 65° ones is one way to increase by 3dB antenna gain.
30° horizontal beam antennas allows a high gain considering 7° vertical beam width like
current 65° H beam antennas used in urban areas.
To keep same overlap between sectors of a same site is to increase the number of
sector per site.
The mix cabinet (classical Node-B and RRH, see the picture below) can also be used in
order to increase the number of sectors per sites.
It is a classical cabinet modified to create mix cabinet able to deploy 1 standard 3 cells
site and in addition up to 3 RRH modules.
All the details of 900MHz coverage impact have been analyzed in parts 6.2 to 6.3.
In the design strategy the solutions to support capacity increases should be analyzed according
to the customer
Features
RRH
900 MHz
HCS network
Eb/Io gain vs. no Tx Div speed “<10km/h” speed “~50km/h” High speed “>50km/h”
Open loop 1 dB 1 dB 1 dB
Close loop 3 dB 1 dB 0 dB
HSDSCH traffic use same Tx div scheme as associated DCH - scheduler selection impacts
performance gain as Tx div reduces power variance
Use Cases
HSDPA throughput increase for slow UE: +15% to +30% (system level simulation)
DCH capacity: +15 to 44% - STSR1+1: one carrier for HSDPA, one for DCH
The gain expected is around 10% of capacity improvement, with a main constraint of adding a
second PA. This feature is not supported yet.
In case of large area lacking capacity, a second or third carrier can be added.
The capacity with 2 carriers is expected to be a little bit less than 2 times the capacity of 1
carrier with an optimized management of the carriers.
Using 2 or 3 carriers over one MCPA may imply reduction of power available per carrier; this
method is essentially used in interference limited areas. If the power reduction per carrier has an
impact on the DL cell radius because the pilot Ec/Io or RSCP is not reached, the solution is then
to add a new PA dedicated to the new carrier, in order to keep the cell size and to improve
sector capacity.
The table below presents the capacity for different PA power and different number of carrier.
UE NF=7dB
2610,4
3183,4
3973,4
Figure 47: DL capacity obtained with different number of carriers and power configurations
Dual Cell feature allow to a UE to use two carriers in order to increase the potential peak
throughput.
The capacity gain between Dual cell vs two carriers without dual cell feature has been estimated
by simulation around 7%. The simulation conditions are detailed below:
• 10 UE per cell
• FTP with file size of 20GB (i.e. the FTP continues throughout the simulation)
• Composite channel:25% AWGN 1.5km/h, 37% Ped.A 3km/h, 13% Ped.A 30km/h, 12%
Veh.A 30km/h, 13% Ped.A 120km/h
• No UE Rx diversity
0,9
0,8
0,7
0,6
CDF
0,5
cat-14 user
0,4
cat-24 cell
0,3 cat-14 user
0,2 cat-14 cell
0,1
0
0 1000 2000 3000 4000 5000 6000 7000 8000
kbps
The comparison between RRH and Macro NodeB is done in the table below, what is the main
information to extract from this table is more the relative capacity comparison between each
solution instead of the absolute capacity value per comparison which depends one user profile
and other parameters
Without TMA
RRH 40W with 0.4dB jumper loss 660 779 1322 1434.8 1568.3 2309
RRH 60W with 0.4dB jumper loss 660 909.9 1481.7 1572.6 1698.2 2445
With TMA
Figure 48: Capacity comparison between Macro NodeB and RRH configuration
All the details of 900MHz capacity impact have been analyzed in part 6.4.2.
In part 8.7, small cell-cell layer deployment is analyzed. The gain expected, the constraints and
limitations of each strategy are detailed.
The current Node-B can support 6 sectors by adding 3 MCPA to an existing 3 sector site.
The mix cabinet (classical Node-B and RRH, see figure 34) can be used in order to increase the
number of sectors per sites. It is a classical cabinet modified to create mix cabinet able to
deploy 1 standard 3 cells site and in addition up to 3 RRH modules.
If more than 4 sectors are implemented on a same roof top, then it is advised to use 30°
horizontal beam width antenna, in order to avoid a too important overlap between sectors.
As 30° antenna gain is 21dBi, this increase the sector cell radius, and a down-tilt
optimization.
All the hardware constraints and drawbacks are the same than those described in part 7.1.7.2.
Anyway considering the heavy impact on RF optimization, site negotiation; the first solutions to
improve capacity is increasing the number of carrier, by adding if necessary another new PA
(STSR2+1) or changing the existing one to a powerful one (STSR3 with 60W PA).
Such configuration provides 47% capacity improvement, and requires less HW change and RF
activities.
The last solution to increase network capacity is to density the number of sites.
This solution is the most expensive, and the most complex, that’s why it is only used when all
others have been applied.
Traffic load
• Site selection
• Antenna tilt and azimuth optimization
• Neighbors Planning realization
• Scrambling code planning realization
The site acquisition problems must be taken into account at the beginning of the radio planning,
as the sites analyzed must correspond to those which are going to be implemented on the field.
Generally the customer should provide the site candidates positions, and the GSM existing sites
positions which generally should be re-used.
Based on this request, it is better to have two UMTS sites implemented in existing GSM sites
positions instead of having 1 new site implemented only for UMTS. Anyway sometimes, there is
no other way than implementing a new 3G site, in order to reach radio QOC and QOS, the aim is
to reduce a number lowest as possible.
Even if the sites positions are proposed by the customer, they must be analyzed and validated
by Nokia.
Radio site search and validation is an important step in the UMTS radio network planning and
implementation.
It allows identifying
If a site is good for coverage required, by detecting some masks which have not been
seen with radio planning tool.
Detection of pattern distortion risk due to the environment and the structure near the
antenna
Analysis of the antenna height compare to the nearest environment. A site must not be
too high relative to its environment, except if only coverage is the constraint, because it
increases inter-cell interferences.
All the type of equipment available on the rooftop, and their frequency band. Based on
that it can be deduced what type of antenna is required, and where to position it in order
to avoid interferences created by co-sitting.
The difficulty in finding sites is a critical issue for UMTS projects since UMTS come after GSM and
several other radio systems; the available sites are very limited in most of the cities. The
constraint on site construction is more and more important.
[R8] & [R10] linked in part 11 of this document intend to give radio engineers the guidelines
and rules on radio site engineering, search and validation.
Antenna selection and co-sitting rules are described in the following parts.
The basic rules for antenna selection are summarized in the following items:
• Frequency band
• Gain
• Polarization:
Antenna & OEM basics are detailed in the presentation referenced as [R25] in part 11 of this
document.
In the document [R24] the process to validate antenna link is detailed, even with TMA.
• XXpol antennas are already available and well known to support a technology in
1800MHz band & another one in 2100MHz band.
• XXpol antennas are already available and well known to support a technology in
900MHz band (or 850MHz band) & 1800MHz( or 1900MHz or AWS or 2100MHz )
band.
• XXXpol antennas are already available and well known to support a technology in
900MHz band (or 850MHz band), another one in 1800MHz (or 1900MHz) band &
another one 2100MHz band.
• RFS
• Kathrein
• Comba
But other providers can be used like KMW; Powerwave &Jaybeam wireless.
Antenna guidelines are respectively in [R24] and [R25] linked in part 11 of this document.
Several rules have been studied in order to be able to do co-sitting between different
technologies without having some interference problems.
A minimum isolation must be guaranteed between technologies. This isolation depends on the
technologies.
[R8] and [R30] linked in part 11 of this document; provide an analysis on antenna decoupling
solutions and GSM-UMTS co-location solution; and also UMTS vs other technologies isolation
required
This isolation can be obtained by different way; antenna sharing which provides 30dB isolation
between antenna ports,. In case it is not sufficient some filters modules are required, or a
second antenna must be implemented on the roof top with a significant spacing distance to
ensure isolation required.
The table below shows the isolation measured for different horizontal separation and vertical
separation and for different frequency bands.
Antenna Frequency Band Combination Vertical Isolation
0.5m 3m
Antennas in the same frequency bands
700MHz with 700MHz 43 51
850/900MHz with 850/900MHz 48 59
1800/1900MHz with 1800/1900MHz 58 72
AWS/2100MHz with AWS/2100MHz 60 73
2300MHz with 2300MHz 62 76
2600MHz with 2600MHz 63 78
Figure 51: Isolation provided by horizontal separation for different frequency bands
Geographical databases are not necessarily provided by the customer. This paragraph provides
some fundamental rules to select the correct database type corresponding to our need.
The database resolution and precision correspond to those expected after considering
the original data scale
The provider has re-sampled the database in order to obtain the resolution required
based on a worst database resolution.
Figure 52: Correspondence between original data scale and final resolution product
Effective Resolution
original Resolution Accuracy
delivered
Carte 1/250K DTM @ 100m Re-sample @ 50m
X,Y = 100m
X,Y = 20m
As for DTM databases, resolution is linked to original data (same correspondence table)
There are two type of clutter database: Raster and building outlines
8.3.1.2.1 Raster
Include geographic environnent distribution:
Clutter types must be chosen the most efficiently for radio propagation study.
Clutter types can be different between a country database and a city one
A minimum number of clutter types must be defined to have enough details for clutter
description
Below an example of the difference in clutter details due to the original data scale
Figure 56: Original data: Figure 55: Original data: 20-30m Figure 54: Original data: 10m
1/200k topo map satellite or 1/50k topo satellite or 1/25k topo map
A resolution higher than 30m is not accurate enough to describe a big city. Below there is an
example of clutter type definition which must be delivered by the geo database provider.
Figure 58:10m satellite imagery Figure 59: Automatic Extraction Figure 60: Manual Extraction
The data extraction method has an important impact on the clutter distribution.
The pictures above show the extraction method has a big impact on the clutter distribution, so
on the database accuracy. Automatic extraction method provides a worst result than the
manual one.
Resolution <= 5m
Height accuracy: 1m
Only 3D ray-tracing models can use the 3D vectors in order to predict small cell-cell and mini-
cell coverage.
Buildings are defined as clutter class per height range with 1m step. For forest or other
clutter the step is 5m
Two files are available, one clutter class for the clutter distribution, and another clutter
height for the height value in each bin of the area analyzed.
3D vector database file, which is necessary when using ray tracing model
These database types are very expensive and generally used by operators more than by network
vendors.
Building outlines without 3D vectors, if a small cell-cell design is not required. These
databases are described in the previous section
Or
Each raster database should have at least 12 clutter types, with at least 8 different housing
clutter types.
In this part the different model solutions for outdoor macro-cells and outdoor small cells are
mentioned. All links to useful guidelines on propagation models are detailed
The aim of a radio coverage simulation with a propagation model is to provide a prediction of
the coverage, which has to be as close as possible to reality.
In order to make reliable simulations with an accurate propagation model, the propagation
modeling is of capital importance.
For more details concerning propagation models principles overview see [R9] linked in part 11 of
this document.
In [R44] all the default ALU recommended models proposed in each technology template are
described. It is highly recommended to refer to this guideline before using them
Models are recommended depending on the database type, the frequency bands and the site or
cell surrounded configuration.
• Micro cell configuration: Antennas more than 5m below surrounded roof tops
Model to use depends on these radio configurations and not on the type of product used
CW measurements must be done very precisely, as these measurements are reference to tune a
propagation model.
A model cannot be tuned based on scanner drive tests or UE drive tests, result is never
acceptable for this kind of activity.
The measurements must be done following NEA/ RF Design team recommendation in order to
guarantee a good accuracy in the signal strength measured, and also a very precise geo
positioning of the samples on the database used, whatever the database precision is.
The essential points which should be respected are detailed in chapter 2 of the document
referenced as [R38] linked in part 11 of this document.
This model is specific to 9955 planning tool, but similar model are available on each planning
tool.
SPM can be used on all databases except 3D building vectors and all environments and the most
commonly used.
-Having a bad model calibration is generally worst that using default models
-To avoid any problem in the model tuning, ask ARFD team support by contacting Paula
Sauer or Pierre Sabatier
-This model can be automatically calibrated in 9955. It has been tested and validated in
all the databases types (see the complete calibration method and model description in
[R40] & [R44] linked in part of this document)
-The frequency domain of default SPM parameters is from 300MHz to 2GHz, but this can
be extended as SPM can be tuned on measurements
-The validity domain of such model depends on measurement range of distance, and
measurement antenna height on which it has been calibrated.
In case no tuned model is available, and there is no time to make CW measurements and to
tune one, some default SPM models are available in default UMTS templates which can be
created with 9955V730.
These default models are detailed in xls file referenced as [R44] in part 11.
The SPM default models are proposed for 2D and 2.5D geographical databases
SPM 2D macro-cells models and SPM 2.5D macro cells models must be used on cells with
same configurations than in macro cell and mini cell description in part 8.3.2.1
SPM 2D metro-cells models and SPM 2.5D metro cells models must be used on cells with
same configuration than in micro cell description in part 8.3.2.1
-These models have been tuned on CW measurements all around the world.
-They are not as accurate as a specific tuned model on a given area but ensure a limited risk
I<ERR>I <2dB & standard deviation <8dB
• these models should be used like they are set in 9955 template without adding
any clutter losses as all the attenuation is compute with diffraction. The user
must use them like they are otherwise results are not guaranteed
• SPM 2D macro and SPM 2D metro default models required to have clutter losses
updated by the user as K1 to K5 values are set.
o K1 to K7 values must not be changed the user has to set K clutter values
based on document [R44] in part 11.
The two following propagation models should be used for micro-cell or mini-cell configurations.
Micro-cell configuration corresponds to antenna height between ground level to 10m in dense
urban and urban areas.
It is not a free model by as NOKIA has a global agreement with AWE for unlimited number of
license, any NOKIA RF designer can ask for one. For a license please send an email at
rnp-tools@nokia.com
Winprop UDP uses a dominant path calculation method which is not a full ray tracing but allow
the same accuracy for a shorter computation time.
As any ray tracing model Winprop can be tuned. We recommend anticipating and planning model
calibration any project, and project funding and planning should consider it. This model has an
automatic calibration module. Winprop model calibration still quite specific and requires
experience.
• Having a bad model calibration is generally worst that using default models
• To avoid any problem in the model tuning, ask ARFD team support by contacting
Paula Sauer or Pierre Sabatier
• In case it is tuned, the validity domain of such model depends on measurement
range of distance, and measurement antenna height on which it has been
calibrated
• The Winprop default models are proposed for 3D geographical databases only
• Winprop macro-cells models must be used on cells with same configurations than in
macro cell and mini cell description in part 8.3.2.1
• Winprop metro-cells models must be used on cells with same configuration than in micro
cell description in part 8.3.2.1
Winprop macro and Winprop metro default models defined in template should be used like they
are set in 9955 template.
Winprop default models include indoor losses so in 9955 no indoor losses should be included
in clutter table or for each study “Indoor coverage” must not be selected
The user must not change some coefficient except indoor losses ( see explanation in [R45])
otherwise results are not guaranteed
To be sure that the models used in project atl are the default ones, the user should open a
default template , by clicking on File/New/From a Document Template/ UMTS then go in
Parameters/Propagation models, and compare the default model settings with the one used in
the project
Winprop default models standard deviation should be considered as 8dB with 0dB mean error
Crosswave is a propagation model developed by Orange Labs and distributed and supported By
Forsk as an option in Atoll (9955). Each ALU entity fund their own licenses and should send an
email to rnp-tools@nokia.com to ask for a Crosswave license.
It is a complete model as it can perform predictions for any cell type, recommended for macro
and metro cell type of deployment.
As Winprop, Crosswave uses dominant path method. Crosswave has much better performances
than Winprop but is much more expensive
• Having a bad model calibration is generally worst that using default models
• To avoid any problem in the model tuning, ask ARFD team support by contacting
Paula Sauer or Pierre Sabatier
• In case it is tuned, the validity domain of such model depends on measurement
range of distance, and measurement antenna height on which it has been
calibrated
As model calibration is not systematically planned, existing tuned model may not be available,
and/or there is no time to make CW measurements and to tune one, we (ARFD team) have
defined some default Crosswave models are available in V7.3.2 9955 default UMTS template.
Crosswave macro-cells models must be used on cells with same configurations than in macro
cell and mini cell description in part 8.3.2.1
Crosswave metro-cells models must be used on cells with same configuration than in micro cell
description in part 8.3.2.1, but for such metros cells configurations Crosswave default models
must have 3D databases
Crosswave macro and Crosswave metro default models defined in template should be used like
they are set in 9955 template.
Crosswave default models include indoor losses so in 9955 no indoor losses should be
included in clutter table or for each study “Indoor coverage” must not be selected
The user must not change some coefficient except indoor losses ( see explanation in [R47])
otherwise results are not guaranteed.
To be sure that the models used in project atl are the default ones, the user should open a
default template , by clicking on File/New/From a Document Template/ UMTS then go in
Parameters/Propagation models, and compare the default model settings with the one used in
the project.
Crosswave default models standard deviation should be considered as 8dB with 0dB mean error.
The goal of the uplink budget review is to validate the customer thresholds for the design and
for the acceptance. If link budget with customer assumptions is more constraining, there will
have some reservation on KPI commitments or exclusion areas. In the opposite case, a full
validation of customer thresholds is required.
Service dimensioning
In the following parts the parameters detailed are those currently used, but they can be
modified after discussion with the customer.
This chapter also describes how some parameters are taken into account in the Link Budget, and
their settings in 9955.
Tot = Outdoor
2
Set total standard deviation (calculation detailed in part 2.2) in model standard deviation
column in 9955 clutter table (see part 9.4).
Set indoor penetration recommended values (detailed in part 2.2) in 9955 clutter table
(see part 9.4).
Set Ec/Io (or Eb/No) standard deviation recommended values (calculation detailed in part
2.2) in the corresponding column in 9955 clutter table (see part 9.4).
Select Handoff supported in service table (detailed in part 9.5.3) for R’99, and HSUPA
(macro-diversity supported).
The standard values are not changed except for σOutdoor which has the value of the tuned
propagation model used for the deployment.
Tot = Outdoor
2
With
A= n1
B= .n1 + n2 (1 − ²)
C / I = Tot 2(1 − )
The range values for C / I varied between 2.5dB to 3.5dB depending on standard deviation
These values have been validated with system level simulation and field measurements
This has been confirmed by system level simulation. Here is an example of results that
shows that a 1dB standard deviation is a good approximation:
8 dB 2 dB
7 dB 1 dB
6 dB 0 dB
The penetration loss values are detailed in the tables below. Penetration losses depends on how
the equipments and antennas are implemented see the description in part 8.3.2.1
Dense Urban 18 20
Urban 15 17
Suburban 12 14
Dense Urban 13 15
Urban 12 14
Suburban 11 13
Figure 64: Penetration loss according type of radio configuration, environment type and frequency
band
* This table refers to Macro radio configuration and not necessary to equipment used. Macro
radio configuration doesn’t depends on the product but more on where is placed the equipment
antennas. Indoor losses assigned to outdoor macro layer are applied in reference of the radio
configuration description in part 9 ( antennas should be above surrounded rooftops or less than
5m lower)
Example: RF design of a small cell placed above surrounded rooftops should consider
indoor losses for outdoor macro layer
** This table refers to Metro radio configuration and not necessary to equipment used. Macro
radio configuration doesn’t depends on the product but more on where is paced the equipment
antennas. Indoor losses assigned to outdoor macro layer are applied in reference of the radio
configuration description in part 9 ( antennas should be more than 5m below surrounded
rooftops)
The penetration loss values recommended to use with 9955 are the same than those used in
the LB. considering the clutter description of the table below
Open in urban streets and open places in Dense urban /urban areas
Figure 65: Mapping between penetration losses with clutter types find in clutter geo databases
The fast fading margin models the impact of fast power control which is not any more efficient
at cell edge (the mobile reaching its maximum transmit power).
It is extracted by making the difference between the DL TX Eb/No and DL Rx Eb/No with and
without power control.
Environment parameters
Uplink load with R’99 only 65% (75%) 65%(75%) 65%(75%) 65%by default but it
(with HSUPA) can be reduced,
UL radio performances
Usually, all sites are integrated with mounted head amplifier (TMA) in order to push back uplink
coverage limitation. Node-B radio performance in uplink is mainly characterized by two
parameters, noise figure and Eb/No values or SIR performance.
Typical Node-B (or RRH) noise figure value is increase from 2.5dB without TMA to 3.3 dB with
TMA configuration.
In the configuration with TMA recommended, the total NF stays equal to BTS NF without TMA.
Node-B noise figure is impacted by TMA using .Anyway the overall calculation with Friis formula
find a global NF degradation equals to 0.2dB between with and without TMA
Nokia considers TMA annihilates the feeder loss without any additional reception gain, and the
remain loss is the jumper loss between TMA and antenna connector
To go faster with Radio Planning tool and to avoid a Friis formula calculation for each site, which
takes time when it concerns huge urban area; TMA impact can be approximated. In that case it is
considered that the UL feeder losses are reduced to 0.4dB which is the jumper loss between
TMA and antenna.
Below for each frequency band, the minimum cable length value to use MHA considering 2
jumpers (0.3dB each) required
850-900MHz 15m 18 m
2100MHz 8m 10m
2600MHz 5m 6m
Eb/No performances
Nokia Networks performances are computed for the different standard ITU channel profiles,
focusing on Vehicular A (VehA), Pedestrian A (PedA) and Pedestrian B (PedB).
PedA is a multi-path channel with a dominant path when PedB is composed of two main paths.
For macro cell application, according to measurements in typical European city, the profile PedA
is statistically observed 45% of the time while the profile PedB represents 55% of the samples.
Looking at Eb/No performances, this mixture of multi-path channel profile is well represented by
the performances of the VehA multi-path performances.
All the Eb/No values are detailed in the Link Budget tool, or in the RNP tool. The values come
from R&D are internal only, and should not be changed
UMTS downlink budgets are always built the same way. In fact there are two downlink budgets,
first one deals with pilot dimensioning and second one with service dimensioning.
The pilot dimensioning phase allows as well fixing common channels contribution as their power
is relative to pilot one.
Global recommendations
Basically there is no maximum power limitation, when there is one MCPA per sector, it means
that max power setting can be set to any value, generally it is recommended to set it to max PA
power
If there are several MCPA per sector, each MCPA must have the same max power setting.
Special case: several MCPA per sector with different number of carriers (STSR2+1)
To obtain that power per carrier is defined based on carrier power available with MCPA where
the highest number of carrier is supported.
Then MCPA max power which supports the lower number of carrier must be limit to
Example:
For STSR2+1 configuration, with 2 45W MCPA; the power available in STSR2 configuration,
considering an equal power distribution, is 22.5 W per carrier.
Then on second MCPA which support only one carrier; max power must be limited to 22,5W.
Remark:
It possible to increase CPICH power & max Tx Power of the 3rd carrier but the interest is very
limited
No gain for HSUPA carrier (FDD3) as the limitation comes from UL load
Gain only in areas with DL limitation and low interferences => Deep Indoor
Possible to favor HSDPA carrier (FDD2) : improve Indoor coverage and PA power
for HSDPA capacity
The aim is to determine the minimum CPICH power ratio to insure a certain pilot quality at cell
edge. The max allowable path loss is fixed by the uplink budget and usually based on the
dimensioning service. Pilot quality is estimated by the ratio of energy chip over interference
Ec/Io.
The recommended CPICH power ratio is 10%, this value has been deduced considering
Under such value ALU considers pilot good detection by the mobile can be problematic.
Extra-cell interference has been evaluated to DL Ie/Ii= 200% compared to intra-cell, considering
• two neighboring macro-cells received with the same RSCP than the serving cell
• pilot ratio is the same for the serving cell and the two neighbors
• serving cells and the two neighbors have the same traffic load
Mobile noise figure and antenna gain are used to determine the noise floor
In an UMTS network where TMA are implemented, the variation of the cable losses between
neighbor sectors has a tiny influence on UL coverage whereas the DL coverage is strongly
dependent of cable losses.
UL/DL Unbalanced
The aim is to have the same equivalence path loss point in UL and DL between two cells. This
requirement is important for HSDPA as macro-diversity is not available. The aim is to avoid to
receive cell A in DL and to be received by cell B in UL, because in that case there is an HSDPA call
drop, instead of having a hard handover.
UL analysis:
This requirement is resolved, between Node-B with quite the same cable length, by setting the
same CPICH power value at the antenna connector for all the cells.
The CPICH must be tuned differently for all the neighboring relations like:
Between two 45W Node-B if the cable length are very different (20m difference) without
TMA
Between two Node-B, one without TMA, and the other with TMA
The table below contains the common channel power settings recommended.
As mentioned in the table above the recommended % of CPICH power ratio is between 8% to
12%;
• 8% for Dense urban, urban and some suburban areas which are interference limited
For instance with 5% CPICH power ratio the Ec /Io target value for 80% network traffic load
should be -18dB instead of -15dB with fALU recommended power settings.
SHO margin is considered equal to 15% of the traffic power, which represents around 10% of
the total power. Based on this the max downlink power allowed including Common channels is
90% of the PA power.
MS noise figure 8 dB 9 dB
The orthogonality factor values have been deduced considering PedA profile 45% of time and
PedB profile 55% of time as it has been statistically evaluated on the field.
The DL shift values represent the delta dB in path loss due to UL and DL frequency differences.
The values presented below come from Cost-Hata model.
DL/UL conversion factor
UL user bit rate 12.2 kbits/s 64 kbits/s 64 kbits/s 128 kbits/s 384 kbits/s
2100MHz
CS 64 5 5.4 5.5
1850MHz
CS 64 5 5.3 5.35
900MHz
CS 64 4.95 5 4.95
The process of the design is based on the coverage requirements of the worst link budget in the
uplink (or the dimensioning service negotiated with the customer).
Dimensioning with this service allows having an enough path loss to support
DL PS128 service at cell edge for 90% traffic load with 95% area probability
DL PS384 service at cell edge for 50% traffic load with 95% area probability
The available UL path loss has been calculated in the LB with the parameter settings validated
with the customer.
The traffic analysis is the second stage of the simulation. The traffic load should have been
defined or validated by the customer.
In the following are presented in detail the service and traffic information necessary to complete
an UMTS design. Traffic assumption has a large impact on the cell count and radio planning
results. They should thus be detailed for each of the required services.
• A % of traffic load or total power used is specified between 75% to 90% of traffic
load including common channels and shared power are usually considered
according to the customer this value can change. This is taken into account in
9955, by setting in Cells table, the Total Power column to power values in dBm
corresponding to the % of total PA power
• A mixed traffic is provided by the customer. This mixed traffic is based on a user
call profile which contains each service activity and their duration or data volume
transmitted during a busy hour. First it is necessary to determine exhaustively all
the services the network will have to support. A service is the conjunction of a
data type (voice or data), a transport mode (circuit switched or packet switched),
and a data rate.
The area where each service has to be provided must be clearly defined (like high data rate is
required in urban center.)
These areas are generally defined per environment type, but traffic can also be defined per
service and per user type, and per sector.
The user type (indoor, outdoor, in car 50km/h, 120km/h) per service, per area is necessary as
well.
Indeed the required Eb/No for a given service in a given environment may be different according
to the profile of the user.
UMTS simulations must be run, in order to take into account the right traffic in 9955
Based on dimensioning service and traffic assumptions, the RF design can be completed and
optimized.
The first step is then to select the correct candidate sites in order to fulfill the required QoC and
QoS per area.
During the design, it is important to keep in mind that a W-CDMA network is an interference
limited system, and that in such cases, the sites should be positioned so as to limit inter site
interference.
Therefore, high sites or sites on top of hills or mountains should be avoided (in very hilly
environment, some bi-sector sites could be chosen) the sites should be regularly positioned
over homogeneous areas, electrical down tilt based on the site height and the cell size should be
used…
The RF performances are analyzed with 9955 using the predictions templates and UMTS
simulation module.
The ways how to realize, in 9955, the predictions and simulations to complete the following
design targets are detailed in parts 9.9 & 9.10
It depends on the dimensioning service selected (see part 8.4). The % of QoC is 95%, except for
rural areas where it is 90%.
It means that, when shadowing margin is taken into account, for a 100% covered area obtained
in prediction with radio planning tool for RSCP>=RSCP target criteria, 95% of this area has
RSCP>=RSCP target criteria on the field.
With 9955 there is no solution to take into account the SHO gain for coverage signal analysis.
So only for this study, the easiest way to realize a prediction is to use LB available path loss
without including penetration loss
Set the standard deviation of the propagation model used (If it is the ALU default model, then
the STD is 8dB). Here the example is for a 7dB tuned model
To use this RSCP target with 9955 penetration losses should be set in 9955 clutter table (see
part 9.4), in line with the values detailed in part 8.3.3
Select indoor loss in 9955 coverage prediction study (see part 9.10.1)
Do not select “Shadowing taken into account” in 9955 coverage prediction study
Acceptance level is the outdoor RSCP target value to at least measured or above over 95% of
the area to be sure that UL service at cell edge targeted will be ensured in indoor over 95% of
the zone.
The Ec/Io target value depends on the traffic load. The values detailed below considers CPICH
power ratio between 8% to 12%; if CPICH ratio is out of this range then Ec/Io target should be
modified in consequence (5% CPICH ratio induce a 3dB decrease on the target values detailed
below)
• 95% of the design area should have a Ec/Io value higher than -15dB, with 85% DL
load
• 95% of the design area should have a Ec/Io value higher than-14dB, with 75% DL
load
• 95% of the design area should have a Ec/Io value higher than -12dB, with 50% DL
load
• 95% of the design area should have a Ec/Io value higher than -9dB, with 25% DL load
Such Ec/Io criteria allow managing interferences in order to obtain a RF network design able to
support
• PS384 service at cell edge when surrounded cells have 50% traffic load or
PS128 at cell edge when surrounded cells have 80% traffic load.
• 400kbps HSDPA throughput at cell edge when surrounded cells have 75%
traffic load
Scrambling code planning must be done, using the automatic code planning in 9955 (see part
9.8).
Coverage target for each service must complete 95% of the design area. A service is considered
to be supported when UL and DL Eb/No are reached. As we have seen Eb/No values take into
account all the profile characteristics, speed, orthogonality factor…
The aim is to optimize tilt in order to reduce the number of servers covering the same area,
thereby minimizing interference without impacting coverage, quality and at the same time
enhancing the SINR
The optimization method to reach these targets is detailed in the part below.
This analysis should be done after tilt and azimuth optimization so after several ACP iterations
analysis to reach the coverage quality and overlapping targets, and before neighbors list and cell
ID attribution
This analysis consists in verifying the several following aspects over the area which has just been
optimized
• Ensure that best server area is continuous and there is no interruption in best
serving zone by a neighbor cell which would have higher signal strength.
• Verify that all targets are respected and try to improve them manually on the
areas where it is not the case
• Check that the sites the most far away which create some interference have the
maximum down-tilt value applied.
o If it is not the case on the site has a bad impact on Ec/Io, overlapping, or
Eb/Nt, increase its down-tilt by 2° steps and see the impact.
• Verify that the solution propose is applicable; the highest sites must be the most
down-tilted ones compare to low height sites.
8.6 OPTIMIZATION
8.6.1 METHODOLOGY
The methodology consists in analyzing the area optimized; verify the interference and quality
targets, neighbor plan and scrambling code plan.
The scrambling code plan and polluters reduction should take care about the following points
• The SC of the optimized area re-used in the 2nd external site ring or further are
not necessarily the last polluters in term of ranking (4th, 5th … servers).
• If all polluters are belonging to the first ring, it can be due to a design issue
(missing site, non homogeneous design, high sites) or an environment issue
(Open in urban, hill, mask). In this case, the goal is to create dominance from one
or two sectors increasing their influence and decreasing other influences. Most of
the time tilt manipulations are not sufficient and azimuth changes are required.
• There are also some fine tuning changes that consist in reducing the sector BTS
side while keeping the coverage in the main direction. It’s realized by electrical
down-tilts associated with a mechanical up-tilt.
These methods are identical in all radio system in order to limit inter cell interference, or to
improve the QOC.
In case of pilot unavailability pilot power may also be tuned, in order to provide a dominant
server. The pilot signal should be decoded in every location, as it is required to enter a cell, to
start a communication, and for handoff. When the Ec/Io is insufficient then the interference of
the neighbors should be minimized, and/or the best server received energy increased.
Such optimizations can be done manually, for a low number of sites but the optimization can be
done by an Automatic Cell Planning tool (ACP) allowing a treatment of a higher number of sites
in a short time delay.
This tool extracts/uses from the RF planning tool, all the information necessary to the
optimizations:
Path loss
Traffic assumptions
Mobile types
Node-B Characteristics
Antennas
…..
Based on these information, and with some KPI targets values like
RSCP level,
QOS target
Ec/Io target,
GOS target,
Pollution window,
The ACP is able to optimize the Network, by changing the different parameters the user allowed
(CPICH power, mechanical tilt, electrical tilt, azimuth, antenna type).
The user is able to define a % for each KPI, and to weight them.
The network optimization stage is done according to the simulation results in order to reach the
target quality thresholds, and to optimize the final radio planning results.
The optimization of the network radio planning is based on the results of the analysis.
Depending on the blocking level, or coverage holes location and size, different solutions can be
used.
The main part of tilt, azimuth, CPICH power adjustments is done in this phase.
CPICH power must be configured without using ACP in order to have the same value at the
antenna port.
Examples:
60°inter-sector angle between two 60° H beam width antennas, or one 60°H beam width
antenna and one 30°H beam width
Or
90° inter-sector angle between two 90° H beam width antennas, or one 60°H beam
width antenna and one 90°H beam width
o Identify the cluster to optimize, the number of sites for this cluster should not
exceed 500 sites.
• This can be done with RNP tool and ACP tool using focus zone
which includes sites to optimize; and computation whose sites
o Then when everything has been verified and is ok , neighbors and Scrambling
Code allocation can be done
At the end of this phase a neighboring plan is automatically generated with 9955. This
neighboring plan is implemented on the field.
The main planning rule is to re-use cell ID with a sufficient distance in order to avoid radio
interferences.
This distance depends on propagation path loss, the environment and the frequency.
A correct RF design and optimization allow having an average number of neighbors per cell
between 16 and 22 .
Considering two cells cell A and cell B, on the same frequency carrier using the same scrambling
code, the distance between those must satisfy the following criteria:
To complete radio constraint, a distance criterion can be added but it can not be used alone as
it assumes a homogeneous environment and a flat terrain area
The distance detailed below represents the minimum distance to consider a cell not as a
neighbor, or to consider that scrambling code can be re-used
Considering mean cell radius around 700m @ 1800MHz and 2100MHz, the minimum
inter-site distance is 2,4km
Considering mean cell radius around 1,5km @ 850MHz and 900MHz, the minimum inter-
site distance is 5,5km
Suburban
Considering mean cell radius around 2,2km @ 1800MHz and 2100MHz, the minimum
inter-site distance is: 7km
Considering mean cell radius around 5,5km @ 850MHz and 900MHz, the minimum inter-
site distance is 18km
Rural
Considering mean cell radius around 7km @ 1800MHz and 2100MHz, the minimum
inter-site distance is: 21km
Considering mean cell radius around 18km @ 850MHz and 900MHz, the minimum inter-
site distance is: 53km
Scrambling code information is provided by SCH broadcast physical channel which contains P-
SCH and S-SCH. The cell search process is the following one:
• Third step the scrambling code is found by correlating symbol by symbol over P-
CPICH with the codes in the group found in step2.
A more complete analysis and explanation is done in the [R17] linked in part 11 of this
document
Planning strategy:
Two strategies are possible. The first one is to use a low number of groups with a maximum
number of codes on each group; the second one is to use a high number of groups with a low
number of codes per group.
The table below summarizes the advantages and drawbacks of each strategy.
To improve synchronisation time vs. UE processing load since bad synchronisation will finally
impact the network performance and the quality of service
On one site, it is possible to apply an offset using tcell parameter setting in order to
desynchronize the SCH channels. Default parameters are 0, 3, and 6 for the 3 cells. Moreover
the SC group is identical over the 3 cells.
The following recommendation has been extracted from the presentation [R18] linked in part
11 of this document.
Recommendation:
To keep different tcell values in order to better distribute the interference generated by the
SCH channels over the time.
When all the azimuth, electrical and mechanical tilts have been optimized during the pre-
optimization, then the next step is the field implementation.
WARNING! :Particularly attention should be done for azimuth and mechanical tilt settings
• Azimuth settings must be also considered with a particular attention and we must
take into account that Radio Planning tool recommendations are geographical azimuth
and on the field we measure magnetic azimuth.
o The method to evaluate the difference and define the magnetic azimuth
corresponding to the recommended geographic ones is described in document
linked as [R48] in part 11.
• Mechanical tilt as it is also manual operation and requires measurements to ensure field
match recommendations
This part describes the field measurement process used to finalize and validate the RF
parameters in order to obtain the best network performances.
RF measurements must be done with scanners and reception antenna positioned on car
rooftop
WARNING!
In the document [R32] some indoor measurements have been performed on the same roads, to
compare values measured with calibrated scanner vs with UE considering 10 different UE. The
average UE error is around 13dB. This should not be considered as a stable value as depending
on UE the error variation is between 7dB to 22dB lower than the scanner. Such results can be
extrapolated for outdoor scanner and UE measured same macro along the same roads
All the runs done for RF design control and validation must be done with scanner (JDSU scanner)
and rooftop antenna as UE cannot be reliable for absolute measured values.
The tilt and azimuth can be changed only after the analysis of these runs. No tilt and azimuth
can be updated based on a UE drive test where the load is not fixed and constant
Any UE measurements are not reliable for RF design since, their measurements accuracy is very
poor as mentioned above and most of them don’t have the possibility to use external antennas.
Test mobiles, are only used for the other optimization steps!
Post-treatment process
• The complete description of scanner using and measurement process is done in the
document [R11] of part 11 of this guideline.
• Post treatment can be done using EDAT software (internal fALU) or other software
like Gladiator (Agilent). An internal macro with associated guidelines has been
developed for the scanner measurement post-treatment. It is available for JDSU &
PCTEL scanner, as Edat may face issues to extract some RF design criterion.
• Internal macro with associated guidelines is available in link [R39]. All measurement
process is detailed in [R11]
• Also a detailed description and presentation of equipments Nixt and MT8222 is done
on the referenced documents [R49]& [R50] in part 11.
• Done just after the implementation of the pre-optimization results (see part 8.6.2).
• Fixed load must be applied with OCNS on the entire zone which should be validated. The
recommended OCNS value are 50% or 100%
• This has been tested and validate for trials and POC over more
than 150 metros.
• Overlapping criteria
• Ec/Io
• The targeted values detailed in parts 8.5.2 must be ensured over more
than 95% of the area.
• With drives test measurements, the scanner measured all the servers in 20dB
range. The analysis of such data provides the right information to see if
Neighboring and Cell ID plans done with the RNP are correct or if some cells must
be added
• Post processing will mainly be on radio conditions to check which change should be
applied to optimize network
• For the areas where the targets are not ensured this can be improved by decreasing
interferes impact with down-tilt optimization. This can also be done by increasing best
server coverage area
• In the list of the significant interferers, the one which should be tilted in priority is the
further one not necessarily the one with the lowest level
• The best tilt change configuration suggested must be the one which respects all targets
criteria and/or provides the best improvement for all over the area analyzed.
2nd Run: RF design control fine tuning + neighbour list control tuning,
• Done just after the implementation of change derived from 1st Run, to validate change
efficiency
• Fixed load must be applied with OCNS on the entire zone which should be validated. The
recommended OCNS values are 50% or 100%, and obviously the same load should be
used for run 1 and 2.
• Azimuth
A 3rd run may be necessary to validate the RF design in case new changes have been done after
the 2nd run. This 3rd run is done in the same conditions than the two previous one, fixed load
scanner.
• UTRAN parameters settings are detailed in [R12] for LR15.x and in [R13] for
LR14.x linked in part 11 of this document
• HSxPA parameters settings are detailed in [R37] for UA8.0 linked in part 4 of this
document
Small cells deployment allows offloading macro layer and support traffic & network capacity
increase.
Small cells can be deployed in shared channel conditions with macro-layer, or in dedicated
channel. .
All the capacity gain analysis and constraints for a small cells deployment in shared or dedicated
channel are detailed in 3G small cells guideline linked as [R41] in part 11.
For outdoor small cells RF design cell radius vs macro signal strength can be extrapolated from
the tables detailed in document [R51] or in guideline linked as [R41] in part 11.
Outdoor Small cells Link Budgets can provide a first view of small cells footprint vs macro signal
strength . It is based on the tables detailed in [R51]. For more information concerning indoor
and outdoor small cells link budgets using, check the doc following documents whose link are in
in part 11:
To define the macro signal strength on a given position there are two options.
• RNP analysis
▪ Since the user has all the information we recommend to perform the
complete footprint analysis as detailed in the points below
o With RNP the user can also analyze any candidates positions and check the
small cells radius obtained using coverage by transmitter table in 9955. For
more information see all 9955 settings in part 8.7
▪ The minimum recommended cell radius value for metro which cover a
road is to have 50m cell radius along this road to allow the incar UE
staying a minimum of time on the cell.
▪ To check quickly if a small cell is above after the analysis with coverage
by transmitter, the user has the possibility to define a grid per site see
part 9.2.2.
▪ A new module has been created to evaluate on automatic way the cells
radius achievable for each candidate position for each azimuth range.
This module description and settings are available in the link [R55] in
part 11
o For RNP analysis the user should take care about the propagation models
used for macros and metros, see the documents [R44] and [R9] linked in part
11.
• Scanner measurements
All interference RF design targets detailed in part 8.5 must be followed and respected.
In case where RNP analysis is possible interference criteria can be checked and RF design
updated to match them in a pre-optimization phase for each candidates position where cell
radius is acceptable. Then after implementation the verification is performed on the field, during
RF design field validation phase see part 8.6.3.2
All the analysis and constraints for a small cells deployment in shared or dedicated channel are
detailed in UMTS small cells guideline linked as [R41] in part 11.
All the documents detailing tools using method for small cells placement and optimization
[R52], [R53] and [R56] referenced in part 11
Some documents detailing offload analysis area referenced as [R57] in part 11.
ATD module has been developed in order to create a focus zone an traffic map based on geo-
location.
Based on PCDM data, such module allows creating traffic map inside 9955. Then based on the
traffic map created ASP/ACP can load balanced the traffic by optimizing tilts and azimuth
depending on the settings.
With ATD module a traffic map can also be created based on data volume transferred per cell
and demographic map.
Such optimization allows a better traffic distribution, users will be scheduled more times so
higher user throughput is expected, and better network efficiency.
All the process and method is detailed in the document linked as [R56] in part 11.
A new module named ATD; has been developed in order to create a focus zone an traffic map
based on geo-location.
ATD internal module has been developed inside 9955 in order to create a focus zone an traffic
map based on geo-location.
Based on PCDM data, such module allows creating traffic map inside 9955. Then based on the
traffic map created ASP/ACP can place and optimize small cells in the areas where there is the
highest traffic load. Depending on the settings macro-cells can have their electrical tilts
optimized too.
Such analysis allows optimum small cell placement; traffic offload from macro layer to small cell
layer distribution, users will be scheduled more times so higher user throughput is expected,
and better network efficiency
All the process and method is detailed in the document linked as [R56] in part 11.
For RF design done with macro-cells, principles, targets and rules are the same than those
detailed in previous parts. The main constraint of such deployment is capacity and users/km²
density to support.
To better manage interference small cells can be a better approach vs. macro-cells
All recommendations; constraints and RF design rules for small cells deployment in stadium, in
shared or dedicated channel with macro-layer, are detailed in document [R41] in part 11.
For more details and information concerning 9955, see [R21]and [R22] linked in part 11 of this
document.
A 9955 default template is available when the user clicks on File/New/ ‘From a Document
Template’/’UMTS ALU’; all the recommended settings have been defined including all the
bearers, equipment settings, including Quality tables, MUG & MIMO tables, and Bearer selection
thresholds.
Since release V7.1.0 , 9955 has its interface completely updated. Now, in V7.3.2, as it is shown in
the picture below , there are 4 different tab.
• Parameters:
o Traffic parameters: contains Services, Mobility Type, Terminals, User profiles and
Environment folders
o On-line maps
• Network:
o Drive test Data: updated contains and possible settings as in previous 9955
release
• Site Explorer:
Do not set the altitude; it is done automatically with the DTM database
Columns for Pylon height or Support types are just information and not used for any calculation
Parameter Unit Value Remarks
Name Set site name
X Set longitude
Y Set latitude
Altitude Automatically set when there is a DTM database
Max nb of CE UL See document [R28]
Max nb of CE DL See document [R28]
Max Iub UL Backhaul 12288
Throughput (kbps)
Max Iub DL Backhaul 12288
Throughput (kbps)
Pylon Height (m)
Equipment Macro Used to indicate the site equipment type
Figure 79: Sites table parameters description
This is used more for small cells, to quickly check if their footprint is sufficient after running
coverage by transmitter analysis ( see part 9.10.1.2)
In Network right click on Site, select Grid the following window appears
To visualize if the small cell footprint respect a critical threshold, 50m in vehicular traffic area
recommended as a minimum, set 50m as Maximum cell radius , and 20 m for circle spacing.
Azimuth degrees
Additional Electrical
degrees Calculated with 3D interpolation
Downtilt
Inter-carrier power If selected the power between two carrier can be shared in order
sharing to support an optimum number of users in the cell
Number of Reception
if set to 1 then reception diversity is not applied
Antenna ports
BTS Noise Figure dB 2.5 set the typical value according to the chosen equipment
Receiver antenna 0
dB Already included in the Eb/No values
diversity gain
Transmitter
Indicates the transmitter’s type of equipment
equipment
Possibility to select a TMA type if it has been specified in the
directory by Right click on Transmitters\equipment\TMA
TMA equipment
Warning: if TMA impact has already been considered in Reception
& transmission loss column, nothing must be set here
Possibility to select a feeder type if it has been specified in the
Feeder equipment
directory by Right click on Transmitters\equipment\Feeder
Defined the cable length, only used if no value has been
Transmission feeder
meter specified in transmission loss, and if a feeder equipment has
length
been specified
Defined the cable length, only used if no value has been
Reception feeder
meter specified in reception loss, and if a feeder equipment has been
length
specified
Use the tuned model established for the area where the Rf
Main propagation design is done, and if none is available the use the default
model models detailed in document
Main Calculation Should be equal to the geo database resolution, or to the half of
meter
Radius the database resolution
Number of
1,2,4; Set if it is set to 1 then TX diversity can not be applied
Transmission Antenna 1
Ports
Total PA power:
3 possibilities:
dB or
SCH power Pilot PA power -7dB
dBm
dB or
Other CCH power Pilot PA power+0.6 dB
dBm
Power used by the cell for all common channels and traffic.
It represents the DL load of the cell taken into account during the
prediction.
Total Power dBm For a design it can be set to 50%, 75%, 90% of the total PA
power, this impacts the Ec/Io target (see part 8.5.2).
If the user does not set any value, the default value is 50.
Scrambling Code Define the SC domain corresponding to the cell. SC domain are
domain defined
Primary Scrambling
Cell SC defined using automatic allocation
Code
Scrambling Code
domain
HSDPA Dynamic Power Allows automatic remain HSDPA power calculation when a UMTS
selected
Allocation simulation is launch. Must be selected
Max number of HSDPA Input used in simulation to have a limitation in the number of
15
users users per cell
Number of HSDPA Output from simulation to indicate the number of users for each
users cell
Max number of HS- Limit the maximum number of PDSCH codes allocated to a HSDPA
15
PDSCH codes user
Max number of HSUPA Input used in simulation to have a limitation in the number of
15
users users per cell
Number of HSUPA Output from simulation to indicate the number of users for each
users cell
DL Max Throughput
kbps 1000 Input for simulation
per user
Number of HSUPA Output from simulation to indicate the number of users for each
users cell
Warning!!! to ensure 3dB noise rise 50% must be set even if 50%
is set in the column UL load column dedicated to R’99
HSDPA Scheduler Proportional HSDPA Scheduler type from the Schedulers table. Select
Algorithm Fair ‘Proportional Fair’ to be in-line with ALU product.
MIMO Support None Allows to select: None, Transmit Diversity, Spatial Multiplexing
Inter-technology DL
dB 0 Indicates the level of inter-technology interference in DL
noise rise
Inter-technology UL
dB 0 Indicates the level of inter-technology interference in UL
noise rise
A new cell is automatically created at an adjacent frequency of the cell on the same sector
This table contains all the information concerning the clutter type characteristics.
Model standard dB
8 Set total standard deviation ( calculation detailed in part 8.3)
deviation
Ec/Io standard dB
deviation 2 Set Ec/Io (or Eb/No) standard deviation recommended values
(calculation detailed in part 8.5)
Eb/Nt UL standard dB
2
deviation
Eb/Nt DL standard dB
2
deviation
Indoor loss Set indoor penetration recommended values (detailed in part 8.3) in
9955 clutter table.
Or set the values provided by the customer
Orthogonality factor Set the values detailed in part 8.3 of this document
Spatial Multiplexing linear 1 Linear gain applied to the throughput when MIMO is supported
Gain Factor
Additional Diversity dB 0
gain (DL)
Pilot finger % 100 Used to evaluated an orthogonality factor impact on the pilot
Orthogonality factor Set the values detailed in part 8.3 of this document
Losses dB 0
Noise figure ( main dB 8 The terminal noise figure for the frequency band identified
frequency band) as the main one. Depends on customer inputs otherwise use
default values
Noise figure ( second dB 8 The terminal noise figure for the frequency band identified
frequency band) as the second one. Depends on customer inputs otherwise
use default values
Noise figure ( third dB 8 The terminal noise figure for the frequency band identified
frequency band) as the third supported one. Depends on customer inputs
otherwise use default values
Rake factor 1 Enables to model rake receiver
In Parameters tab / Traffic Parameters double click on the folder Mobility Types
As specified previously, CPICH Ec/Io target is set to -15dB, the new parameter is HS_SCCH Ec/Nt
target which is used by 9955 to define the power used for the HS_SCCH based on the radio
conditions. This power is deduced from the HSDPA power specified in the Transmitters/Cells
table.
The HS_SCCH Ec/Nt target value recommended is -13dB based on the calculation method.
A service can be created and saved without any R99 radio bearer assigned, but for any
simulation a R99 radio bearer must be assigned to a service.
Parameter Unit Value Remarks
Type of service: Circuit(R99), Packet(R99),
Packet(HSDPA-Best effort), Packet(HSPA-Best
Type Effort), Packet(HSPA-Constant bit rate),
Packet(HSDPA-Variable bit rate), Packet(HSPA-
Variable bit rate),
Carrier Support -1 all the carriers, or selection between the
In this part several tables are described, all the parameters are not detailed, but only the most
useful and significant for RF design.
For more details and information concerning 9955, see [R21] & [R22] linked in part 11 of this
document
These parameters are already set in the default template; click on File/New/ From a Document
Template/ UMTS; all the recommended settings have been defined including all the bearers,
equipment settings, including Quality tables, MIMO tables, and Bearer selection threshold.
For R’99 only, Nokia recommends calculation of the pilot Ec/Io with the
Io Total
signal pilot considered.
Without useful For R’99 only, Nokia recommends calculation Eb/Nt without the useful
Nt
signal signal.
RSCP Activation
dBm -100dBm RSCP signal threshold for compressed mode start
threshold
Eb/Nt UL & DL
dB Eb/Nt increase due to compressed mode introduction.
Target increase
Without useful For HSDPA only, Nokia recommends calculation Eb/Nt without the
Nt
signal useful signal.
As UE ACLR and ACS values are worst than Node-B ones, and equal to 33dB; the recommended
value to set in 9955 is 33dB.
In Parameters tab/Radio Network Settings/ Radio Bearers double click on R99 Bearers
Nothing except
interactive for ADPCH For info only it corresponds to the traffic class the bearer
Type
used for R’99 RAB can fulfill
assign for HSDPA
DL coding factor 1
UL DPCCH/DPCH Used for simulation and prediction in order to know the part
0.125
power ratio of power used for Control channel
DL DPCCH/DPCH
0.179
power ratio
Define the max power assign to one user. the user updates
Max TCH power dBm Depends on service this value if necessary in order to reduce the user rejection
% during an UMTS simulation
DL spreading factor
Depends on service Possibility to define the spreading factor for active users
(Active users)
DL spreading factor
Depends on service Possibility to define the spreading factor for inactive users
(Inactive users)
In Parameters tab/Network Settings/ Radio Bearers double click on HSDPA Radio Bearer
The Max RLC Throughput values don’t take into account BLER.
The BLER is taken into account in prediction and simulation if the “application throughput” is
considered, instead of “peak rate” where BLER is not taken into account.
In Parameters tab/Radio Network Settings/ Radio Bearers double click on HSUPA Radio Bearer
The Max RLC Throughput values don’t take into account BLER.
The BLER is taken into account in prediction and simulation if the “Effective throughput” or
“Application Throughput” are considered, instead of “peak rate” where BLER is not taken into
account.
In Parameters tab /Radio Network Settings/ Radio Resource Management double click on Sites
Equipment
Rake factor 1
DL Overhead
resources for common
channels/cell
UL Overhead
resources for common
channels/cell
Compressed mode Not selected Used for hard handover from 2G to 3G.
Overhead Iub
Throughput /cell
Throughput supported Not selected Used for hard handover from 2G to 3G.
per E1/T1 Ethernet
link (kbps)
Multi-cell HSDPA Proportional HSDPA scheduling algorithm is selected when in multi-cell mode of
scheduling algorithm Fair operation
In Parameters tab/Radio Network Settings/ Radio Resource Management double click on R99
Consumption Resource
In Parameters tab /Radio Network Settings/ Radio Resource Management double click on HSUPA
Consumption Resource
In Parameters tab /Radio Network Settings double click on Reception equipment, the table
below appears
• MIMO
They should not be changed without RAN Expertise support as linked to product characteristics
Setting recommended
Carriers: set the carrier corresponding, if several carriers should be selected, set them
like follow: 1; 2; 3
Suburban
Rural
Force co-site as neighbours: force in the candidate neighbour list cells located on the
same reference cell site. Selected
Force adjacent cells as neighbours: force in the candidate neighbour list cells
geographically adjacent to the reference cell. Selected
Force exceptional pairs: force forbid some neighborhood relationship defined by the
user. No specific recommendation
Reset neighbours: if selected all the existing neighbours are deleted before computation.
No specific recommendation
Ec/Io -15dB
Ec/Io margin equals to 10 dB, margin should be between 10dB and 15dB
Then assign the different groups as follow. In the picture below each scare is a minimum of 100
sites cluster.
Constraints recommended
Suburban
Rural
Strategy
A cluster is a number of consecutive codes. This number is set in “Nb of codes per
cluster”.
Distributed per cell: allocate a maximum number of cluster. 9955 can allocate codes
from different clusters.
One cluster par site: 1 cluster per site, and 1 code of cluster par cell. If there aren’t
enough clusters then they are reused without using “Default reuse distance” but trying
to assign clusters space out as possible.
Distributed per site: 3 adjacent clusters per site, 1 cluster per transmitter, 1 code per
cell. If there aren’t enough clusters then they are reused without using “Default reuse
distance” but trying to assign clusters space out as possible.
9.9 SIMULATIONS
For information concerning monte carlo simulation settings and method , see the
document [R22] in part 11 of this guideline.
The coverage study has a separate setting in 9955, as it is not possible to take into account
easily the SHO gain for coverage if the shadowing margin is calculated directly with 9955.
On the other side, if the RSCP target are only based on LB, as there one RSCP target per
environment it is not easy to set it in 9955 study to have a clear coverage map in one shot.
Do not select “shadowing taken into account “icon as it is already done in RSCP target calculated
below (see figure below)
Coverage by transmitter is a useful study to verify if there are some overshoots and also check
if small cells footprint is sufficient particularly above 50m in vehicular traffic area
Right click on Predictions folder in 9955 then go in Standard Predictions and select “Coverage by
transmitter”.
Verify that the values set in clutter table is in line ( +/-1) with the default value set in 9.4.
indicated values
In the box “with a margin”, set the different values (4dB and 10dB).
Minimum signal level should be set: it is Pilot signal level which is considered
9.10.4 RSSI
Set total standard deviation value (calculation detailed in part 8.3) in model standard
deviation column in 9955 clutter table (see part 9.4).
Set indoor penetration recommended values (detailed in part 8.3) in 9955 clutter table
(see part 9.4).
Set Ec/Io (or Eb/No) standard deviation recommended values (calculation detailed in part
8.5) in the corresponding column in 9955 clutter table (see part 9.4).
Select Handoff supported in service table (detailed in part 9.5.3) for R’99, and HSUPA
(macro-diversity supported)
Considering the % of area reliability target and the equivalence in cell edge reliability detailed in
part 8.4, set the following % in front of “Cell Edge Coverage Probability”.
Figure 117: % of cell edge probability for each environment to set in 9955
Select “indoor coverage” checkbox in 9955 coverage study (see figure above)
For each study, if no simulation has been done, the load taken into account is the lowest
between:
Total power value and max DL load (%) set in Cells table
UL load factor (%) and max UL load (%) set in Cells table
In case of no value set in Total power and UL load, the default value taken into account is 50%
Average RLC throughput is RLC Peak throughput taking into account BLER:
( Peak .rate.(1 − BLER )) / TTi
Application throughput use the throughput scale factor and throughput offset (see [R22]) in
order to calculate throughput for higher protocol level than RLC.
Prerequisites:
- ‘Multi-cell mode’ parameter in the Transmitters table allows ‘DL multi-cell’ operation
mode, see part 9.3
- Throughput prediction is run for a terminal type which is with a configured HSDPA UE
Category supporting one of the multi-cell modes of Dual-cell, 3C (3-cell), 4C (4-cell), 6C
(6-cell) or 8C (8-cell), see part 9.6.7
- There are defined in the main operating frequency band at least as many active HSDPA
cells as the number of cells supported by the terminal
Prerequisites:
- ‘Multi-cell mode’ parameter in the Transmitters table allows ‘DL multi-cell’ operation
mode, , see part 9.3
- Terminal is configured with a HSPA support mode of either DB‐HSDPA or DB‐HSPA, see
part 9.5.1
- Throughput prediction is run for a terminal type which is with a configured HSDPA UE
Category supporting one of the multi-cell modes of Dual-cell, 3C (3-cell), 4C (4-cell), 6C
(6-cell) or 8C (8-cell), see part 9.6.7
- There are defined in the main and second operating frequency bands at least as many
active HSDPA cells as the number of cells supported by the terminal
The difference between peak rate, throughput and application throughput are exactly the same
than those detailed in the HSDPA study part, just before.
WARNING!!!
Soft handover gain is not taken into account in HSUPA study even, if it is selected in the service
table
9.10.6 MULTI-STOREY
Air to ground design, where predictions for several altitude ranges are requested
High RSSI levels observed in upper floors during some indoor projects
• Click on the Multi-storey prediction icon and a dialog box will appear
You can enter a name other than the name of the selected coverage prediction.
The predictions will be created using this name suffixed with the receiver height
of each prediction.
You can enter a resolution other than the resolution of the selected coverage
prediction.
Under receiver heights, enter the nb of storeys and the Rx heights for the first
and next storeys. The nb of coverage predictions calculated by the add-in will
correspond to the nb of storeys.
The add-in extrapolates Rx heights corresponding to the next storeys from the
values you enter for the first two storeys when checking the automatically adjust
receiver heights for the next storeys box.
Below the process on how to import Drive Test data is presented, and it is quite the same than
the one done for CW measurements, when there is a difference, it is specified.
Select the file corresponding to the drive tests measurements done; the format should be .txt
or .csv
Then specify the first measurement row, (here in the example it is the line 2), it will automatically
update the name in the “Field” line, as the values set in line 1
Then click on Setup and select the right name corresponding to X and Y column, also set
manually the column name where Primary Scrambling Code (PSC) or, alternatively, Physical cell ID
is provided (for CW measurements it is not Physical Cell ID which should be indicated manually
but column where CW measurements are set).
To select a criterion measured right click on “Drive Test Data” (for CW measurements it should
be on “CW measurement” folder) select properties, then go on Display table, select in Display
Type “Value Intervals”, and then the measured criterion required in “Field”
A much accurate comparison is possible by doing it point by point based on the steps
Right click on Drive Test Data and select Calculations, Create a New Prediction
For CW measurements the comparison with predictions is done in a similar way, point by point,
by following some similar steps
Then right click on CW measurements and select Display Statistics to have the results
• clutter classes file ( can be named clutter or DLU) imported as Clutter Classes in 9955
For 2.5D database, there is a third file which contains each database bin height, it can be named
Clutter height, or DLH and should be imported as Clutter Heights in 9955
Sites table is exactly the same as described for UMTS template in part 9.2
LTE Transmitters table has the same structure than what is described in part 9.3.
Also same functionalities are kept, by right clicking on LTE Transmitters, the user can:
• Open Cells table whose structure is the same than in part 9.3.1
• For LTE predictions, like for any LTE template, right click on
Predictions/New the following window appears
Then display “ Customized Predictions UMTS” to have all ALU prediction settings, like in the
following window
To perform Monte Carlo simulations see the process described in document linked as [R22] in
part 11.
9.13.2 PARAMETERS
Network settings structure and data are the same than the description in part 9.6
In Services the only difference is all the services are defined including GSM, UMTS and LTE ones,
otherwise UMTS field in services table are the same than those described in part 9.5.3
In Terminals table each technology terminal is defined and related to its Reception equipment
performances like described in part 9.5.1 for UMTS.
In User Profiles, there is the possibility to define a multi-technology user, for services common
to technology download; each profile will be attached to a technology by the type of terminal
used
Also for Radio Network Equipment, there is no modification compare to what is available in UMTS
template
9.13.3 GEO
This part has not changed, only the clutter table contains separated column per technology for
Model standard deviation, C/I standard deviation...
• For small cells analysis it is highly recommended to have accurate database to position
small cells in the streets. So 2.5D databases is the best option to place and optimize
small cells using SPM2.5D small cells default models.
• All the process and description for Forsk ASP/ACP settings and using is available on the
document linked as [R55] in part 11 of this guideline.
To take into account backhauling constraints Small Cell Selection module can be used. All
settings are detailed in document [R53] referenced in part 11.
For small cell placement based on geo-localized data all the process is detailed in document
referenced as [R53] in part 11.
• provided as a new integrated modules they reuse 9955 project network design, zones,
clutter classes and templates
• all pathloss (PL) calculation done in ASP/ACP can be reused in the 9955 platform
ASP can be used to deploy new sites for a greenfield network. This can be the case for
countrywide coverage projects (regulatory aspects), where a 2D database is used.
Before launching ASP, make sure that there is a site template per environment (DU / U / SU / RU)
configured according to your project’s need:
• Frequency band;
• Antenna types;
• Sites’ height;
• Propagation model
• Launch calculation
• Commit results
Figure 138: Configure scenario: choose target area and Station template to use
Figure 140: Configure scenario: review/update target objective, (optional) set clutter specific design
thresholds
Note: in case of indoor coverage is used please review/update accordingly all the design
thresholds configured for the existing ‘Gold’, ‘Silver’ and ‘Bronze’ objectives.
ASP can be used to deploy new sites for an existing 3G network (network densification). This can
be the case when coverage/capacity performances of existing network need to be improved.
Note that there is no restriction regarding type of new sites to add to the current design (it can
be either macro or metro) it is specified in the same way as for the Greenfield option, see
section 10.1, by choosing an appropriate existing ‘Station templates’.
Before launching ASP, make sure that there is an appropriate site template per environment (DU
/ U / SU / RU) configured according to your project:
• Frequency band;
• Antenna type;
• Sites’ height;
• Propagation model
This part details the method how to prepare 9955 environment before accessing 9955 ACP
functionalities.
The aim is to select the right candidate sites and to optimize tilt and azimuth. This exercise can
be done in the all along the site selection phase, and during the pre-optimization phase.
In order to realize an optimization in line with RF design targets recommended in part 8.5, 9955
transmitters table and cell tables should be set as it is recommended in parts 9.2 & 9.3 ; and
ACP targets should be those detailed in part 10.4
• Identify the cluster to optimize, the number of sites for this cluster should not exceed
500 sites.
o The cluster must be surrounded by a ring of sites which have already been
implemented on the field but which would not be optimized and used as
interferer.
Figure 150: Example of computation zone & focus zone setup in 9955 before using ACP
o A focus zone can be defined which contains exclusively the sites to optimize. The
focus zone must be inside the computation zone.
Note: ACP is very flexible with using the different zones at Geo tab, in case there are some hot-
spot zones defined they can be used as well for target area to be optimized (we can have several
different non-intersecting polygons defined!) or in case of focus zone is not defined it can
optimize all the sites found inside the computation zone.
• Run ACP optimization and commit optimization results in-line with a chosen optimization
strategy
After the computation or focus zone definition which delimits the area where the
optimization should be done; the user should respect the following steps
Please note, if a site selection is expected, then it is recommended the sites concerned not
to be activated. Then, these sites will be considered as ‘Current Candidate’ ones while a
current candidate selection process will take place with applying some defined selection
policies as of described in section 10.4.5. Hence we can ensure a very selective and rf
optimized site activation (example strategy: the most appropriate one single site is activated
out of five possible candidates).
‘Optimisation’ tab of a new ACP setup allows designers to set/choose the main design
optimization parameters:
- Set ‘Number of iterations’; value shall be higher than then minim need one in order to
achieve convergence (minim value is indicated as ‘Recommended value)
- Optimization environment, analysis options and other global parameter options grouped
into the subfolders of:
o Technology layers
o Zones
o Cost Control
o Constraints
By default only used in the project technology layers will be displayed and be available to
work with. In case of a MultiRAT project you may have several entries, belonging to different
technologies, available. Be aware that ACP will work simultaneously with all the selected
layers, hence ensure that you have the complete network setup for all the considered
technology layers before creating ACP setup.
There are two important optimization mode selection options available as simple
checkboxes:
- ‘Reconfiguration’ – only active sites/sectors from within the optimization area will be
considered for optimization (reconfiguration). Only this option selected will allow to
chose one or more of the transmitter/cell parameters of an active site to be optimized:
antenna type, electrical tilt, mechanical tilt, antenna height, sector azimuth and pilot
power, see 10.4.5.
- ‘Site selection’ – with this option we might go for current site selection (even with
shouting down some currently active sites, sectors) or for candidate site selection
(activate some new candidate sites) with a possible option to import/add new possible
candidates into the working site list, see 10.4.5.
At the ‘Zones’ subfolder you can specify options for area accounted for objectives’
evaluation but as well option for area to be optimized. By default you are recommended to
do the evaluation over the full computation zone but optimize only the sites falling inside
the focus area. At ‘Zone parameters’ you will be indicated with some details of these existing
zones but as well the option to change priorities of these.
It is possible to limit the max number of active sites when working with candidates. Although
this is only a global limit on total number of sites, would you need more control on site
candidate’s activation please check section 10.4.5.
At ACP Setup/Objectives the optimization targets to be used can be specified. By default the
3G objectives of ‘UMTS RSCP Coverage’ and ‘UMTS EcIo’ objectives are defined, see figure
above.
o For Coverage target, the user should define the RSCP target based on Link
budget max available path loss (MAPL). As ACP use the clutter losses set in RNP
clutter table with setting detailed in part 9.4, a common target can be set for all
clutter types.
• The user has to identify each clutter to the four case in the LB,
which are Dense urban, urban suburban or rural
o For Ec/Io target, for all the clutter set -15dB, in case of 90% cell load assumption
(90% of load in 9955 Total power column in cell table), which is our
recommendation but if the customer required another target, the user should
set this one.
To ensure that 4dB overlapping criterion is respected the objective of ‘UMTS 1st –Nth
difference’ should be used and set as below in order to ensure a minimum of 4dB delta
between the serving cell and the 2nd server in more than 65% of the area (max. 35% of the
area with 2 and more servers in 4 dB range)
Alternatively, the objective of ‘UMTS Pilot pollution’ can be used, configured according to a
formerly created ‘Overlapping Zones’ prediction type defined with a margin of 10dB (as well,
you can define this manually but mandatorily before creating the objective), and set as below
in order to ensure a minimum of 10dB delta between the serving cell and the 7th server in at
least 98% of the area (max. 2% of the area with 7 or more servers in 10dB range).
Before setting any optimization parameter by accessing the ‘Reconfiguration’ tab you may
define the way ACP will work with the antennas during optimization, i.e. for example, in case of
antenna electrical tilt optimization you may want to have all the pattern variants of a given
antenna type indicated. This can be buy grouping antenna patterns. Select all antenna patterns
belonging to the same antenna (different electrical tilts and click on the arrow to group them
under a specified antenna element name. You may define several different antenna elements
ACP to work with. Note that antenna elements shall contain patterns standing for given
frequency band. In case of a multiband antenna, you may define antenna elements as per
different supported frequency band and then grouping them into a physical antenna definition.
Unless multiband is your case physical antennas will be defined based on single antenna
elements.
This is the place where you can indicate design parameters which are going to be optimized.
They are grouped into three groups: belonging either to Sites, Transmitters or UMTS Cells.
Option only available if ‘Site Selection’ checkbox at ACP Setup/Optimisation/Tehnology layer was
selected. If not selected there will be no sites displayed in the list at Reconfiguration/Sites.
Having Site reconfiguration disabled, i.e. not allowing site optimization can be useful when
handling small cells design where macro layer shall not be changed. Otherwise existing site
sector or entire site can be removed, candidate sites can be activated (they were defined but
formerly not active) or even new candidates can be added into the project.
Note that adding new candidate sites is not equivalent to ASP which is with positioning new
UMTS sites. ACP’s new candidate site selection is requiring a site list with known site locations
indicated. This list can be populated either from the current Sites table, by considering current
entries with no transmitters linked to, or importing new site locations from a file or, in case of a
MultiRAT project, imported from the other technology layers’ sites tables.
A too high tilt value will induce to have the main path impacted by side lobes antenna
pattern zone which is not stable and so not predictable
WARNING!
• Minimum step between two tilt values must be at least 2°, having 1° doesn’t change the
design and increase computation time
Notes:
- the electrical tilt optimization required to set in 9955 an antenna pattern per antenna type and
tilt value. Antenna pattern grouping, in-line with recommendations of 10.4.4, is recommended.
Doing this allows for a given sector to optimize tilts only based on the pattern included in the
same group (under same antenna element name) than the one already assigned before ACP
launch.
- Do not select too much groups and patterns, it increases exponentially computation time
- “Electrical Tilt range should be from 0° to 10° or 12°, and sometimes 16° (depends on
environment & availability of pattern files as per tilt)
- Mechanical tilt optimization should be combined with Electrical tilt optimization if the pattern list
per antenna model is not exhaustive or if the tilt range on the antenna models which must be
used is not sufficient
- Mechanical tilt max ranges shall be specified in order not to exceed the recommended total tilt
value (in case of electrical tilt is used) or set the same range as for electrical tilt (see above) in case
of only mechanical tilts alone (no e-tilt pattern files are available for different e-tilt values!) are
optimized
ACP may take into account a known traffic distribution in the optimization process via ‘pixel
weighting’ of one or more defined objectives.
Right click on a newly created ACP Setup and select ‘Run…’ This will launch the ACP optimization
process in-line with the ACP Setup options/selections.
During the ACP run you can check the real-time evolution of the optimization process, not only
as a graph on the progress of Objectives’ achievement (in %-age) but in detailed views by
accessing the successive tabs of “Changes’, ‘Quality’ and ‘Objectives’
As soon as ACP run is finished an ‘Optimisation’ result folder is created under the Setup one with
details on ACP results.
The presented information is allowing you to check all the results in forms of statistics, detailed
plots on before-after performances of different objectives, histograms, implementation plan
proposal and option to commit on results.
ACP can be used to optimize the number of outdoor small cells implemented in a macro-layer.
All ACP settings and different steps are described in part 11.2 of guidelines linked as [R54] in
section 11. This includes the tables which make the link between macro cell RSRP level where
small cells are placed and metro cells radius achieved.
To take into account backhauling constraints Small Cell Selection module can be used. All
settings are detailed in document [R53] referenced in part 11.
For small cell placement based on geo-localized data all the process is detailed in document
referenced as [R56] in part 11.
Livelink: https://wcdma-ll.app.alcatel-
lucent.com/livelink/livelink.exe?func=ll&objId=68125231&objAction=browse&viewType=1
Livelink: https://wcdma-ll.app.alcatel-
lucent.com/livelink/livelink.exe?func=ll&objId=68125231&objAction=browse&viewType=1
https://wcdma-ll.app.alcatel-
lucent.com/livelink/livelink.exe?func=ll&objId=67696328&objAction=browse&viewType=1
Livelink: https://wcdma-ll.app.alcatel-
lucent.com/livelink/livelink.exe?func=ll&objId=68551080&objAction=browse&viewType=1
https://wcdma-ll.app.alcatel-
lucent.com/livelink/livelink.exe?func=ll&objId=68549511&objAction=browse&viewType=1
https://wcdma-ll.app.alcatel-
lucent.com/livelink/livelink.exe?func=ll&objId=68125231&objAction=browse&viewType=1
Livelink: https://wcdma-ll.app.alcatel-
lucent.com/livelink/livelink.exe?func=ll&objId=53160351&objAction=browse&sort=name&viewType=1
Livelink: https://wcdma-ll.app.alcatel-
lucent.com/livelink/livelink.exe?func=ll&objId=68080962&objAction=browse&viewType=1
Livelink: https://wcdma-ll.app.alcatel-
lucent.com/livelink/livelink.exe?func=ll&objId=68078989&objAction=browse&viewType=1
Livelink: https://wcdma-ll.app.alcatel-
lucent.com/livelink/livelink.exe?func=ll&objId=68080961&objAction=browse&viewType=1
Livelink: https://wcdma-ll.app.alcatel-
lucent.com/livelink/livelink.exe?func=ll&objId=68080970&objAction=browse&viewType=1
Livelink: https://wcdma-ll.app.alcatel-
lucent.com/livelink/livelink.exe?func=ll&objId=68170433&objAction=browse&viewType=1
Livelink https://wcdma-ll.app.alcatel-
lucent.com/livelink/livelink.exe?func=ll&objId=67320862&objAction=browse&viewType=1
Livelink: https://wcdma-ll.app.alcatel-
lucent.com/livelink/livelink.exe?func=ll&objId=68404785&objAction=browse&viewType=1
Livelink: https://wcdma-ll.app.alcatel-
lucent.com/livelink/livelink.exe?func=ll&objId=68078685&objAction=browse&viewType=1
[R26] “UMTS900 GSM900 Co-existence Simulation Study Based on ALU UMTS900 ProductV4”
Livelink https://wcdma-ll.app.alcatel-
lucent.com/livelink/livelink.exe?func=ll&objId=68079333&objAction=browse&sort=name
[R30] " Guidelines for Antenna Isolation between ALU UMTS and other techno“
Kiosklive reference: UMT/IRC/APP/032700
Livelink https://wcdma-ll.app.alcatel-
lucent.com/livelink/livelink.exe?func=ll&objId=68079333&objAction=browse&sort=name
[R33] ” Guidelines for Antenna Isolation between LTE and other technologies ”
Kiosklive reference: LTE/IRC/APP/032647
Livelink: https://wcdma-ll.app.alcatel-
lucent.com/livelink/livelink.exe?func=ll&objId=68079333&objAction=browse&sort=name
Livelink: https://wcdma-ll.app.alcatel-
Livelink: https://wcdma-ll.app.alcatel-
lucent.com/livelink/livelink.exe?func=ll&objId=68078989&objAction=browse&viewType=1
Livelink: https://wcdma-ll.app.alcatel-
lucent.com/livelink/livelink.exe?func=ll&objId=68081433&objAction=browse&viewType=1
[R44] 9955 default propagation models recommended for 400 - 5500 MHz frequency bands
Kiosklive : LTE/IRC/APP/032646
Livelink : https://wcdma-ll.app.alcatel-
lucent.com/livelink/livelink.exe?func=ll&objId=68078989&objAction=browse&viewType=1