ECC Report 296

ECC Report 296
National synchronization regulatory framework options in

3400-3800 MHz: a toolbox for coexistence of MFCNs in
synchronised, unsynchronised and semi-synchronised
operation in 3400-3800 MHz
Approved 8 March 2019

ECC REPORT 296 - Page 2
0 EXECUTIVE SUMMARY
The purpose of this Report is to support Administrations in setting up the synchronisation frameworks at
national level for the introduction of 5G-NR in the 3400-3800 MHz band in a multi-operator environment
leveraging on the synchronised, unsynchronised and semi-synchronised modes.
This Report extends the contents in previous ECC Report 216 [1] and in ECC Report 281 [2] to account for
the following new aspects:
 5G-NR new frame structures bring new compatibility and performance aspects to be considered in the
case of synchronised operation between 5G-NR and LTE-TDD, which make it desirable to also consider
unsynchronised operation;
 The adoption of Active Antenna System (AAS) technology to MFCN base stations brings new challenges
for unsynchronised operation (in terms of cost-effectiveness and spectrum efficiency of the LRTC's
implementation), which makes it desirable to consider synchronised operation;
 Semi-synchronised operation is a new mode of operation that was not studied previously (LRTCs defined
were conservative and aligned with unsynchronised operation in ECC Decision (11)06 (October 2018)
[3].
Section 2 provides a general introduction about synchronised, semi-synchronised and unsynchronised

operation including definitions, benefits and challenges.
Section 3 assesses the performance impacts (in terms of spectrum efficiency, UL/DL throughput and latency)
for a few selected 5G-NR frame structures. The analysis also addresses options for the “LTE compatible”
5G-NR frame structures, suitable for the cross-technology synchronised operation between a 5G-NR and
LTE-TDD networks.
Section 4 assesses the applicability of different mechanisms to manage the cross-link interference deriving
from simultaneous UL/DL transmissions in case of unsynchronised and semi-synchronised operation when
the ECC “baseline” out of block power limit (defined for the synchronised operating mode in ECC Decision
(11)06 (October 2018)) is applied. This section analyses the impact of geographic separation between
Macro-cellular networks as well as the impact associated with the adoption of alternative topologies (Micro
BSs and Indoor BSs networks).
Leveraging on the analysis and studies performed in the previous sections, Section 5 proposes a “toolbox”
with options to support Administrations and operators in identifying the most appropriate synchronisation
regulatory framework at national level. The key elements from the “toolbox” are summarised hereafter.
Synchronised operation avoids any BS-BS and MS-MS interference therefore allowing coexistence
between adjacent networks without the need for guard bands or additional filters. This operating mode
simplifies network deployment because no additional interference mitigation is required. However, in order to
1
implement this, within each deployment area/region, all MFCN licensees operating in the same band should
use:
 A common phase clock reference (e.g. UTC), with proper accuracy/performance constraints that depend
on the underlining technology, and permanent monitoring and agreed remedies in case of accuracy loss.
Those aspects and challenges are detailed in ECC Report 216;
 A compatible frame structure to avoid simultaneous UL/DL transmission, which determines a specific
DL/UL transmission ratio and frame length. The chosen frame structure will contribute to the network
2
performance (e.g. latency, spectral efficiency, throughput and coverage ). The feasibility and
performance impacts of synchronised operation between different radio technologies have to be
assessed on a case-by-case basis depending on the specific technologies. As assessed in this Report,
1
Not limited to the licensees with adjacent blocks.
2
For example: the size of the guard periods between DL and UL transmissions will have an impact on maximum cell radius. Increasing
the number of UL transmissions has an impact on the UL coverage performance.
the synchronised operation of 5G-NR and LTE-TDD may imply a cost in terms of user plane latency and
performance, especially with regards to 5G URLLC latency targets. Agreements on synchronised
operation between operators will be simplified when the same type of services are targeted with the
associated desired user plane latency and performance targets.
Unsynchronised operation does not require the adoption of a compatible frame structure among licensees.
Licensees select the most appropriate frame structure independently and can adapt the frame structure to
service and end user requirements, which may change depending on the location and on time. However, in a
multi-operator scenario, the flexibility in operators’ frame structure selection leads to a number of interference
scenarios that need to be assessed and managed.
The cross-link interference deriving from simultaneous UL/DL transmissions could be managed with the
adoption of the ECC restricted baseline out of block power limit which, based on currently available
technology, would imply operator-specific filters and inter-operator guard bands. It is assumed that it will be
challenging to implement operator-specific filters cost effectively in case of AAS BS, and that significant
guard bands would be needed in such case (i.e. in case of implementation of the ECC restrictive out of block
power limits). Therefore, this Report emphasizes on the need to implement a framework that would not
require the implementation of the ECC restrictive out of block power limits, and provides an analysis on
whether and under which conditions the unsynchronised operation can be used when base stations
implement the ECC baseline out of block power limits. Here follows a concise summary of the outcomes:
 Unsynchronised Macro-cellular networks: Based on currently available filtering technology, guard
bands and operator specific filters are necessary to enable unsynchronised operation between
operators. Alternatively, geographic separation distances could be necessary but a specific
recommendation or single set of trigger values cannot be provided due to the dependency on various
factors. The studies show minimum distances required between unsynchronised Macro-cellular networks
could be up to 60 km when operating co-channel and up to 14 km when operating in the adjacent
3
channel without guard bands .
 Unsynchronised Micro BS networks and Macro-cellular networks: the studies show that, in general,
adjacent channel unsynchronised operation of Macro-cellular networks and Micro BS networks might not
be feasible in the same area. Separation distances have not been assessed in this Report. If there is no
Macro-cellular network, adjacent channel unsynchronised operation between two Micro BS networks
might be feasible with careful planning avoiding line of sight.
 Unsynchronised Indoor BS networks and Macro-cellular networks: under specific assumptions,
4
adjacent channel unsynchronised operation should be possible with careful installation of the indoor
BSs. Such planning seems to be feasible in case of industrial – type of use case (e.g. smart factory
indoor coverage). In the case of co-channel operation of Macro BSs and indoor BSs, the lack of out of
block filtering on the Macro BS and on the indoor BS transmitters' sides will need to be considered.
Semi-synchronised operation corresponds to the case where part of the frame is consistent with
synchronised operation as described above, while the remaining portion of the frame is consistent with
unsynchronised operation as described above.
This leads to a limited degree of frame structure flexibility at the expense of some additional interference that
can be controlled to some extent (for example, this operating mode could be implemented to avoid
simultaneous UL/DL transmissions for control channels). All MFCN licensees operating in the same band
and same coverage area/region should use:
 A common phase clock reference, as for synchronised operation;
 Partial frame alignment: the agreement shall define a default frame structure as for synchronised
operation (for which UL/DL directions are defined across the whole frame) and at the same time the part
of the frame where each operator is allowed to reverse the default transmission direction (flexible part). If
3
These studies have assumed a "fully unsynchronised", also called "anti-synchronised" situation between the cellular networks (see
Section 4).
4
For example, ”careful installation” would include measures like ceiling-mounted installation, placement of indoor BS away from
windows, additional shielding around buildings in the worst case. Such measures may be more appropriate for professional
installations which seem less suitable for consumer-type of scenario (without further mitigation schemes implemented in the indoor
BS).
no changes are applied to the default frame structure, the semi-synchronised operation is identical to the
synchronous case;
 The terms and conditions under which the ECC baseline out of block power limit can be applied to the
semi-synchronised operation. These terms and conditions could be agreed between all operators, or
they could be facilitated by the reguIator. Some semi-synchronised scenarios might not require
regulatory intervention.
With respect to semi-synchronised operation with the ECC baseline out of block power limit, it is useful to
distinguish DL to UL and UL to DL modifications compared to the reference frame:
 "DL to UL modifications": the default DL transmission direction in the flexible part is modified into UL
 From BS-BS interference perspective, the network that modifies the default DL into UL will not interfere
with the other network(s) but it will receive additional interference from the other network(s);
 In most circumstances, MS-MS interference will be negligible because terminals typically transmit
intermittently and many will be mobile so any interference would be transient. It is expected that some
5
5G use cases will imply the deployment of MSs that are in fixed positions and close to each other which
6
would result in permanent MS-MS interference . No specific studies were performed on MS-MS
interference, therefore, in case of MSs that are in fixed positions and close to each other, no conclusion
can be derived. In any case, MS RF requirements are handled by SDOs and associated harmonised
standards.
 "UL to DL modifications": the default UL transmission direction in the flexible part is modified into DL
 From BS-BS interference perspective, the network that modifies the default UL transmission direction
into DL will interfere the other network while it will not receive additional interference from the other
network;
 Coexistence is facilitated if semi-synchronised operation is applied to Micro and indoor BS but it could be
technically challenging for indoor BS to be semi-synchronised with outdoor networks; and
 Coexistence could be more challenging if semi-synchronised operation is applied to Macro BS before
efficient interference cancellation algorithms have been developed and implemented. At the date of
publication of this Report, 3GPP is studying such algorithms.
As a consequence, and considering that:

 With reference to the unsynchronised operation, it is assumed that the implementation of the ECC
restricted baseline limit would imply operator-specific filters, which are challenging to implement cost
effectively in the case of AAS BS with currently available technology. In addition, the implementation of
the ECC restricted baseline limit would also require significant inter-operator guard bands, which are
highly undesirable. Without operator-specific filters, it may not be possible to rely on guard bands alone
to enable unsynchronised operation between Macro cells, and significant separation distances are likely
to be required (up to 60 km when co-channel and up to 14 km when operating in the adjacent channel).
Based on currently available AAS BS technology, it is therefore assumed that equipment will only
implement filters designed to comply with the ECC baseline out of block power limit;
 Synchronised operation requires operators to agree on a compatible frame structure (which determines a
specific DL/UL transmission ratio) and frame length, which contribute to the network performance (e.g.
latency, spectral efficiency, throughput and coverage). Compatible frame structure may also introduce
new operational constraints and additional costs: for instance, inter-operator synchronisation may lead to
a less flexible DL/UL ratio selection, resulting in suboptimal spectrum utilisation for an individual operator
(see ECC Report 216, section 3.3). Agreements on synchronised operation between operators will be
simplified when operators adopt the same technology (e.g. 5G-NR) and target the same type of services
with the associated desired user plane latency and performance targets;
 Synchronised operation and semi-synchronised operation both require a multilateral agreement between
all MFCN licensees in the same band and coverage area / region (not limited to the licensees with
adjacent blocks). Such agreement could be facilitated by the regulator (e.g. for reasons described in
ECC Report 216 section 3.3).
5
E.g. crowded stadiums, trains, buses, (home) CPEs in fixed wireless access (FWA) systems.
6
The MS-MS interference will occur for MSs close to each other while communicating with different BSs / networks.
A general framework could be defined at the national level by Administrations wishing to do so specifying:
 The technical parameters for synchronised operation , and for semi-synchronised operation if appropriate
(including reference clock and reference frame structure);
 The scope of synchronised, semi-synchronised and unsynchronised operation in terms of geographical
areas and type of cells: E.g. in the case of AAS BSs, the general framework could specify that
unsynchronised operation could only be implemented in those cases where additional isolation is
available (e.g. separation distances would still allow the use of the ECC baseline out of block limit); in the
case of semi-synchronised operation the general framework could specify in which scenarios DL slots
may / may not be unilaterally converted to uplink slots, depending on national circumstances;
 Mechanisms to ensure the periodic review of the agreed conditions.
TABLE OF CONTENTS
0 Executive summary ................................................................................................................................. 2
1 Introduction ............................................................................................................................................ 10
2 Synchronisation modes for TDD networks coexistence ................................................................... 12

2.1 Synchronised operation ................................................................................................................. 12
2.1.1 Definition ........................................................................................................................... 12
2.1.2 Benefits and challenges of synchronised operation ......................................................... 14
2.2 Unsynchronised operation ............................................................................................................. 16
2.2.1 Definition ........................................................................................................................... 16
2.2.2 Benefits and challenges of unsynchronised operation ..................................................... 16
2.3 Semi-synchronised operation ........................................................................................................ 18
2.3.1 Definition ........................................................................................................................... 18
2.3.2 Benefits and challenges of semi-synchronised operation ................................................ 19
2.4 Operating modes summary ........................................................................................................... 20
3 Performance impacts from frame structure selection ....................................................................... 22

3.1 General information about 5G-NR frame structure ....................................................................... 22
3.2 Performance impacts of 5G-NR frame structures ......................................................................... 22
3.3 Compatibility of 5G-NR frame with LTE-TDD frame structure....................................................... 24
3.3.1 Summary from Study #1 (grant-based UL transmissions)................................................ 25
3.3.2 Summary from Study #2 ................................................................................................... 27
3.3.2.1 DL capacity assessment (grant-based UL transmissions) ................................ 29
3.3.2.2 UL capacity assessment (grant-based UL transmissions) ................................ 30
4 Interference mitigation for unsynchronised operation with ECC baseline out of block power limit
................................................................................................................................................................. 31
4.1 Guard band requirement for unsynchronised operation ................................................................ 32
4.2 Geographic separation of Networks .............................................................................................. 32
4.2.1 Proposed methodologies .................................................................................................. 33
4.2.1.1 Method #1: Separation distance calculation ..................................................... 33
4.2.1.2 Method #2: Trigger Values calculation .............................................................. 34
4.2.2 Summary of the studies .................................................................................................... 35
4.2.3 Conclusions from studies .................................................................................................. 36
4.3 Coexistence between unsynchronised Micro BS and Macro BS deployments ............................. 36
4.3.1 Coexistence between unsynchronised Micro BSs and Macro BSs – Study #5................ 36
4.3.2 Coexistence between unsynchronised Micro BSs and Macro BSs – Study #6................ 36
4.3.2.1 Network topologies and main assumptions ....................................................... 36
4.3.2.2 Simulations results and conclusions ................................................................. 40
4.4 Coexistence between unsynchronised Indoor BSs and Macro BS - Study #7 .............................. 40
4.5 Study on semi-synchronisation for Microcell and Macro-cell cases - study #8 ............................. 41
4.6 Summary of all studies performed ................................................................................................. 43
5 Toolbox: options for administrations to support the design of synchronisation frameworks at

national level .......................................................................................................................................... 48
5.1 Key aspects to be considered when setting up the synchronisation framework at national level . 48
5.2 Enabling synchronised, unsynchronised and semi-synchronised operation at national level based
on multilateral agreements among MFCN licensees ..................................................................... 48
5.2.1 Synchronised operation .................................................................................................... 48
5.2.2 Unsynchronised operation based on the ECC baseline out of block power limit ............. 49
5.2.2.1 Options for enabling the unsynchronised operation involving macro-cellular
networks ............................................................................................................ 50
5.2.2.2 Options for the unsynchronised operation involving Micro BS networks in the
same area: ......................................................................................................... 50
5.2.2.3 Options for the unsynchronised operation involving indoor BS: ....................... 51
5.2.3 Semi-synchronised operation based on the ECC baseline out of block power limit ........ 51
5.3 Options for agreements ................................................................................................................. 53
6 Conclusions............................................................................................................................................ 55
ANNEX 1: Synchronisation frameworks in recent C-band award procedures........................................ 59
ANNEX 2: Technical solutions for network synchronisation.................................................................... 62
ANNEX 3: 5G-NR and LTE-TDD Frame structures, options and assessments - Study #1 and #2 ........ 64
ANNEX 4: MFCN parameter values and assumptions for simulations .................................................... 91
ANNEX 5: Unsynchronised operation of two Macro BS MFCN networks not in the same area- Study
#3 and #4 ................................................................................................................................................. 96
ANNEX 6: Coexistence studies between unsynchronised Micro BSs and Macro BSs - Study #5 and #6
............................................................................................................................................................... 109
ANNEX 7: Coexistence study between unsynchronised indoor BS and Macro base stations - Study #7
............................................................................................................................................................... 125
ANNEX 8: Coexistence between semi-synchronised MFCN networks - Study #8................................ 131
ANNEX 9: Status of unsynchronised and semi-synchronised operation in 3GPP ............................... 134
ANNEX 10: List of References .................................................................................................................... 136

LIST OF ABBREVIATIONS
Abbreviation Explanation
3GPP 3rd Generation Partnership Project

AAS Active Antenna System
ACI Adjacent Channel Interference
ACIR Adjacent Channel Interference Ratio
ACK Acknowledgment
ACLR Adjacent Channel Leakage Ratio
ACS Adjacent Channel Selectivity
ARQ Automatic Repeat reQuest
BEM Block Edge Mask
BS Base Station
CA Carrier Aggregation
CEPT European Conference of Postal and Telecommunications Administrations
CQI Channel Quality Indicator
CSI Channel State Information
DL Downlink
DwPTS Downlink Pilot Time Slot
EC European Commission
ECC Electronic Communications Committee
e.i.r.p. Equivalent Isotropically Radiated Power
FDD Frequency Division Duplex
GP Guard Period
GSM Global System for Mobile Communications
HARQ Hybrid Automatic Repeat reQuest
IMT International Mobile Telecommunications
ISD Inter-Site Distance
LoS Line of Sight
LRTC Least Restrictive Technical Conditions
LTE Long Term Evolution
eMBB Enhanced Mobile Broadband
MIMO Multiple Input Multiple Output
MFCN Mobile/Fixed Communications Network
MS Mobile Station
Abbreviation Explanation
MU Multi User
NLoS Non Line of Sight
NR New Radio
OFDM Orthogonal Frequency Division Multiplexing
OOB Out of Band
PRTC Primary Reference Time Clock
RAN Radio Access Network
RF Radio Frequency
RTT Round Trip Time
SCS Sub Carrier Spacing
SEM Spectrum Emission Mask
SLA Service Level Agreement
SLAV Side Lobe Attenuation
SR Scheduling Request
SUL Supplemental UpLink
TAI International Atomic Time
TDD Time Division Duplex
TRP Total Radiated Power
TSG Technical Specification Group
UL Uplink
UMa Urban Macro-cellular
UMi Urban Micro-cellular
UMTS Universal Mobile Telecommunications System
UPT User Perceived Throughput
UpPTS Uplink Pilot Time Slot
URLLC Ultra Reliable Low Latency Communications
UTC Coordinated Universal Time
WCDMA Wideband Code Division Multiple Access
1 INTRODUCTION
The 3400-3800 MHz band has already been harmonised for MFCN in CEPT and is recognised to be the 5G
primary band in Europe. The 3400-3800 MHz band has more contiguous spectrum than lower frequency
bands and can allow wide channels which are necessary to make the 3400-3800 MHz band effective for 5G
deployments. Compared to lower frequency bands and taking into account its expected introduction in most
countries in the world this primary 5G band offers a combination of higher bandwidth and higher capacity as
well as a good potential to become a future worldwide band for 5G.Many national Administrations within
CEPT plan to enable initial 5G deployments in the 3400-3800 MHz band. At the date of publication of this
Report, several European countries have awarded frequencies in at least a portion of the 3400-3800 MHz
band since 2015 (e.g. Czech Republic, Finland, Hungary, Ireland, Italy, Latvia, Norway, Poland, Romania,
7
Slovakia, Spain and the UK ). Other Administrations (e.g. Austria, Cyprus, France, Germany, Sweden,
8
Switzerland, Portugal and the UK ) are planning to conduct similar auctions. Information on the
synchronisation aspects during latest 3400-3800 MHz spectrum award procedures is provided in ANNEX 1.
While preparing the assignment procedures, Administrations may find ways to ensure the definition of the
most appropriate synchronisation framework accounting for the local circumstances and local market
demand.
The aim of this Report is to support the setup of the most suitable synchronisation framework at national
level. This Report relies on the previously published ECC Report 216 (August 14) [1], ECC Report 281 (July
’18) [2] and ECC Decision (11)06 (October 2018) [3].
 ECC Report 216 [1] provides band-neutral practical guidance for synchronisation of TDD networks. The
Report addresses specific BS-BS and MS-MS interference scenarios in case of unsynchronised
operation and provides background about synchronised operation, definitions, technical aspects for clock
and phase / time, cross-technology frame alignment between WiMAX / LTE-TDD, and options for
Administrations for designing a general framework at the national level for synchronised operation in a
multi-operator context;
 ECC Report 281 (July ’18) [2] and ECC Decision (11)06 (October 2018) [3] define the Least Restrictive
Technical Conditions (LRTCs) applicable to 5G MFCN using non-AAS and AAS based station systems
in the 3400-3800 MHz band. Such LRTCs extend the LRTCs defined in ECC Report 203 [5] (which was
based on IMT-Advanced / 4G). The LRTCs include the baseline out of block power limit and the
transitional regions power limits to be used in case of synchronised operation as well as the restricted
baseline out of block power limit to be used in case of unsynchronised and semi-synchronised operation;
 Synchronisation and coordination of TDD MFCN networks across national borders is addressed by ECC
Recommendation (15)01 [4].
This Report extends the contents in ECC Report 216 and in ECC Report 281 to account for the following
new aspects:
 5G-NR new frame structures bring new compatibility and performance issues in case of synchronised
operation between 5G-NR and LTE-TDD, which make it desirable to consider unsynchronised operation;
 The adoption of Active Antenna System (AAS) technology to MFCN base stations brings new challenges
for unsynchronised operation (in terms of cost-effectiveness of the LRTCs implementation), which makes
it desirable to consider synchronised operation;
 Semi-synchronised operation is a new mode of operation that was not studied previously (LRTCs defined
were conservative and aligned with unsynchronised operation in the aforementioned ECC Decision).
This Report provides:

 An analysis for the specific circumstances under which the unsynchronised operation could be allowed
when the ECC baseline out of block power limit is applied instead of the ECC restricted baseline out of
7
3410-3480 MHz and 3500-3580 MHz.
8
3600-3800 MHz.
block power limit (see ECC Decision (11)06 Table 3 and Table 4 [3]) when adopting specific interference
mitigation techniques or geographical isolation between networks;
 An analysis of the specific circumstances under which semi-synchronised operation could be allowed
when the ECC baseline out of block power limit is applied, providing some degree of flexibility in the
selection of UL/DL transmission direction at the cost of increased cross-link interference;
 An analysis of the performance impact of cross-technology synchronised operation between LTE-TDD
and 5G-NR;
 A general toolbox to help Administrations define a regulatory framework for synchronised, semi-
synchronised and unsynchronised operation for TDD networks at the national level.
2 SYNCHRONISATION MODES FOR TDD NETWORKS COEXISTENCE
Starting from the definitions provided in ECC Report 281 [2] for the synchronised, unsynchronised and semi-
synchronised operation, this section highlights benefits and challenges associated with each mode and
provides an overview on the interference mechanisms that characterise each operating mode.
Different interference scenarios may occur when two TDD networks are deployed in blocks within the same
band (including the co-channel case and the adjacent channel case). Cross link interference will occur when
simultaneous transmissions in uplink (UL) and downlink (DL) directions take place in different TDD networks
(i.e. one BS (or MS) belonging to one network transmits while another BS (or MS) belonging to the other
network receives (this will be referred to as "simultaneous UL/DL transmissions" throughout this Report).
As explained in the following Sections, simultaneous UL/DL transmissions do not take place in case of
synchronised operation while such kind of transmissions take place in case of unsynchronised and semi-
synchronised operation.
Figure 1 illustrates the interference scenarios in case of simultaneous UL/DL transmissions: the green
arrows represent the desired links, while the potential interference is represented by the yellow arrows. BS-
MS interference happens in all cases (FDD and TDD, whether synchronised or not) and is handled as part of
the standards. MS-MS and BS-BS interference in unsynchronised and semi-synchronised TDD networks are
within the scope of this Report.
Figure 1: Interference scenarios in case of simultaneous UL/DL transmissions in MFCN TDD

networks
2.1 SYNCHRONISED OPERATION
2.1.1 Definition
The word “synchronisation” is used in many different contexts with different meanings. For example, BS-MS
synchronisation within the same network, frequency and phase synchronisation at the carrier level for
demodulation purposes, frequency synchronisation for FDD networks such as GSM, etc. This Report will
only focus on phase / time synchronisation at the frame level between TDD networks for interference
mitigation purposes.
ECC Report 281 provides the following definition: “synchronised operation in the context of this Report
means operation of TDD in several different networks, where no simultaneous UL/DL transmissions occur,
i.e. at any given moment in time either all networks transmit in DL or all networks transmit in UL. This
requires non-simultaneous UL/DL transmissions for all TDD networks involved as well as synchronising the
beginning of the frame across all networks”.
In order to deploy synchronised TDD mobile networks in a multi-operator context, operators need to reach
agreement on:
 A common phase clock reference (e.g. UTC - Coordinated Universal Time) and accuracy / performance
constraints that depend on the underlining technology (e.g. +/- 1.5 μs for LTE-TDD and 5G-NR), either
using their own equipment to provide the clock, or sharing the same phase / time clock infrastructure;
 Permanent monitoring of the agreed clock source. When losing the primary reference time clock (PRTC)
equipment may continue operation for a period of time ("holdover period") that has to be agreed and
which depends on the quality of the local oscillator in the BS and on the wireless network accuracy
requirement. If the PRTC is lost for a duration longer than the holdover period, the system shall no
longer be considered in synchronised operation and may start interfering other channels, and therefore
proper action shall be taken (e.g. the BS shall be shut down until the PRTC is recovered);
 A frame structure (including TDD DL/UL ratio and frame length) in order to avoid simultaneous UL/DL
transmissions (guard periods may be different, as illustrated in Figure 3). The assessments in ANNEX 3
(summarised in Section 3) provide information on the implications associated with some specific but
representative frame structures in terms of throughput performance, spectrum efficiency and latency.
The following figure illustrates the frequency, phase and time synchronisation concepts, which are described
in [1].
Figure 2: Frequency, phase and time synchronisation concepts
The following table provides the synchronisation requirements for MFCN technologies including 5G-NR in
terms of frequency and phase accuracy.
Table 1: Frequency and phase synchronisation requirements for different MFCN technologies
Parameter
Technology Frequency accuracy

Phase accuracy relatively to
relatively to the reference
the reference clock
oscillator
GSM, UMTS, WCDMA, LTE-FDD 50 ppb N/A
50 ppb ±1.5 μs (for cell radius ≤ 3 km)

LTE-TDD
50 ppb ±5 μs (for cell radius > 3 km)
5G-NR 50 ppb ±1.5 μs
The following figure provides examples for simultaneous and non-simultaneous UL/DL transmissions in TDD
networks.
Figure 3: Examples of simultaneous UL/DL transmissions in TDD networks
In TDD networks, the maximum cell radius depends on the guard period between DL and UL transmissions:
the examples above show how operators may implement guard periods of different durations (enabling
different coverage radii) while maintaining compatible frame structures (i.e. while avoiding simultaneous
UL/DL transmissions.
2.1.2 Benefits and challenges of synchronised operation
The ECC has defined the baseline and transition region out of block power limits for synchronised operation
of MFCN BSs (see ECC Decision (11)06 (October 2018) Table 3 [3]). The ECC baseline accounts for the
fact that BS-BS and MS-MS interference scenarios do not take place in case of synchronised operation.
ECC baseline regulatory limit does not introduce additional constraints compared to the spectrum emission
mask as defined by the standards.
The purpose of synchronised operation is to prevent BS-BS and MS-MS interference scenarios.
Synchronised operation avoids performance degradation due to such interference without requiring
additional mitigation techniques such as additional filtering (that may be challenging to implement in AAS
BSs and MSs as will be described in section 2.2.2), inter-operator guard bands, geographical separation
between BSs, etc.
Synchronised operation therefore simplifies operators’ network deployments since less coordination for BS
radio planning is required among synchronised operators.
However, the requirements associated with synchronised operation as described in the previous Section also
lead to some challenges:
 Setup of the clock reference: operators have to agree on a common reference clock and common
accuracy / performance. The +/- 1.5 µs accuracy might be challenging to achieve in some cases.
Operators might consider deciding to share the clock infrastructure. Operators will in any case need to
setup such accurate clock solutions within their own networks regardless on the possible need to
synchronise their network with other networks;
 Clock quality monitoring and enforcement: since any imperfection in synchronisation affects other users
in the band, operators must constantly monitor their reference clock quality (depending on the
performance of the BS local oscillator) and take proper action (e.g. equipment shutdown if the reference
clock is lost for more than an agreed amount of time). Operators (and/or Administrations) should
therefore be able to test and enforce whether the clock quality is met;
 Compatible frame structure across operators: the frame structure determines a specific DL/UL
transmission ratio and frame length, which contribute to the network performance (e.g. latency, spectral
9 10
efficiency, throughput and coverage ). Therefore, the selection of a compatible frame structure will
provide the same contribution to the performance of all operators involved, with similar impacts on the
services to end users.
The compatible frame structures can be renewed over time, subject to the agreement. There are already
11
precedents for this . Some new mechanisms might be specified to review and periodically (involving
regulators if needed) or dynamically adjust such parameters (this option is currently considered as
challenging). For example, the agreement on a common DL/UL ratio could be based on a compromise
taking all operators' performance requirements into account.
The agreement between a small number of operators, potentially using the same technology, is easier to
achieve than an agreement between multiple operators, potentially using different technologies and
potentially targeting different services.
It is to be noted that the adaptability of DL/UL ratios in time and according to different geographic
locations may or may not be a market requirement in a given market.
Depending on the regulatory framework in place, the possible regulator choice for a “preferred frame
structure” could lead to problems in terms of compliance with the technology neutrality principle if the
chosen format would not be supported by some candidate TDD technology for the band.
In case of existing unsynchronised networks with locally / regionally assigned spectrum, the unsynchronised
operation of the 5G network would result in interference with the local / regional networks while the
synchronised operation of the 5G network with the local / regional networks would lead to interference within
the national network. In this case, it might be desirable to consider the synchronised operation also for the
existing local / regional networks.
All issues above apply in all cases of TDD coexistence, including in 5G-NR / 5G-NR and LTE-TDD / 5G-NR
coexistence cases. In case of LTE-TDD / 5G-NR synchronised operation, 5G-NR may be negatively
impacted in terms of latency performance of 5G; section 3.3 and ANNEX 3 provide detailed assessments on
these matters.
9
For example: the size of the guard periods between DL / UL transmissions will have an impact on maximum cell radius. Increasing the
number of UL transmissions has an impact on the UL coverage performance.
10
As described in ECC Report 216 section2.1, 8 the frame structures do not need to be exactly identical provided that the last
transmitter stops before the first receiver starts, taking into account the propagation delay (e.g. in LOS non co-sited cases).
11
Two of the Italian operators (Tiscali and Linkem) that acquired spectrum usage rights from the 3400-3600 MHz assignment procedure
in 2007 have agreed on common synchronisation and non-simultaneous UL/DL transmissions by agreeing on a common frame
structure. The following format was chosen to facilitate coexistence between LTE-TDD and the existing WiMAX system: LTE
configuration #2 with Special sub-frame structure #5 3:9:2 - “WiMAX compatible”. At a latter stage, the two operators eventually
agreed to change to a new common format after WiMAX migration to LTE-TDD. With the progressive refarming of WiMAX
technology towards LTE-TDD, operators started their migration towards a different frame structure which is the one that is now more
commonly adopted: LTE configuration #2 with Special sub-frame structure #7 10:2:2.
2.2 UNSYNCHRONISED OPERATION
2.2.1 Definition
The “unsynchronised operation” terminology refers to the general case where neither time synchronisation
between operators’ MFCNs nor inter-operator frame alignment is implemented (of course, this does not
prevent an operator to use the synchronised operation within its own network to avoid co-channel
interferences). More precisely, ECC Report 281 has provided the following definition “the unsynchronised
operation in the context of this Report means operation of TDD in several different networks, where at any
given moment in time at least one network transmits in DL while at least one network transmits in UL. This
might happen if the TDD networks either do not align all UL and DL transmissions or do not synchronise at
the beginning of the frame".
The ECC has defined the restricted baseline out of block power limit for unsynchronised and semi-
synchronised operation of MFCN BSs (see ECC Decision (11)06 (October 2018) Table 4 [3]).
2.2.2 Benefits and challenges of unsynchronised operation
The benefit of unsynchronised operation is in the fact that it does not require the adoption of a compatible
frame structure among operators. Operators can select the most appropriate frame independently and can
adapt the frame structure to service and end user requirements in space and time domains. This allows more
flexibility in the execution of operators’ business models.
However, in a multi-operator scenario, the flexibility in operators’ frame structure selection leads to a number
of interference scenarios that need to be assessed and managed.
As illustrated in Figure 4, BS-BS interference is a result of two separate and independent phenomena.
Spectral leakage from the interfering BS transmitter side:
This is where a BS radiates unwanted emissions into adjacent channels, thereby effectively increasing the
noise-plus-interference floor at a victim BS and resulting in desensitisation. The extent of spectral leakage of
the interfering BS is defined by its adjacent channel leakage ratio (ACLR) and unwanted emission
specifications.
BS-BS interference due to spectral leakage can be mitigated by restricting the unwanted emissions of
unsynchronised BSs through the specification of regulatory block edge masks (BEMs). An example is the
restricted baseline out of block power limit defined in the ECC Decision (11)06 (October 2018)[3]: the
regulatory upper limit of -43 dBm/5MHz on the out-of-block TRP of unsynchronised AAS BSs in the 3400-
3800 MHz band applicable at the frequency boundary (block edge) with another operator. The ECC
restricted baseline out of block power limit is significantly more restrictive than the ECC baseline limits for
synchronised BSs, and compliance with it would require the installation of costly operator-specific transmitter
filters in non-AAS systems. It is even more challenging to achieve in AAS systems where additional internal
filters would be required.
Therefore, based on currently available filtering technology for AAS, unsynchronised operation could be
implemented only in those cases where additional isolation (e.g. separation distances) or specific network
configurations (e.g. indoor low power BSs) would still allow the use of the ECC baseline out of block limit as
12
defined in ECC Decision (11)06 (October 2018) Table 3 . The identification of such specific cases is
addressed in Section 4, which highlights deployment challenges.
12
With reference to the restricted baseline limits defined for AAS base stations, ECC Report 281 [2] states: ”For unsynchronised and
semi-synchronised operations, if no geographic or indoor/outdoor separation is available, the restricted baseline limit must be
respected. However, agreements at national level (including bilateral agreements among any pair of adjacent MNOs) may be
concluded to allow the definition of a different BEM. ”With this respect, ECC Report 281 refers to the possibility to account the
information provided in this toolbox Report.
Blocking of the victim BS receiver:
This is where the victim BS’s receiver is unable to decode a weak wanted signal when simultaneously being
exposed to a relatively high received carrier power radiated by an interfering BS operating in another
channel. The impact would be a desensitisation of the victim BS or, in an extreme case, the complete
overload of the victim BS’s RF front-end. The extent of susceptibility of a victim BS receiver is defined by its
adjacent channel selectivity (ACS) and blocking specifications.
Figure 4: BS-BS interference mechanisms in case of simultaneous UL/DL transmissions
It should be noted that the application of stringent regulatory limits on the interfering BS wanted emissions
alone may not be sufficient to mitigate BS-BS interference with the currently available equipment. This is
because the in-band blocking phenomenon can only be avoided through installation of additional operator-
specific RF receiver filters at the victim BSs receiver to suppress the received adjacent channel carriers. As
such, a regulatory framework for unsynchronised BSs should take into account for the level of the victim BS
receiver selectivity. For the same reason, implementing a guard band within a TDD band does not solve all
interference cases if equipment does not implement operator specific hardware filters in their RF front-end to
protect from in-band blocking. These RF filters would have to be operator specific, which would not be
implementable from an economical or mechanical point-of-view. In addition this approach is totally not
applicable on MS side to solve MS to MS interference
Unsynchronised operation therefore requires all of the operators in a band in the same geographical area /
region to comply with the ECC restricted baseline out of block limit over the frequency blocks of other
operators. Furthermore the addition of inter-operator guard band and operator-specific RF filters on both BSs
transmit and receive sides is required to avoid blocking.
 In case of non-AAS BSs, it is possible to deploy external custom filters specifically designed for each
operator spectrum;
 In case of AAS BSs, as illustrated below the BS RF and antenna units are integrated without an
accessible interface between the RF unit and the antennas. The regulatory requirements would therefore
need to be met by product design and any filters would need to be internal, integrated by the vendor
during the manufacturing process.
At the time of the publication of this Report, AAS systems can neither achieve cost-effectively the restricted
ECC baseline out of block limit defined for unsynchronised (and for semi-synchronised) operation on the
transmitter side, nor implement the required operator-specific filters to protect from blocking on the receiver
side, both in adjacent and non-adjacent channels in the same band.
Figure 5: AAS and non-AAS base stations architecture
It is therefore expected that, based on currently available technology, AAS systems will have generic band
filters aligned with SDOs RF requirements (and with the ECC baseline and transitional regions out of block
limits). The introduction of guard bands alone would not be sufficient to allow unsynchronised operation;
meeting the ECC restricted baseline limit would also require operator-specific filters for AAS BSs which are
not currently seen as cost-effective. Based on the above, unsynchronised operation with AAS BS would
require additional mitigation techniques, which are assessed in section 4.
Unsynchronised operation also leads to MS-MS interference as a result of both spectral leakages from the
interfering MS and blocking of the victim MS. Out of band emissions and adjacent channel requirements for
MS are defined in the relevant harmonised standards for synchronised operation rather than for
unsynchronised operation.
ECC identifies the BS-BS interference scenario as the most critical and, for the interference resulting from
transmitter spectrum leakage, regulates it accordingly. Blocking is taken into account in 3GPP standards in
the case of synchronised operation. This is justified by the fact that MS activity is more intermittent than BSs’,
and by the fact that statistical factors mitigate the criticality of the MS-MS interference mechanism since
devices are typically mobile.
MS-MS interference in the 2.6 GHz band was studied in ECC Report 131 [6], and. ECC concluded that MS-
MS interference was handled through standardisation. Therefore, ECC did not adopt BEMs for terminals.
The situation in the 3400-3800 MHz band is more favourable due to the higher propagation losses, which
further limit MS-MS interference.
2.3 SEMI-SYNCHRONISED OPERATION
2.3.1 Definition
ECC Report 281 provides the following definition: ”the semi-synchronised operation corresponds to the case
where part of the frame is consistent with synchronised operation as described above, while the remaining
portion of the frame is consistent with unsynchronised operation as described above. This requires the
adoption of a frame structure for all TDD networks involved, including slots where the UL/DL direction is not
specified, as well as synchronising the beginning of the frame across all networks”.
A very generic description of semi-synchronised operation is depicted in Figure 6 where Operator A and
Operator B operate in adjacent channels. The operators can designate portions of the frame to have
synchronised fixed duplex direction, i.e. they are always DL or always UL. For the remainder of the slots, the
operators may choose semi-static but different, or time-varying duplex directions.
Fixed DL Flexible DL Flexible UL Fixed UL
Operator A
Fixed DL Flexible DL Fixed UL
Operator B
Figure 6: Example of semi-synchronised operation
Semi-synchronised operation is therefore a mode of operation similar to synchronised operation, with the
exception that the frame structure alignment is relaxed to allow some controlled degree of flexibility at the
expense of some additional interference that can be controlled to some extent. Taking into account the
challenges to implement unsynchronised operation with the ECC restricted baseline out of block power limit
as it was described in the previous Section, semi-synchronised operation aims to find a balance between
more flexibility (compared to synchronised operation) and some acceptable data-loss. The part of the frame
with flexible UL/DL transmissions may suffer from BS-BS and MS-MS interference with respect to both
leakage and blocking interference mechanisms as described in section 2.2.2, therefore the conditions where
semi-synchronised operation will be considered acceptable with regard to the data-loss have to be carefully
discussed and agreed at the national level.
In a specific implementation of semi-synchronised operation, the control plane can be protected by ensuring
that the control signals never belong to the flexible part of the frame. This is different from the case of
unsynchronised operation where both control and data channels can be interfered leading to potentially
larger loss (e.g. inability to decode the whole frame resulting in large throughput degradation).
Semi-synchronised operation between TDD networks requires the following agreements between operators:
 Time synchronisation – as in the case of synchronised operation;
 Partial frame alignment: the agreement shall define a default frame structure for synchronised operation
(for which UL/DL directions are defined across the whole frame) and at the same time the part of the
frame where each operator is allowed to reverse the default transmission direction.
Semi-synchronised operation can also be applied in case of coexistence between different technologies
operating in adjacent frequency blocks if the operators involved agree on a frame structure, which could
contain some flexible portions of the frame. A different degree of flexibility in the assignment of UL/ DL
transmission directions to the different portion of the frame (e.g. in granularity, dynamic vs. static) and in the
ability to protect control channels can be achieved by different features.
2.3.2 Benefits and challenges of semi-synchronised operation
Semi-synchronised operation allows for some degree of frame structure flexibility when compared with
Just like synchronised operation, semi-synchronised operation requires operators to find an agreement with
all other concerned operators in the band and in the same area if they want to deploy without any other
additional coexistence mitigation. An agreement between two operators, potentially using the same
technology, is easier to achieve than an agreement between multiple operators, potentially using different
technologies and potentially targeting different services.
Semi-synchronised operation introduces an upper limit to the BS-BS and MS-MS interference when
compared with unsynchronised operation.
Operators may trade-off between frame flexibility and risk of interference. In some circumstances, semi-
synchronised operation of BSs meeting the ECC baseline out of block limits (defined for the synchronised
operation) will be possible in the same geographical areas without guard bands and operator-specific filters
13
with an increase in lost packets that operators may consider acceptable . The applicability of the ECC
baseline out of block limit is investigated in section 4.5.
Should operators agree to allow semi-synchronised operation based on the ECC baseline out of block power
limit (waiving the requirement for ECC restricted baseline), then the part of the frame with flexible UL/DL
transmissions may suffer from BS-BS and MS-MS interference with respect to both leakage and blocking
interference mechanisms as described in section 2.2.2.
 There are some 5G use cases that imply the deployment of MSs that are in fixed positions and close to
each other (e.g. crowded stadiums, trains, busses, (home) CPEs in fixed wireless access (FWA)
systems, fixed machinery/robots in factories). In such scenarios, the MS-MS interference might not be
negligible anymore: no studies were performed with this respect.
In order to support semi-synchronised operation, BSs may have to implement interference mitigation
techniques. For example, in a scenario where a portion of the DL periods can be used for UL:
 Using zero forcing to create a null in the direction of the interference coming from the neighbour network
operating in the adjacent band although it is currently unclear to what extent such techniques will be
effective and additional implementation costs that still need be determined;
 Limiting the UL transmission to part of the occupied bandwidth far from the edge of the operator block
and using a robust modulation and coding scheme although the extent to which this would reduce
spectral efficiency still needs to be determined.
In terms of market availability, some features needed to support some semi-synchronised operation
scenarios are optional in 3GPP specifications. The latest updates on the status and future plans in 3GPP
(Rel. 15 and Rel. 16) on the unsynchronised and semi-synchronised operating modes are available in
ANNEX 9.
2.4 OPERATING MODES SUMMARY
The following table provides a summary for the options associated with the synchronised, unsynchronised
and semi-synchronised operating modes.
Table 2: Options associated with the synchronised, unsynchronised

and semi-synchronised operating modes
Options Synchronised Semi-synchronised Unsynchronised
Mandatory (with proper SLA)

 A common phase clock reference (e.g. UTC) and
accuracy/performance constraints (e.g. +/- 1.5 μs), either
Common using their own equipment to provide the clock, or
clock across sharing the same phase / time clock infrastructure. Not needed.
networks  Operators shall monitor their clock source and shutdown
their equipment if they lose the reference clock for
duration longer than the holdover period of their
equipment (which depends on the quality of the local
oscillator in the BS).
13
With reference to the restricted baseline limits defined for AAS base stations, ECC Report 281 [2] states ”For unsynchronised and
semi-synchronised operations, if no geographic or indoor/outdoor separation is available, the restricted baseline limit must be
respected. However, agreements at national level (including bilateral agreements among any pair of adjacent MNOs) may be
concluded to allow the definition of a different BEM”. With this respect, ECC Report 281 refers to the possibility to account the
information provided in this toolbox Report.
Options Synchronised Semi-synchronised Unsynchronised
A compatible frame
Compatible structure (including Same as for synchronised operation
frame TDD DL/UL ratio and with the additional possibility for
Not needed, full
structure frame length) in operators to agree on the parts of the
flexibility.
across order to avoid frame when flexible UL/DL
networks simultaneous UL/DL transmissions can occur.
transmissions.
The ECC restricted baseline limit (as defined in ECC Decision

(11)06 (October 2018) Table 4 [3]) applies by default. In this
case:
 The interfering BS transmitter side requires custom filters and
guard band;
 The victim BS receiver side also requires custom filters to
avoid blocking.
Both requirements are difficult to be cost effectively implemented

in AAS BSs with currently available technology.
The baseline out of
block power limit as
However, as described below, this Report indicates those cases
defined in ECC
where the ECC baseline limit could be applied to the semi-
Decision (11)06
synchronised mode and those cases where the ECC baseline
(October 2018) Table
limit can be applied to the unsynchronised operation mode.
3 [3] applies.
Out of block
No additional filtering The ECC baseline out of block power
power limits:
required on top of limit is expected to be applicable to
need for
3GPP RF the semi-synchronised operation in a However, at national
filters, guard
specifications (e.g. wider range of circumstances level, operators, in
bands …
AAS BS may be compared to the case of specific scenarios (e.g.
provided with generic unsynchronised operation. power restrictions,
band filters in their Coexistence between operators can geographic isolation) that
RF frontend). No be managed by agreeing which are in the scope of this
need for any guard portions of the frame may be used for Report - section 4 and in
bands or additional flexibly for UL/DL transmissions. the associated Annexes),
isolation. may mutually agree to
Operators may find the most apply the baseline out of
appropriate balance between block power limit (as
improved frame structure flexibility defined in ECC Decision
and throughput degradation caused (11)06 (October 2018)
by cross-link interference. Table 3) to
unsynchronised
Coexistence studies (see section 4.5 operation under specific
and ANNEX 8) assessed the impacts circumstances.
in case of Macro BS and Micro BS
deployments.
With reference to the synchronised and semi-synchronised operation which require a common clock
synchronisation and the initial agreement among operators on compatible frame structure:
 ANNEX 2 provides an overview on the mainstream technical options to implement network
synchronisation.
 Section 3 and the associated ANNEX 3 provide performance assessments (in terms of UL/DL throughput
spectral efficiency and latency) associated with different 5G-NR frame structures;
 Section 5 provides more information on the operator agreements required at national level to enable the
synchronised operation mode;
3 PERFORMANCE IMPACTS FROM FRAME STRUCTURE SELECTION
One of the main scenarios for the rollout of 5G will be based on the 5G-NR air interface and on Macro BSs
implementing AAS technology in a multi-operator environment. The unsynchronised operation in such
scenario would pose additional challenges compared to the existing TDD uses in the 3500 MHz or 2600 MHz
which:
 Were mostly based on Macro non-AAS BSs (see section 2.2.2), where unsynchronised operation with
guard band and custom filters was feasible cost-effectively;
 Were often involving one operator per geographic region with limited inter-operator synchronisation
issues;
 Were mostly based either on WiMAX or LTE-TDD.
Inter-technology synchronised operation between WiMAX and LTE-TDD has been achieved in a number of
14
cases by adopting the “WiMAX compatible” LTE-TDD frame structure without significant performance loss
for LTE-TDD (see ECC Report 216 section 2.3.2). However, synchronised operation between WiMAX / LTE-
TDD and 5G-NR may imply a cost in terms of performance with regards to 5G latency targets especially.
This Section focusses on the implications associated with the selection of a compatible frame structure in a
5G-NR multi-operator context. There may also be a need to ensure coexistence with LTE-TDD base stations
for some Administrations. This section summarises the results from two performance assessments (see the
full studies in ANNEX 3) between three possible examples of 5G-NR frames (“DDDSU”, “DSDU” and
15
“DDDDDDDSUU” ). Other 5G-NR frame structures can be considered for selection, although the
16
performance has not been assessed in this Report, such as: "DDSU", "DDSUU", "DDDSUUDDDD " or
combinations like "DDDSUDDSUU".
3.1 GENERAL INFORMATION ABOUT 5G-NR FRAME STRUCTURE
Compared to LTE-TDD, 5G-NR allows significantly more flexibility in the frame structure with the ability to
configure uplink / downlink / mixed transmission at the symbol level. This is necessary for some solutions to
fulfil IMT-2020 compliance on URLLC latency.
This section provides a brief description for the 5G-NR frame structures while sections A3.1 and A3.2
include more details on LTE-TDD and 5G-NR frame structures respectively.
5G-NR downlink and uplink transmissions are organised into frames with 10ms duration, each consisting of
ten sub-frames of 1ms duration. The number of consecutive OFDM symbols per sub-frame is given by:
subframe, µ subframe, µ
N symb = N symb
slot
N slot
Each frame is divided into two equally-sized half-frames of five sub-frames each with half-frame 0 consisting
of sub-frames 0 – 4 and half-frame 1 consisting of sub-frames 5 – 9. The UL or DL transmissions are
configured within each slot. With reference to the transmission directions, OFDM symbols in a slot can be
classified as 'downlink' (denoted 'D'), 'flexible' (denoted 'X'), or 'uplink' (denoted 'U') see ANNEX 3.
3.2 PERFORMANCE IMPACTS OF 5G-NR FRAME STRUCTURES
The frame structure selection has an impact on several aspects of network performance, including:
14
LTE frame structure #2 with Special Sub-frame structure #5 3:9:2.
15
5G-NR frame structure compatible with LTE frame structure.
16
5G-NR frame structure compatible with LTE frame structure.
 DL/UL traffic ratio;

 Spectrum utilisation efficiency;
17
 Round-trip time (RTT) latency ;
 Coverage (DL synch. coverage and UL coverage).
 As described below, there are links between the aspects that are listed above. The selection of a certain
frame structure will improve performance in some aspects whilst reducing it in others. The selection of a
certain frame structure therefore aims at reaching the most appropriate performance trade-off for the
specific operator's needs and targets in terms of services to the end users.
The frame structure determines a specific DL/UL ratio: the frame structure selection shall therefore carefully
account for the expected traffic patterns. The DL/UL ratio typically relates to the traffic generated by the
services proposed by the operator and therefore can be linked to the business model of the operator.
The more frequent the DL/UL and UL/DL switching, the lower the RTT is. A short latency improves the
channel estimation quality (CQI feedback) using TDD channel reciprocity properties and also enables fast
HARQ retransmissions. More frequent switching therefore has a positive impact on spectrum efficiency in
high mobility conditions. On the other hand, considering that guard periods (GP) are required at each DL/UL
switch, more frequent switching increases the GP overhead that can have a negative impact on spectrum
efficiency.
The frame structure impacts coverage performance. The guard period (GP) between downlink and uplink
must be large enough to compensate the propagation delay for large cells (and for coexistence with other
cells in line of sight). If a TDD cell can interfere with another cell up to 60km away in co-channel, then this
means that the GP may need to be larger than 200µs.
As illustrated in Figure 7, different frame structures correspond to different trade-offs relatively to key
performance aspects. Operators in different markets will assess the behaviour of the key network
characteristics associated with the different frame structure options in order to decide the most appropriate
frame structure for their own networks and when discussing the options for a compatible frame structure with
other operators. Operators owning other MFCN frequency bands (e.g. 700, 800, 900, 1800 MHz or
mmWave) will have the possibility to use jointly such frequencies with the 3400-3800 MHz band through the
Carrier Aggregation or Supplemental Uplink schemes (CA/SUL). Such combined use will provide additional
ways to meet the target network characteristics. The terminals supporting the CA/SUL schemes will require
to support another band in addition to the C-band.
Carrier Aggregation (CA) is a technique that aggregates various component bands into an overall wider
bandwidth. Supplementary UpLink (SUL) makes it possible to use another frequency carrier for NR UL
transmission instead of NR's dedicated UL carrier in a switchable manner. SUL is similar to CA, however,
unlike CA concept, simultaneous data transmissions are not possible in SUL carrier and NR UL carrier it is
linked to. Additionally, there is no possibility for precise estimation by the UE of the coupling loss needed for
the open loop power control.
17 As defined in Section A3.1.7.

18
Figure 7: Factors for frame structure selection
All of those effects have been studied in order to quantify them, and the detailed studies are in ANNEX 3.
3.3 COMPATIBILITY OF 5G-NR FRAME WITH LTE-TDD FRAME STRUCTURE
With reference to the synchronised operation of 5G-NR BSs and LTE-TDD BSs, noting that every LTE-TDD
frame configuration has at least one compatible 5G-NR equivalent configuration, the 5G-NR TDD pattern
should be based on the following sequence of DL, UL and special slots: "DDDSUUDDDD". Two example
19
variants may be considered:
 Variant 1: LTE-TDD and 5G-NR have an aligned frame start, e.g. "DDDSUUDDDD";
 Variant 2: non-zero frame start offset between LTE-TDD and 5G-NR, e.g. “DDDDDDDSUU”.
These variants, with 30 kHz subcarrier spacing (SCS) can be aligned to LTE-TDD “DSUDD” frame structure
with 15 kHz SCS (LTE-TDD frame configuration #2).
It is to be noted that there should also be a compatible structure for the symbols within the LTE-TDD "S" sub-
frame. For the studies considered in this Report, the “DDDDDDDSUU” frame configuration is used to
represent the performance that 5G-NR would have in case of synchronised operation with a neighbour LTE-
TDD network in the same band and in the same area using LTE-TDD frame configuration #2. Note that
similar results apply in case the non-shifted variant, i.e. "DDDSUUDDDD", is used.
18
The examples in this figure do not refer to the frame structures that are addressed in the studies from this Report.
19
Applicable in case of LTE-TDD configuration #2 frame.
Figure 8: Variant 1 - Synchronised operation of 5G-NR ("DDDSUUDDDD" frame)

and LTE-TDD (“DSUDD” frame)
Figure 9: Variant 2 - Synchronised operation of 5G-NR (“DDDDDDDSUU” frame)

and LTE-TDD (“DSUDD” frame)
The performance assessment of 5G-NR in case of adoption of the “LTE-TDD compatible” frame structure is
provided in the following sections where the results from studies # 1 and # 2 are provided.
3.3.1 Summary from Study #1 (grant-based UL transmissions)
Study #1 provides an assessment in terms of latency and capacity performance for two frame structures
provided, namely DSDU and LTE-TDD compatible frame structure.
The analysis carried out in this study assumes grant-based UL transmissions.
Detailed assumptions and a full set of simulation results are presented in Appendix A.3.3.1; in the following a
summary of latency and capacity comparison between the two analysed frame structure is reported.
Latency assessment
A summary of the latency analysis results is reported in Table 3.

Table 3: Simulation analysis – latency
20
Parameter DSDU LTE-TDD compatible frame structure
# required HARQ processes 4 8
DL HARQ RTT 2-3 ms 5 ms
UL HARQ RTT 2 ms 5 ms
UL scheduling delay 1-2 ms 4.5-9.5 ms
It can be observed that when 5G-NR has a frame structure aligned with LTE-TDD (configuration #2 as in this
specific example), considering scheduling and MS / network processing latency, this frame structure will lead
to L1 latency > 4 ms. As already mentioned, the IMT-2020 eMBB latency requirement is 4 ms and URLLC
latency requirement is 1 ms. Therefore, 5G-NR deployments in 3400-3800 MHz using the LTE-TDD frame
structure would not be able to meet some of the IMT-2020 requirements, including the deployment of
innovative services such as URLLC, unless spectrum in other bands can also be used.
With the assumptions provided in Appendix 3.3.1, the DSDU configuration shows significant benefits over
the LTE-TDD compatible 5G-NR frame structure with respect to HARQ RTT and UL scheduling delay, as
reported in Table 3. The simulations result in more than twice the time (5 ms) that is needed to complete one
HARQ round trip as compared to DSDU (2-3 ms).
Table 3 also shows the improved scheduling delay (1-2 ms) over the time required in the case of LTE-TDD
synchronisation (4.5-9.5 ms). This is achieved by more frequent transmit opportunities for UL Scheduling
Requests (SR) and UL data, and is suitable to multiplex low latency services with existing eMBB traffic.
Capacity assessment
The increased flexibility of the DSDU frame structure also has a direct impact on the overall capacity of the
network. The more frequent UL opportunities can allow a higher spectral efficiency due to the fast channel
feedback. The UL symbols allows MS to send sounding reference signals (SRS) and channel quality
information (CQI) every 1 ms, allowing the BS to have an up-to-date estimate of the channel conditions. A
more accurate channel estimation allows for a more efficient usage of beamforming and better rate control
through more accurate modulation and coding scheme (MCS) selection.
The result is improved cell capacity, as shown for a heavily loaded scenario in in Figure 10. The figure has
been obtained considering an outdoor user with different moving speeds running a full buffer DL traffic
pattern. More frequent opportunities to transmit SRS leads to better spectral efficiency over the PDSCH
symbols in a fast fading channel. Faster sounding allows better tracking of channel fluctuations, thus allowing
improved demodulation performance. Figure 10 compares the simulated spectral efficiency at 5 ms and 1 ms
SRS transmission opportunities. The median and 5%-tile spectral efficiency are shown in Figure 10. It can
clearly be seen that the fast switching of DSDU achieves a better spectral efficiency across all speeds as
compared to LTE-TDD compatible 5G-NR frame structure. While the median gain is 30 to 40%, the gain at
the lower percentile (e.g. cell edge conditions) rises to 70%.
20
The DDDDDDDSUU frame structure was used in the simulations.
Figure 10: Spectral efficiency gains vs. speed
To simulate the effect of the slot structure on user perceived throughput in a realistic scenario, a bursty traffic
pattern (bursty FTP model 3, 0.5 MB file size, variable file arrival time) was simulated. The results are shown
in Figure 11. The shorter DL/UL switching periodicity of DSDU creates more transmission opportunities. The
improved spectral efficiency enables the use of larger transport blocks. With these advantages, the gain of
the median throughput can be as high as 50% (593 Mbps for DSDU vs. 394 Mbps for LTE-TDD compatible
5G-NR frame structure). Even in cell edge conditions, a 23% gain can still be achieved.
Figure 11: Bursty traffic – perceived throughput vs. file arrival rate
3.3.2 Summary from Study #2
Study #2 provides an assessment in terms of latency and capacity performance for the three frame
structures shown in Figure 12.
Among the three frames that are addressed in this study, the DDDDDDDSUU frame structure is the only one
to be LTE-TDD compatible.
DDDSU frame D D D S U
DDDDDDDSUU frame D D D D D D D S U U
DSDU frame D S D U
Figure 12: 5G-NR frame structures for evaluation – slot level
The analysis carried out in this study assumes grant-free UL transmissions for the latency assessment and
grant-based UL transmissions for the assessment of UL and DL capacity and spectral efficiency.
The more complete set of results and detailed assumptions for this study are available in section A3.3.2.The
round-trip time (RTT) for the frame structures is presented in Table 4. Due to the shorter DL/UL switching
period, the DSDU frame structure has a lower RTT than the other frame structures considered.
The RTT associated with DDDSU and LTE-TDD compatible 5G-NR frame structures can be reduced by
using lower frequencies (e.g. 700 MHz, 800 MHz, 900 MHz, 1800 MHz) in combination with the 3400-3800
MHz band (e.g. through Carrier Aggregation (CA) or Supplemental Uplink (SUL) schemes). The resulting
RTT will meet the most stringent latency requirement for URLLC and eMBB simultaneously. It is to be noted,
that licensees for the 3400-3800 MHz frequencies do not necessarily have access to lower frequency bands
and so may not be able to take advantage of CA or SUL to enable URLLC.
In this evaluation, it was assumed that one MS always uses either the SUL for all its UL transmissions, or
always uses the 3400-3800 MHz UL for all its UL transmissions, based on measured RSRP. NR
specifications allow configuring the MS to use both uplinks in a TDM manner, which provides more flexibility
in the operation than simulated here. Taking this into account, the simulation results presented here should
be interpreted as a "best case" scenario., as the impact of the signalling required to assign the lower band
UL to CA/SUL operation has not been taken into account. Accounting for partial resource availability for NR
users in the SUL band (or in the uplink portion of the band in case of CA) would reduce the improvement in
latency to some extent.
The latency assessment results in Table 421 account for the “grant-free” UL transmissions feature (also
known as “configured grant”) which is an optional feature for 5G-NR in 3GPP. The “grant-free” UL is
beneficial for low latency since it avoids the need to first transmit a scheduling request on UL followed by a
scheduling grant on DL before UL data transmission can take place. It is to be noted that decisions to
mandate features for Rel-15 5G-NR MSs were made in consideration of eMBB services, which are first
services that are likely to be delivered using early 5G. Most of the features relating to low latency and/or
reliability are optional. This includes not only the “grant-free” feature, but also other features such as the
mini-slots (frequency control monitoring and short transmission durations), MS processing capability #2
(necessary for the "self-contained" slot operation), dynamic signalling of slot format (see Table 16), etc.
Simulations results with grant-free based UL transmission are valuable in deriving the lowest possible user
plane latency performance.
It is also worth noticing that a comparison between Carrier Aggregation (CA) and SUL has not been
analysed in this report. When CA and SUL are applied to FDD lower frequency bands, at least the same
latency improvement presented for the SUL case can be achieved in case of Carrier Aggregation.
Both CA and SUL have their own merits and drawbacks, which are not addressed in this Report. Carrier
Aggregation (CA) is a technique that aggregates various component bands into an overall wider bandwidth.
Supplementary Up Link (SUL) makes it possible to use another frequency carrier for NR UL transmission
instead of NR's dedicated UL carrier in a switchable manner. SUL is similar to CA, however, one of the main
differences between CA and SUL is the possibility to use the DL carrier in the Carrier Aggregation case.
Therefore with CA there is an increase in both DL and UL throughput compared to SUL scenario. In addition,
21
In case of grant-based transmission, the latency is expected to be higher.
the DL carrier can be used e.g. for DL measurements, power control calculation, mobility management DL
CA requires the UE to have one additional receiver for the downlink.
Table 4: DL and UL latency evaluation results for 5G-NR frame structures (SCS = 30 kHz)
Non-slot RTT (ms) for 5G-NR frame structure (GP: 2 OFDM symbols)
DL/UL based
scheduling DDDSU DDDDDDDSUU DSDU DDDSU+SUL DDDDDDDSUU+SUL
2 OFDM
DL 1.77 3.02 1.12 0.78 0.82
symbols
2 OFDM
UL 1.71 2.95 1.05 0.82 0.86
symbols
3.3.2.1 DL capacity assessment (grant-based UL transmissions)
Grant-based UL transmissions are assumed for this assessment, therefore the latency results presented
in Table 4 (which assume grant-free UL transmissions) do not apply here.
Spectral efficiency with different MS moving speeds and the user-perceived throughput (UPT) with
different arrival rates are evaluated.
It is observed that the DSDU frame structure performance benefits from fast CSI measurement and
feedback, however the frequent DL/UL switching brings about the extra GP overhead.
Figure 13 provides the cell average and cell-edge spectrum efficiency under 10 km/h moving speed can
achieve 15% and 23% gain for DDDDDDSUU vs. DSDU, due to the lower overhead.
Figure 13: DL spectrum efficiency with different speeds

Figure 14 shows the balance between overhead and feedback delay, DDDSU frame structure has the best
performance in most cases and the gain compared to DSDU can be achieved by more than 10%.
Figure 14: DL user-perceived throughput (UPT) with different file arrival rates
3.3.2.2 UL capacity assessment (grant-based UL transmissions)
According to the study performed:

 DDDSU frame structure with the good balance of overhead and delay has better performance in most
cases under different traffic loads. Compared to DSDU, the gain of user average UPT can achieve to
17%;
 If using SUL band, up to 5-times gain can be reached for 5% user throughput since such users benefit
from the lower path loss and the sufficient bandwidth in the SUL band.
4 INTERFERENCE MITIGATION FOR UNSYNCHRONISED OPERATION WITH ECC BASELINE OUT

OF BLOCK POWER LIMIT
As it was mentioned in section 2.2, the ECC has defined the restricted baseline out of block power limit for
unsynchronised and semi-synchronised operation of MFCN BSs (see ECC Decision (11)06 (October 2018)
Table 4 [3]. However, the ECC Decision allows CEPT Administrations to define a "relaxed alternative
“restricted baseline limit” applying to specific implementation cases to ensure a more efficient usage of
spectrum […] depending on national circumstances." ).
22
The practical interference criteria adopted to derive the ECC restricted baseline limit is 5% degradation in
the mean UL throughput of the victim MFCN due to ACLR of interfering BS, with the understanding that
interference is not dominated by the adjacent channel selectivity (ACS) of the victim BS. The limits were
derived from a study (see ANNEX 3 in ECC Report 281) which considered the interfering and victim MFCNs
consisting of Macro BSs in a hexagonal grid (19 sites with three cells each) with an inter-site distance (ISD)
of 500 metres. The study considered a shift of the victim MFCN with respect to the interfering MFCN by 70
metres (representing a conservative, not worst case, scenario) and by 288 metres (the best-case scenario).
The restricted baseline limit was then derived considering the 70 metres shift. Coexistence in case of
uncoordinated collocated sites (e.g. two base stations installed in different corners on the same rooftop and
possibly pointing at each other) would correspond to the worst-case scenario.
Figure 15: Best case and conservative network shifts
The study has assumed the required ACLR to be nominally equal to the required ACIR, with the
understanding that interference is not dominated by the adjacent channel selectivity (ACS) of the victim base
station. Therefore the study does not account for the blocking effect due to BS-BS interference.
The study does not assess the MS-MS interference while it is expected that some 5G use cases will imply
the deployment of MSs that are in fixed positions and close to each other (e.g. crowded stadiums, trains,
busses, (home) CPEs in fixed wireless access (FWA) systems, and fixed machinery/robots in factories). In
such scenarios, the MS-MS interference might not be negligible anymore.
It is worth noting that ECC restricted baseline out of block power limit defined in ECC Report 281 was
derived assuming the specific case of two adjacent operators with misaligned duplex directions for the whole
frame duration which in this Report is referred to as "fully-synchronised". The probability for this specific case
22
The studies have adopted the performance criteria of 5% maximum average UL throughput loss. It is worth noting that, for URLLC
use cases, 5% loss may not be acceptable while the target throughput loss level could be closer to 0%.
23
to happen is low and, therefore, "fully-unsynchronised" is a theoretical worst-case assumption purely for
study purposes.
The implementation challenges, based on currently available technology, associated with the ECC restricted
baseline out of block limit have been described in section 2.2.2.
Interference due to unsynchronised operation can be partly mitigated by adopting the following solutions
individually or in combination:
 Adoption of a guard band and operator-specific filtering between the adjacent spectrum assignments
associated with the interfering network and the victim network;
 Geographic separation between the interfering network and the victim network;
 Alternative network topologies to macro-cellular networks:
 Micro BS networks;
 Indoor BS networks;
 Semi-synchronised operation.
The following sections assess to what extent each of these interference mitigations can improve coexistence
between operators. The main results from coexistence studies are introduced while leaving the full set of
studies to the Annexes of this Report.
24
It is worth noting that the performance criteria in the coexistence studies is maximum 5% throughput loss .
For URLLC use cases 5% loss is not acceptable. For these use cases the relevant throughput loss level
should be closer to 0%.
4.1 GUARD BAND REQUIREMENT FOR UNSYNCHRONISED OPERATION
For non-AAS BSs, according to ECC Report 203 a 5 MHz guard band and operator-specific filtering are
necessary for coexistence between TDD and FDD networks in the 3400-3800 MHz band and it is expected
that a similar guard band and external filtering would be required for unsynchronised non-AAS TDD
networks.
There was no technical analysis on the size of guard band and internal operator-specific filters required for
AAS to meet the ECC restrictive baseline out of block power limit.
With AAS operation, operator-specific RF filters implementation would be very challenging, and the
implementation of a guard band, which would also reduce the spectral efficiency in the band, would not
provide any benefit in practice.
For AAS BSs, ECC Report 281 states that, using current filtering technology, about 20 MHz guard band and
internal filters would be required for AAS to meet the ECC additional baseline out of band power limit to
protect radars below 3400MHz. A similar size of guard band and similar operator-specific internal filters may
be required for AAS to meet the ECC restrictive baseline out of block power limit.
4.2 GEOGRAPHIC SEPARATION OF NETWORKS
This section investigates the coexistence between unsynchronised macro-cellular networks operating in
3400-3800 MHz band.
23
For example, in case of a completely random situation, in which the two adjacent operators are fully uncoordinated and select a
random direction, the probability of Tx/Rx overlapping in adjacent channels is a function of the average DL/UL ratio. For instance,
an average 1:1 DL/UL ratio (i.e. equal DL and UL probability), at a given point in time the probability for the UL slots to be interfered
will be 25%, and the probability for the DL slots to be interfered will be 25%. Nevertheless as slots would not be aligned in case of
unsynchronised use, a given slot of Operator A may overlap in time with two slots of Operator B, so that the number of interfered
slots might be higher.
24
The same throughput loss was assumed to derive the ECC baseline limit in ECC Decision (11) 06 (October 2018).
The objective is to derive the minimum isolation, expressed in terms of separation distance, required
between two unsynchronised networks when all deployed BSs meet the baseline out of block power limits as
defined in ECC Decision (11)06 (October 2018) Table 3 [3].
4.2.1 Proposed methodologies
This section discusses and proposes the methodology when coordinating two unsynchronised TDD macro-
cellular networks at national level.
There are two possible approaches to deal with coexistence between two unsynchronised TDD networks
within a country:
 Method #1: define the minimum required separation distance between the two unsynchronised networks;
 Method #2: define the electric field strength trigger value at the nearest victim BS.
Either of these two approaches can be applied.
With reference to the BS technology options, three possible cases can be considered:
 Non-AAS Network A to non-AAS Network B, which could represent two LTE-TDD FWA networks;
 AAS Network A to non-AAS Network B, which could represent one 5G-NR network and another LTE-
TDD FWA network;
 AAS Network A to AAS Network B, which could represent two 5G-NR networks.
The separation distance can be derived based on different protection thresholds:

 5% network cluster mean UL throughput loss;
 I/N=-6 dB at the nearest victim BS.
4.2.1.1 Method #1: Separation distance calculation
As illustrated in Figure 16 and Figure 17, the separation distance is defined between the two nearest BSs in
network A and network B.
Figure 16: Separation distance between Networks A and B – adjacent channel
Figure 17: Separation distance between Networks A and B – co-channel

If networks A and B are both non-AAS, then the separation distance can be calculated using the protection
ratio of I/N=-6 dB or determined by simulation based on the agreed mean UL throughput loss (e.g. 5%
network cluster mean UL throughput loss) between the two concerned mobile operators.
If either network A or B or both adopt AAS BSs, then the separation distance has to be determined by
simulations based on the agreed mean UL throughput loss (e.g. 5% network cluster mean UL throughput
loss) between the two concerned mobile operators.
4.2.1.2 Method #2: Trigger Values calculation
An alternative approach is to define a trigger value (dBµV/m/5MHz) at the nearest BS receiving antenna or at
3m height above the ground, as shown in Figure 18.
When the trigger value is defined at 3m height above ground, a BS antenna height conversion factor should
be used, the determination of antenna height conversion factor is discussed in ANNEX 5 section A5.2.7.
Figure 18: Electromagnetic field trigger value between Networks A and B
Calculation of trigger values

25
The relation between field strength E (dBµV/m) and power level PR (dBm) can be expressed as :
E = PR + 20 * log10(F) + 77.2 (1)
PR = PTX + G1 – PL (2)
Where:
 F (MHz): frequency;
 PR (dBm): received power level at the receiving BS antenna (before antenna);
 PTX (dB): transmit power before antenna;
 G1 (dB): interfering BS antenna gain including feeder loss in the direction of the receiving antenna;
 PL (dB): path loss at the distance D.
25
The formula is derived from the following relationship (after some units conversions):
𝐸𝐸 2 ∙ 𝑐𝑐 2 ∙ 𝐺𝐺𝑅𝑅
𝑃𝑃𝑟𝑟 (𝑊𝑊) = 𝑤𝑤ℎ𝑒𝑒𝑒𝑒𝑒𝑒 𝑐𝑐 𝑖𝑖𝑖𝑖 𝑡𝑡ℎ𝑒𝑒 𝑙𝑙𝑙𝑙𝑙𝑙ℎ𝑡𝑡 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠, 𝑓𝑓 𝑖𝑖𝑖𝑖 𝐻𝐻𝐻𝐻
480 ∙ 𝜋𝜋 2 ∙ 𝑓𝑓 2
It should be pointed out that the trigger value determination for the case AAS BSs is much more complicated
due to the dynamic behaviour of the AAS antenna pattern.
Figure 19: Relationship between the interferer transmitter power and the electric field strength at the
victim receiver
4.2.2 Summary of the studies
Two studies described in ANNEX 5 have presented the simulation results in terms of separation distance
between two unsynchronised macro-cellular networks. The simulation results are summarised in Table 5 .
Table 5: Summary of the simulation results of separation distance between two macro-cellular
networks
Study # 4
Study # 3
Scenario (5% avg. throughput
(5% avg. throughput loss, loss,ITU-R P.452 [21] 50%
ITU-R P.452 [21] 20% time) time)
AAS-AAS 60 km 50 km
Co-channel
Non-AAS – Non-AAS 50 km
Adjacent AAS-AAS 10.5 km 14 km

Channel Non-AAS – Non-AAS 15 km
The results from the two studies based on 5% network mean UL throughput loss show that:
1 For the co-channel case, the required separation distance is in the order of 50 to 60 km;
2 For the adjacent channel case, the required separation distance is in the order of 10 to 15 km.
For the non-AAS to non-AAS co-channel case, the calculated separation distance based on I/N=-6 dB for
non-AAS to non-AAS case is 44 km.
It is to be noted that Study #4 was based on the SEAMCAT software which at the time of publication of this
Report was using a beta version for the module associated with the AAS system. Future improvements to
this specific SEAMCAT module might lead to different results.
4.2.3 Conclusions from studies
The analysis and the simulations that were carried out in this Section lead to the following conclusions:
 The two methodologies are described in this Section, either based on the separation distance or on the
electric field trigger value, can be applied;
 Specific separation distance values can be defined at national level based on the specific circumstances.
This is because the required separation distance and electric field trigger values calculation depend on
many factors:
 Cellular network technology and topology (LTE-TDD or 5G-NR, non-AAS or AAS, BS antenna height,
environment, cell range);
 Propagation environment and propagation model used in relation with this environment;
 Frequencies and overlap of the blocks, e.g. full overlap for co-channel case, or partial overlap (e.g. in
some cases of coexistence between operators in adjacent areas), or adjacent channel);
 Protection ratio, e.g. I/N, or nearest cell throughput loss at x%, or network mean throughput loss at
y%, etc.
4.3 COEXISTENCE BETWEEN UNSYNCHRONISED MICRO BS AND MACRO BS DEPLOYMENTS
4.3.1 Coexistence between unsynchronised Micro BSs and Macro BSs – Study #5
In Annex A1.1.1, the interference between Micro and Macro BSs is studied. For the macro-cellular network
the BSs have an output power (TRP) of 51 dBm and 500 m ISD while the BS in the Micro BS network has an
output power (TRP) of 40 dBm and an ISD of 166 m. The impact on both types of BSs is studied. The
distance between aggressor and victim BS varies, but for the closest pair the distance is 30 m.
The propagation between BSs is modelled using the Urban Macro-cellular (UMa) model, and this model has
a random component. Performance of one specific realisation of the BS-BS propagation is studied. This is
the best way to model the situation in practical deployments since the BS-BS propagation will not vary over
time. According to the study, in order to limit the throughput loss to maximum 5% the required ACIR between
the networks has to be around 60 dB to protect the Micro BS network and 45 dB to protect the Macro-cellular
network for typical deployments. For the most sensitive pair of BS, the ones with 30m separation, the ACIR
has to be between 50 dB and 70 dB to protect the Micro BS network and between 45 dB and 60 dB to
protect the macro-cellular network.
Considering that the ECC baseline gives an ACIR of slightly less than 45 dB, it can be concluded that there
are a few cases, i.e. deployment scenarios, where standard equipment will result in less than 5% throughput
loss, but in the majority of cases the losses are larger. In these scenarios, synchronisation will be an effective
interference mitigation technique.
4.3.2 Coexistence between unsynchronised Micro BSs and Macro BSs – Study #6
This Section provides the main conclusions from the study in A6.2, which considers the impact of BS-BS
interference between MFCNs with simultaneous UL/DL transmission in terms of the resulting degradation in
the mean UL throughput of the victim MFCN. The MFCNs consist of Macro BSs and Micro BSs.
The study addresses two scenarios according to the specific class of base stations, namely:
 Macro-cellular network (hexagonal grid of outdoor stations) is operating as the interferer and the Micro
BS network (hexagonal grid of outdoor stations) is interfered;
 Micro BS network (hexagonal grid of outdoor stations) is operating as the interferer and the Macro BS
(placed outdoors) is interfered;
 Interference from one Micro BS to another Micro BS (both base stations are placed outdoors).
4.3.2.1 Network topologies and main assumptions
The two interfering deployments operate in the same geographic area on adjacent frequency channels.
All Micro and Macro BSs are assumed to be AAS base stations forming a beam towards a MS (MSs are
assumed to be uniformly distributed within a cell).
The Macro BSs have 25 m high antennas and comprise three sectors per site; the Micro BSs are placed 6 m
above ground, comprising one sector per site with random boresight.
Base stations are assumed to be "fully-unsynchronised".
See Section A6.2.3 for the full list of assumptions and parameters.
a) Macro BSs network as interferer; Micro BSs network as victim
Figure 20 provides the topology used for the coexistence studies in case of a macro-cellular network
(hexagonal grid placed outdoors) operating as the interferer towards a Micro BS network (hexagonal grid
placed outdoors).
Figure 20: Topology for the Macro BS to Micro BS interference scenario
b) Micro BSs network as interferer; Macro BSs network as victim
Figure 21 provides the topology used to support coexistence studies in case of Micro BS network (hexagonal
grid placed outdoors) operating as the interferer towards the Macro-cellular network (hexagonal grid placed
26
outdoors). In line with ECC Report 203 [5] , the simulations address one Macro BS, which is completely
surrounded by the Micro BS network grid.
26
ECC Report 203 page 26: “One important thing to note here is that the results contained in Table 17 are for one reference cell in the
Macro-cellular network, which is overlapped completely by the Micro BS network (Manhattan) grid (see Figure 19).”
Figure 21: Topology for the Micro BS to Macro BS interference scenario
c) Micro BSs to Micro BSs
The two approaches which have been used to assess the interference from AAS Micro BSs to AAS Micro
BSs are described below.
Approach 1: in this analysis the separation distance between the Micro BSs is an input parameter, the Urban
Micro-cellular (UMi) path loss model determines the associated Line-of-Sight (LoS) probability.
The following two settings have been considered:
 Case 1a: 30m separation distance between the two Micro BSs leading to 80% LoS probability based on
the UMi path loss model (the smaller the distance, the greater the probability the two Micro BSs will be
along the same street).
Figure 22: Micro BS to Micro BS interference scenario topology - Case 1a:

30m separation distance leading to 80% LoS probability based on UMi path loss model
 Case 1b: 100m separation distance between the two Micro BSs leading to 25% LoS probability based on
the UMi path loss model (the larger the distance, the greater the probability the two Micro BSs will be
located in different streets).
Figure 23: Micro BS to Micro BS interference scenario topology - Case 1b:

Approach 2: in this analysis, the separation distance between the Micro BSs is an input parameter as well as
the LoS probability.
This approach accounts for the fact that it is difficult to carry out meaningful simulations to assess the
interference between two Micro BS networks in the same urban area since the interference scenario will be
strongly impacted by the LoS/NLoS conditions, which radically change depending on where the Micro BS are
installed with respect to each other in built-up areas.
The study therefore considers two specific set of cases for the deployments of the interfering and victim base
stations:
 Cases 2a, 2b and 2c: two Micro BSs located in different streets at 30m, 50m and 75m separation
distance with 0% LoS probability.
Figure 24: Micro BS to Micro BS interference scenario topology. Case 2a, 2b and 2c:
30, 50, 70 m separation distance and 0% LoS probability (different streets)
Case 2d: two Micro BSs located along the same street (100% LoS probability) at 100 m separation distance.
Figure 25: Micro BS to Micro BS interference scenario topology. Case 2d: 100 m separation distance
and 100% LoS probability (same street)
4.3.2.2 Simulations results and conclusions
This section presents the simulations results expressed in terms of degradation of the mean uplink
throughput of the victim MFCN due to base station to base station interference from the interfering MFCN,
presented as a function of ACIR. In general terms, as expected, the impact of interference on network
performance diminishes with increasing values of ACIR.
Note that the required ACLR is assumed to be nominally equal to the required ACIR, with the understanding
that interference is not dominated by the adjacent channel selectivity (ACS) of the victim base station.
Therefore the study did not take into account the BS blocking effects on the victim BS receiver.
Note that both victim BS and interferer base stations are assumed to operate with 60 MHz channel
bandwidth.
It is important to highlight that the study did not account for MS-MS interference.
Interference from AAS macro-cellular network into AAS Micro BS network:
With reference to the topology proposed in Figure 20, an ACIR greater than 68dB is required to ensure a
mean uplink throughput degradation smaller than 5%.
Interference from AAS Micro BS network into AAS Macro BS network:
With reference to the topology proposed in Figure 21, an ACIR greater than 55dB is required to ensure mean
uplink throughput degradation smaller than 5%.
Interference between two AAS Micro BSs:
With reference to the topology proposed in Figure 22 (Case 1a: 30m separation distance leading to 80% LoS
probability based on UMi path loss model), an ACIR greater than 63dB is required to ensure a mean uplink
throughput degradation smaller than 5%.
With reference to the topology proposed in Figure 23 (Case 1b: 100m separation distance leading to 25%
LoS probability based on UMi path loss model), an ACIR greater than 54dB is required to ensure a mean
uplink throughput degradation smaller than 5%.
With reference to the topology proposed in Figure 24 (Case 2a, 2b and 2c: 30, 50, 70 m separation distance
and 0% LoS probability (different streets), shows how an ACIR greater than 49dB is required to ensure a
mean uplink throughput degradation smaller than 5% for 30m separation distance. If the separation distance
is 50 m, 45 dB ACIR is required to ensure mean uplink throughput degradation smaller than 5%.
With reference to the topology proposed in Figure 25 (Case 2d: 100m separation distance and 100% LoS
probability (same street). An ACIR greater than 70 dB is required to ensure a mean uplink throughput
degradation smaller than 5%.
4.4 COEXISTENCE BETWEEN UNSYNCHRONISED INDOOR BSS AND MACRO BS - STUDY #7
In ANNEX 7, the impact on an indoor system from a macro-cellular network is studied. The indoor BS
network is located in a 50x120m large building, which is located 70m from the Macro BS. For the macro-
cellular network the BS have an output power (TRP) of 51 dBm and 500m ISD while the indoor BS is ceiling
mounted with an output power (TRP) of 24 dBm.
The propagation between BS is modelled using the UMa model and this model has a random component.
We study performance of one specific realisation of the BS-BS propagation. This is the best way to model
the situation in practical deployments since the BS-BS propagation will not vary over time.
Based on the study, in order to limit the mean UL throughput degradation for the indoor network to maximum
5% the ACIR (adjacent channel interference ratio) between the networks has to be in the range 25 dB to 65
dB, depending on the actual channel realisation between the Macro BS and indoor BS.
The ECC baseline gives an ACIR of slightly less than 45 dB. From this it can be concluded that in some
cases standard equipment will result in less than 5% throughput loss and in other cases the losses are
larger. This indicates that, for this type of scenario, unsynchronised operation may be possible for carefully
installed indoor BS, but synchronisation may be required for BS installed in shallow indoor locations...
While the 5% mean UL throughput degradation applies to the eMBB use case, it is worth noting that the UL
mean throughput closer to 0%. According to the results above, the URLLC use cases will require an
additional isolation somewhere in the order of 20-25 dB compared to the 5% loss results. Such additional
isolation could be obtained with accurate indoor BS planning and, for example, with adoption of proper
shielding of the building.
4.5 STUDY ON SEMI-SYNCHRONISATION FOR MICROCELL AND MACRO-CELL CASES - STUDY

#8
The benefits and challenges associated with semi-synchronised operation have been discussed in section
2.3. In the following sub-sections, adjacent channel simulation results are presented. These results are
based on the study available in ANNEX 8.
Table 6: Study on semi-synchronisation for Micro-cell and Macro-cell cases – summary of results
Macro-cellular network to Micro BS network to

Scenario
Macro-cellular network Micro BS network
Minimum
distance 288 m 96 m
among networks
BS-BS
propagation Free space path loss 3GPP TR 38.901 – Umi
model
Semi- 10% 50% 100% 100%

20% 10% 20% 50%
synchronisation (fully (fully
unsynch. unsynch. unsynch. unsynch. unsynch. unsynch.
percentage unsynch.) unsynch.)
ACIR needed for

5% mean UL
54 dB 62 dB 70 dB 75 dB N/A 40 dB 47 dB 54 dB
throughput
degradation
As shown in Figure 15, the 288 m network shift assumption between macro-cellular networks represents the
best case. Similarly, the 96 m network shift assumption between micro-cellular networks represents the best
27
case between the two analysed assumptions .
Differently from the approach followed in ANNEX 8, the recommended approach is to use the separation
distance and the line-of-sight probability as input parameter during the coexistence studies for the macro-
cellular network and the Micro BSs network cases. This approach accounts for the fact that it is difficult to
carry out meaningful simulations to assess the interference between two Micro BS networks in the same
27 According the studies' assumptions reported in Table 33:

-Network shift (Macro BS ↔ Macro BS case): Baseline: 70 m
Additional for reference: 288 m
- Min. separation distance between Macro BS and Micro BS: 30 m
urban area since the interference scenario will be strongly impacted by the LoS/NLoS conditions which
radically change depending on where the Micro BS are installed with respect to each other in built-up areas.
Coexistence between the macro-cellular network and the Micro BS network was not assessed by this study.
It is worth noting that the study assumes that operators do not always decide to modify UL symbols / slots
into DL symbols / slots in the flexible part of the frame. This reflects a real deployment scenario where:
 The two operators adopt the same default frame structure;
 When the default frame structure is not modified (in its flexible portion), the network is actually operating
in synchronised mode;
 An operator might decide to modify the agreed default frame structure in some specific locations (hot
spots) and at specific point in time (specific event or busy hour, for instance). In this particular case, only
28
the base stations in these areas and at these times will be subject to cross-link interference .
Taking this into account the results presented in this section represent a worst-case scenario in terms of
throughput degradation resulting from the semi-synchronised case.
"DL to UL modifications": the default DL transmission direction in the flexible part is modified into
UL
In this case, from BS-BS interference perspective, the network that modifies the default DL transmission
direction into UL will not interfere with the other network while it will receive additional interference from the
other network during the period of the modified transmission direction.
In most circumstances, MS-MS interference will be negligible because terminals typically transmit
intermittently and many will be mobile so any interference would be transient
"UL to DL modifications": the default UL transmission direction in the flexible part is modified into
DL
In this case, from BS-BS interference perspective, the network that modifies the default UL transmission
direction into DL will interfere with the other network while it will not receive additional interference from the
other network.
Under the specific assumptions and methodology used in this study, it can be concluded:
 For macro-cellular network to macro-cellular network results show that the throughput degradation is
~9% when operators are unsynchronised (UL to DL modification) for 10% of the frame (i.e. the flexible
part). The modelling considered that the interfering operator always changes the transmission direction
during the flexible part of the frame, this represents worst-case assumptions. This means that the
throughput degradation will likely be lower in a realistic scenario where the interfering operator will not
always modify the transmission direction of the flexible portion of the frame. 288 m BS - BS separation
distance is assumed;
 From Micro BS to Micro BS interference perspective, it is possible to use the ECC baseline out of block
power limit for synchronised operation as specified if the operators have simultaneous UL/DL
transmissions for at most 20% of the frame(based on acceptable Loss 5% and ACIR 45 dB); for a BS -
BS separation distance at 96 m; No conclusion can be derived for the Macro-cellular network to Micro
BS network case since this scenario was not studied. In the case of Micro BS network to Macro-cellular
network, due to the lower power of the interfering BS it is expected that interference from the Macro base
station will dominate the coexistence analysis.-.
28
For example: assuming a configuration in which the flexible "X" slots represent 20% of the entire frame and are used for DL, the
actual percentage of time with cross DL to UL interference will be lower. Even assuming that the operator will use his flexible part in
DL for 50% of the time and for all gNBs, the actual percentage of slots affected by cross interference will be 10% if and only if the
other operator always switches the UL in the same 50% of time. It follows that the actual cross-link interference will be even lower
than 10%. On top of this, not all the gNBs in the network will need to change the baseline configuration and as a consequence the
cumulative interference will be strongly reduced
Based on the above, since UL to DL flexibility creates additional BS-BS interference to the neighbour
operator, the specific cases in which the semi-synchronised operation (for UL to DL flexibility) could be
allowed require agreement at national level.
4.6 SUMMARY OF ALL STUDIES PERFORMED
The following tables collect the results from the studies performed.
Table 7: Coexistence between unsynchronised/semi-synchronised macro-cellular network and

macro-cellular network – summary of results
Required ACIR (dB)

needed for 5% mean UL t-put degradation
Interference scenario
ECC Report 281 Annex
Study #8
3
AAS to
Unsynch. N/A 83
non-AAS
70 m
Unsynch. N/A 77
shift AAS to
AAS Semi-
N/A N/A
synch.
AAS to
Unsynch. N/A 79
Adjacent non-AAS
channel Unsynch. 75 74
288 m
shift 50%
70 N/A
(best unsynch.
case) AAS to
(*) AAS 20%
63 N/A
unsynch
MACRO 10%
BS 55 N/A
unsynch.
→
MACRO Geographic separation distance (km)
BS needed for 5% mean UL t-put degradation
Study #3 Study #4
AAS to
Unsynch. 10.5 km N/A
AAS
AAS to
Unsynch. 14 km N/A
Adjacent non-AAS
channel
31 km
non-AAS
(5% t-put loss)
to non- Unsynch. N/A
3GPP 12 km
AAS
SEM (50% t-put loss)
(45 dB
ACIR) AAS to
Co- AAS
Unsynch. 60 km N/A
channel AAS to
non-AAS
58 km
non-AAS
(5% t-put loss)
to non- Unsynch. N/A
49 km
AAS
(50% t-put loss)
(*) Best case. Agreed assumption for the network shift (Macro BS ↔ Macro BS case): baseline: 70m, additional for reference: 288 m (see
Table 33)
Table 8: Coexistence between unsynchronised/semi synchronised macro-cellular network and Micro

BS network – summary of results
Required ACIR (dB)

needed for 5% mean UL t-put
degradation
Stud
Study #5 Study #6
y #8
63 (58 to 65)
Micro BS in the
Unsynch. middle: N/A 68
AAS to AAS 50 - 70 (worst
Adjacent
case)
MACRO channel
BS Semi-synch. N/A N/A N/A
30 m
→
separation AAS to non- Unsynch. &
MICRO N/A N/A N/A
BS AAS Semi-synch.
Unsynch. &
AAS to AAS N/A N/A N/A
co- Semi-synch.
channel AAS to non- Unsynch. &
N/A N/A N/A
AAS Semi-synch.
43 (40 - 50)
Macro BS in the
AAS to AAS Unsynch. middle: N/A 55
Adjacent 43 - 50 (worst
MICRO channel case)
BS
30 m AAS to non- Unsynch. &
→ N/A N/A N/A
separation AAS Semi-synch.
MACRO
BS Unsynch. &
AAS to AAS N/A N/A N/A
co- Semi-synch.
channel AAS to non- Unsynch. &
N/A N/A N/A
AAS Semi-synch.
Unsynch. N/A 54 N/A
96 m 50%
Adjacent N/A 47 N/A
separation AAS to AAS unsynch.
channel
(**), (***)
20%
MICRO N/A 40 N/A
unsynch.
BS
→
MICRO Case 1a
BS 30 m
separation, Adjacent
AAS to AAS Unsynch. N/A N/A 63
80% LoS channel
prob.
(***)
Case 1b
100 m
Adjacent
separation, AAS to AAS Unsynch. N/A N/A 54
channel
25% LoS
prob. (***)
49 (30m)
Unsynch. N/A N/A 45 (50m)
AAS to AAS <40 (70m)
Case 2a, 2b, Adjacent
2c channel Semi-synch. N/A N/A N/A
30, 50, 70 m
separation, AAS to non- Unsynch. &
N/A N/A N/A
0% LoS AAS Semi-synch.
prob. AAS to AAS
co- Unsynch. &
& AAS to N/A N/A N/A
channel. Semi-synch.
non-AAS
Unsynch. N/A N/A 70
AAS to AAS
Case 2d Adjacent Semi-synch. N/A N/A N/A
100 m channel
AAS to non- Unsynch. &
separation, N/A N/A N/A
AAS Semi-synch.
100% LoS
prob. AAS to AAS
co- Unsynch. &
& AAS to N/A N/A N/A
channel Semi-synch.
non-AAS
(**) Best case. Agreed assumption for the min. separation distance between Macro BS and Micro BS: 30m (see Table 33).
(***) Differently for this study, it was agreed to use the separation distance and the line-of-sight probability as input parameter during the
coexistence studies between the Macro-cellular network and the Micro BSs network. This approach accounts for the fact that it is
difficult to carry out meaningful simulations to assess the interference between two Micro BS networks in the same urban area since
the interference scenario will be strongly impacted by the LoS/NLoS conditions which radically change as the Micro BSs change
their locations with respect to buildings.
Table 9: Coexistence between unsynchronised macro-cellular network and indoor BS network –

summary of results
Required ACIR (dB)

needed for 5% mean UL t-put degradation
Study #7
Building short side facing BS: 43 (34 to 57)
Unsynch
AAS to Building long side facing BS: 53 (23 to 63)
AAS Semi-
NA
Adjacent synch
MACRO channel.
BS AAS to Unsynch &
→ non Semi- NA
INDOOR AAS synch
BS
AAS to
AAS & Unsynch &
Co-
AAS to Semi- NA
channel.
non- synch
AAS
"not simulated. However we can observe that
70m the indoor BS has lower output power, which
separation Unsynch means that we should see lower impact from
btw. (qualitative) the indoor. On the other hand if there are
MACRO AAS to (****) several buildings with indoor systems
BS AAS deployed there is a need to consider the
and effect of the aggregate interference."
building Adjacent
INDOOR wall channel. Semi-
NA
BS CEILING synch
→ MOUNT
MACRO non- Unsynch &
BS AAS to Semi- NA
AAS synch
AAS to
AAS & Unsynch &
Co-
non- Semi- NA
channel.
AAS to synch
AAS
INDOOR AAS to
Adjacent
BS AAS & Unsynch &
channel.
↔ AAS to Semi- NA
& co -
INDOOR non- synch
ch.
BS AAS
(****) Qualitative assessment was performed for this scenario.

5 TOOLBOX: OPTIONS FOR ADMINISTRATIONS TO SUPPORT THE DESIGN OF

SYNCHRONISATION FRAMEWORKS AT NATIONAL LEVEL
5.1 KEY ASPECTS TO BE CONSIDERED WHEN SETTING UP THE SYNCHRONISATION

FRAMEWORK AT NATIONAL LEVEL
Synchronisation and semi-synchronisation are effective for avoiding (in the case of synchronised) or
minimising (in the case of semi-synchronised) cross-link interference between operators. Additional
interference mitigation techniques may be required (including separation distances, alternative network
topologies, etc.) if operators intend to use unsynchronised AAS BS. Synchronised operation is accompanied
with some challenges related to the selection of common clock and frame structure. Such challenges are
explained in more detail in Section 5.2.1.
At the time of this writing, new 5G AAS systems cannot cost-effectively implement the operator-specific
filtering which would be required to meet the ECC restricted baseline out of block power limits and protect
the receiver from blocking from adjacent and non-adjacent channels in the same band. Based on currently
available AAS BS technology, it is assumed that equipment will only implement filters designed to comply
29
with the ECC baseline out of band power limits .
If interference mitigation due to unsynchronised operation relies on separation distances, the minimum
distances required will depend on network topology, terrain and clutter and will need to be discussed at the
national level. The results from the coexistence studies summarised in Section 4 of this Report show that the
separation distances required between unsynchronised Macro cells could be up to 60 km when co-channel
and up to 14 km when operating in the adjacent channel.
Semi-synchronised operation is similar to synchronised operation, with the exception that simultaneous U
/DL transmissions between networks can be allowed in some defined parts of the frame. This leads to a
degree of flexibility at the expense of some additional interference that can be controlled to some extent.
Compared to unsynchronised operation, semi-synchronised operation reduces the impact from BS-BS and
MS-MS interference. Results from studies in Section 4.5 show that the ECC baseline out of block power limit
can be applied to the semi-synchronised operation in specific circumstances. The interference impact on
network performance associated with semi-synchronised operation is reduced when interference on the
control channels is avoided (e.g. where possible, the flexible portions of the frame do not include control
plane channels). As in the case of synchronised operation, semi-synchronised networks will need a common
accurate phase / time synchronisation and an agreement on a compatible frame structure which identifies
the portions of the frame where transmission direction is flexible.
5.2 ENABLING SYNCHRONISED, UNSYNCHRONISED AND SEMI-SYNCHRONISED OPERATION AT

NATIONAL LEVEL BASED ON MULTILATERAL AGREEMENTS AMONG MFCN LICENSEES
5.2.1 Synchronised operation
Synchronised operation avoids any BS-BS and MS-MS interference therefore allowing coexistence between
adjacent networks without the need for guard bands or additional filters. This operating mode therefore
simplifies network deployment because no additional interference mitigation is required. Synchronised
operation leads to the selection of a compatible frame structure, which determines a specific DL/UL
transmission ratio and contributes to the network performance (e.g. latency, spectral efficiency, throughput
and coverage).
29
ECC baseline out of block power limit is defined in ECC Decision (11)06 (October 2018) Table 3 [3] with reference to the
The out of block power limit associated with the synchronised operation mode was defined in ECC Decision
(11)06 (October 2018) Table 3.
For synchronised operation the following issues should be agreed at national level with a general
30 31
framework involving all MFCN licensees in the band and in the same geographic area . In some cases,
Administrations may get involved in order to reach multilateral agreements in a fair and timely manner:
 A common phase clock reference (e.g. UTC) and accuracy/performance constraints that depend on the
underlining technology (e.g. +/- 1.5 μs for LTE-TDD and 5G-NR), either using their own equipment to
provide the clock, or sharing the same phase / time clock infrastructure. Permanent monitoring of the
agreed clock source is needed. When losing the primary reference time clock (PRTC) equipment may
continue operation for some time ("holdover period") that has to be agreed and which depends on the
quality of the local oscillator in the BS and on the wireless network accuracy requirement. If the PRTC is
lost for a period that is longer than the holdover period, the system shall no longer be considered in
synchronised operation and may cause interference to other operators. Proper action shall therefore be
taken (e.g. the BS shall be shut down until the PRTC is recovered);
 A compatible frame structure to avoid simultaneous UL/DL transmissions (guard periods between DL
and UL transmissions may be different, as illustrated in Figure 3;Figure 3);
 The conditions when synchronisation must apply and/or may not be required (when additional isolation is
available and in case of low power indoor BSs);
 Periodic review of the agreed conditions may be needed to account for possible market and technology
developments (e.g. introducing new technologies, adjusting to new needs in the DL/UL ratio or latency,
etc.).
Synchronised operation between 5G-NR and LTE-TDD/WiMAX systems could imply a cost in term of user
plane latency and throughput performance; the summary from the detailed assessments in is provided in
Section 3.3.2. Operators may have the option to reduce the user plane latency and RTT, under some
circumstances, by using lower frequencies (e.g. 700, 800, 900, 1800 MHz) in combination with the 3400-
3800 MHz band (e.g. through Carrier Aggregation or Supplemental Uplink schemes). Some licensees may
not have access to additional spectrum in lower frequency bands with available capacity (e.g. verticals and
some MNOs) and the user terminals supporting these functionalities may not be available in short term.
5.2.2 Unsynchronised operation based on the ECC baseline out of block power limit
Licensees can select the most appropriate frame structure independently and can adapt the frame structure
to service and end user requirements, which may change depending on the location and on time.
However, in a multi-operator scenario, the flexibility in operators’ frame structure selection leads to a number
of interference scenarios that need to be assessed and managed.
The out of block power limit for unsynchronised and semi-synchronised operation is defined in ECC Decision
(11)06 (October ) Table 4, the "Restricted baseline". However, the ECC Decision allows CEPT
Administrations to define a "relaxed alternative “restricted baseline limit” applying to specific implementation
cases to ensure a more efficient usage of spectrum […] depending on national circumstances.") .
The simulations defined in this report assess the feasibility of unsynchronised operation when using the ECC
32
baseline limit and not the restricted baseline limit because of the practical difficulties in achieving the
restricted baseline limit as discussed earlier . The assumptions in the studies are consistent with the
30
As explained in section 2.2, operators’ agreements must be multilateral, involving all operators sharing a band, because the blocking
effect can happen within the whole band regardless of any frequency separation within that band and is not restricted to the
adjacent channel.
31
“Same geographic area” refers to an area within which two networks can be impacted by mutual interference in case of simultaneous
32
ECC Decision (11)06 (October -2018) Table 3
definition and assumptions in ECC Report 281 [2], including the simultaneous UL/DL transmissions across
the whole frame ("fully-unsynchronised" transmission scenario).
System level simulations do not account for the additional potential data loss that would result from interfered
33
control channels (e.g. inability to decode the whole frame resulting in larger throughput degradation) . Link-
level simulations would be required for a more accurate analysis.
Unsynchronised operation could be allowed at national level in a limited number of specific cases where
sufficient isolation between interferer and victim base stations exists. The associated parameters should be
34
agreed at national level with multilateral agreements among all MFCN licensees in the same geographic
area35 in the band in a fair and timely manner. Such agreements could account for the following options.
5.2.2.1 Options for enabling the unsynchronised operation involving macro-cellular networks
A specific recommendation for the separation distance or a single set of trigger values between
36
unsynchronised macro-cellular networks cannot be provided (due to the dependency on various factors ).
Section 4.2.1 provides the methodology to support Administrations and MFCN licensees in deriving specific
values for separation distances and/or trigger values at national level. MFCN licensees need those values to
establish an agreement when their networks are not fully or semi-partially synchronised.
The results from the coexistence studies summarised in section 4 of this Report show that those distances
37
could be up to 60 km when co-channel and up to 14 km when operating in the adjacent channel . Those
separation distances are based on the analysis detailed in ANNEX 6 for a flat terrain environment. Smaller
distances may be achieved in a different environment and/or with proper mitigation techniques e.g. with
some coordination on the azimuth/down tilt, etc.
In case of coordination within national borders, different coordination parameters may be defined (leading to
different separation distances) compared to the case of international cross border coordination. While the
specific coordination parameters will need to be agreed at national level, the international coordination
approaches defined in Rec. ECC Recommendation (15)01 [4] may be used as a reference to deal with the
case of two unsynchronised macro-cellular networks within a given country when the physical borderline is
defined between two networks.
5.2.2.2 Options for the unsynchronised operation involving Micro BS networks in the same area:
The studies are summarised in section 4.3 and reported in ANNEX 6. Simulation results have shown that, in
general, unsynchronised operation of Micro BSs in case of ECC baseline out of block power limit in the same
geographic area might not be feasible.
33
Simulations have mapped SINR to throughput, which allows accounting for interfered Physical Uplink Shared Channel (PUSCH).
When determining the average UL throughput loss, simulations have not accounted for interference on control channels such as the
Physical Uplink Control Channel (PUCCH), which would have impacts on ACK/NACK transmissions (such mechanism cannot be
covered by the SINR – throughput curve. Simulations, in fact, accounted for cell-to-cell interference).
34
As explained in section 2.2, operators’ agreements must be multilateral, involving all operators sharing a band, because the blocking
effect can happen within the whole band regardless of any frequency separation within that band and is not restricted to the
adjacent channel.
35
“Same geographic area” refers to an area within which two networks can be impacted by mutual interference in case of simultaneous
36
Network technologies and topologies (LTE / 5G-NR, non-AAS / AAS BS, BS antenna height), propagation environment and
propagation model, frequency assignments, protection criteria (I/N or network throughput loss at x%, etc.….).
37
It should be noted that ITU-R M.2374 [7] has performed a study between adjacent channel unsynchronised LTE systems in the 2.3
GHz band, with the conclusion that « without any additional RF improvement, one BS could influence unsynchronised BSs
operating in adjacent spectrum block in an area with a radius of. 2.4 to 5.3 km depending on the propagation environment », which
illustrates that input hypothesis such as the propagation model are of significant importance.
There could be very specific circumstances where two Micro BSs could coexist when using the ECC
baseline out of block power limit. For example, when the adjacent channel Micro BSs are not in line of sight
(i.e. 100% NLoS). These Micro BSs might still face coexistence issues with the macro-cellular network
coverage layer above them because they are likely to be in LoS of Macrocells and Macrocells are higher
power.
5.2.2.3 Options for the unsynchronised operation involving indoor BS:
Studies summarised in Section 4.4 and reported in ANNEX 7 investigated unsynchronised operation of
indoor BS and Macro-cellular network in the same area. Simulation results have shown that (under specific
assumptions, in the adjacent channel case) in order to limit the mean UL throughput degradation for the
indoor BS network to maximum 5%, the ACIR (adjacent channel interference ratio) between the networks
has to be in the range 25 to 65 dB, depending on the actual channel realisation between the Macro BS and
indoor BS.
38
Based on the above, the unsynchronised operation of low power indoor BSs standard equipment in some
cases will lead to less than 5% mean UL throughput degradation and in other cases will lead to larger losses.
39
This indicates that unsynchronised operation should be possible with careful installation of the indoor BS.
Synchronised operation of indoor BS may be difficult in practice because of the challenges involved in
distributing the common clock signal to indoor BSs.
It is worth noting that the performance criteria in the coexistence studies are maximum 5% throughput loss.
For URLLC use cases 5% loss is not acceptable. For these use cases the relevant throughput loss level
should be closer to 0%. The studies' results show that a close to 0% throughput loss the URLLC use cases
will require an additional isolation of around 20-25 dB. Such additional isolation could be obtained with
accurate indoor BS planning and, for example, with adoption of proper shielding around the building.
The case where the macro-cellular network is the victim has not been simulated because the indoor cells will
be lower power and so are expected to pose a lower risk of interference. However, if there are several
buildings with indoor systems deployed, there could be a need to consider the effect of the aggregate
interference.
In the case of co-channel operation of Macro BS and Indoor BS, the conclusions on coexistence between
the two systems should account for lack of out of block filtering on the Macro BS and on the indoor BS
transmitters' side.
Accounting for the above, agreements among MFCN licensees that operate macro-cellular networks and the
Indoor BS in the same area and in the same band could include the conditions that identify the specific
circumstances under which indoor BS networks could operate in unsynchronised mode.
5.2.3 Semi-synchronised operation based on the ECC baseline out of block power limit
Semi-synchronised operation is similar to synchronised operation, with the exception that simultaneous
UL/DL transmissions between networks can be allowed in some defined parts of the frame. This leads to a
degree of flexibility at the expense of some additional interference that can be controlled to some extent.
Compared to unsynchronised operation, semi-synchronised operation reduces the impact on BS-BS and
MS-MS interference. The results from studies in Section 4.5 show that in specific circumstances the ECC
baseline (as defined in ECC Decision (11)06 (October 2018) Table 3), out of block power limit can be applied
to the semi-synchronised operation.
38
24 dBm TRP was assumed in the study included in ANNEX 7 to this Report. 3GPP 38.104 [8] defines 24 dBm as the maximum TRP
for the Local Area BS power class.
39
For example “careful installation” would include measures like ceiling-mounted installation, placement of indoor BS away from
BS). Such measure seems to be feasible in case of industrial – type of use case (e.g. smart factory indoor coverage).
In order to deploy semi-synchronised operation of TDD mobile networks in a multi-network context (without
guard bands or operator-specific custom filters), MFCN licensees need to reach agreement on:
 Time synchronisation, as for synchronised operation;
 Partial frame alignment: the agreement shall define a default frame structure for synchronised operation
(for which UL/DL directions are defined across the whole frame) and the part of the frame where each
operator is allowed to reverse the default transmission direction (flexible part);
 The conditions under which the ECC baseline out of block power limit can be applied to the semi-
Options for the semi-synchronised operation of Macro BSs and Micro BSs:
The studies summarised in section 4.5 and reported in ANNEX 8 have led to the following results (under
40
specific assumptions ):
 If no changes are applied to the default frame structure, the semi-synchronised operation is identical to
the synchronous case;
 In case an operator selects the UL direction in the flexible part while the default frame structure adopts
the DL direction (DL to UL modifications), the operator which follows the default (DL) frame transmission
direction does not receive additional BS-BS interference compared the synchronous case;
 In case an operator selects the DL direction in the flexible part while the default frame structure adopts
UL direction (UL to DL modifications), the operator which follows the default (UL) frame transmission
direction receives additional BS-BS interference compared to the synchronous case.
"DL to UL modifications": the default DL transmission direction in the flexible part is modified into UL:
 In this case, from BS-BS interference perspective, the network that modifies the default DL transmission
direction into UL will not interfere with the other network, while it will receive additional interference from
the other network.
intermittently and many will be mobile so any interference would be transient.
 It is expected that some 5G use cases will imply the deployment of MSs that are in fixed positions and
close to each other (e.g. crowded stadiums, trains, busses, (home) CPEs in fixed wireless access (FWA)
systems, and fixed machinery/robots in factories). In some of those specific scenarios, the MS-MS
interference might not be negligible anymore: no specific studies were performed with this respect.
 A general framework could be defined at the national level specifying the scope of semi-synchronised
operation in terms of geographical areas: defining whether and in which types of scenario downlink slots
may be unilaterally converted to uplink slots should take into account situations when MS-MS
interferences can be considered acceptable and when they cannot, assuming MS RF requirements are
handled by SDOs and associated harmonised standards.
"UL to DL modifications": the default UL transmission direction in the flexible part is modified into DL:
 In this case, from BS-BS interference perspective, the network that modifies the default UL transmission
direction into DL will interfere with the other network, while it will not receive additional interference from
the other network.
A general framework could be defined at the national level specifying the scope of semi-synchronised
operation in terms of geographical areas and type of cells: defining whether and in which types of scenario
40
A) As shown in Figure 15, the 288 m network shift assumption between macro-cellular networks represents a best case assumption.
Similarly, the 96m network shift assumption between micro-cellular networks represents a best case assumption.
B) Differently from the approach followed in this study, the recommended approach is to use the separation distance and the line-of-
sight probability as input parameter during the coexistence studies between the macro-cellular network and the Micro BSs network.
This approach accounts for the fact that it is difficult to carry out meaningful simulations to assess the interference between two
Micro BS networks in the same urban area since the interference scenario will be strongly impacted by the LoS/NLoS conditions
which radically change as the Micro BSs change their locations with respect to buildings.
C) Coexistence between the macro-cellular network and the Micro BS network was not assessed by this study.
uplink slots may be unilaterally converted to downlink or flexible slots. Such framework should take into
account situations when BS-BS interferences can be considered acceptable and when they cannot.
It is worth noting that operators will not always decide to modify the default transmission direction from UL
into DL (and from DL into UL) in the whole flexible part of the frame. In a typical scenario, an operator might
decide to modify the agreed default frame structure in specific locations (e.g. hot spots) and at specific times
(e.g. specific event or busy hour). In this particular case, only cells in areas where the transmission direction
has been changed will be subject to cross-link interference.
Multi-stakeholder agreements will need to target the optimal balance between transmission direction
flexibility and the additional interference (with associated throughput degradation). Such multi-stakeholder
agreements should account for the following options which are based on the results from section 4.5, in case
of UL to DL flexibility:
 Coexistence is facilitated if semi-synchronised operation is applied to Micro and indoor BS but it could be
technically challenging for indoor BS to be semi-synchronised with outdoor networks;
efficient interference cancellation algorithms have been developed and implemented.
5.3 OPTIONS FOR AGREEMENTS

A general framework could be defined at the national level specifying:
 The technical parameters for synchronised and, for semi-synchronised operation if appropriate as
described in previous Sections;
areas and type of cells (e.g. whether indoor cells may operate in unsynchronised operation, and when
semi-synchronised operation may be used);
 The definition of such framework before the spectrum awards would lead to greater market certainty.
Administrations may facilitate the process to ensure fair and timely agreements in cases where agreements
41
could be more challenging, for example :
 Different operators may prefer different frame structures based on the services they seek to provide. As
a consequence, the negotiation to achieve common parameters (especially on the DL/UL ratio and
performance targets) may become challenging;
 Multilateral agreements (involving all licensees in the band that may interfere with each other) are
needed;
 Agreements on more complex synchronisation frameworks are more difficult to be achieved (e.g.
regional/local licensing);
 Agreements may become more difficult in case of asymmetric or non-mutual interference scenarios (e.g.
macro-cellular networks vs. Indoor BS networks, downlink-only configurations) - see also ECC Report
216 section 3.3 [1];
 Licensees operating networks in the band which do not implement AAS technology in their BSs might
have less incentive in synchronised operation (due the possibility, in case of non-AAS BSs, to add
external filters to meet the ECC restricted baseline out of block power limit).
Licensees may seek to periodically update (e.g. every few years) the characterising synchronisation
framework. Such updates may be necessary to adapt to evolving technology and market requirements (e.g.
latency and DL/UL ratio requirements and advances in semi-synchronised operation).
Administrations might consider consolidating similar systems together in specific portions of the 3400-3800
MHz band. Such measures will facilitate unsynchronised operation between 5G networks and existing MFCN
networks by reducing the number of geographic and spectrum “boundaries” (see ECC Report 287 [9]).
41
Such case have also been assessed in section §3.3 of ECC Report 216 [1].
In case of regional/local assignments, Administrations might consider defining stable borders and
coexistence conditions at those borders. Such measures will facilitate unsynchronised operation between
networks not in the same area e.g. based on approaches defined in ECC Recommendation (15)01 [4].
6 CONCLUSIONS
Starting from the definitions provided in ECC Report 281 [2] for the synchronised, unsynchronised and semi-
synchronised operation, this Report supports Administrations, wishing to do so, in setting up a
synchronisation framework at national level for the introduction of 5G-NR in the 3400-3800 MHz band in a
multi-operator environment.
Benefits and challenges of the above mentioned three operating modes are briefly summarised as follows:
Synchronised operation avoids any BS-BS and MS-MS interferences therefore allowing coexistence
between adjacent networks without the need for guard bands or additional filters. This operating mode
simplifies network deployment because no additional interference mitigation is required. Synchronised
operation leads to the selection of a compatible frame structure, which determines a specific DL/UL
transmission ratio and frame length which contribute to the network performance (e.g. latency, spectral
42
efficiency, throughput and coverage ). A common phase clock reference (e.g. UTC) and
accuracy/performance constraints that depend on the underlining technology (e.g. +/- 1.5 μs for LTE-TDD
and 5G-NR) is required and those aspects and challenges are detailed in ECC Report 216 [1].
Licensees can select the most appropriate frame structure independently and can adapt the frame structure
to service and end user requirements, which may change depending on the location and on time. However,
in a multi-operator scenario, the flexibility in operators’ frame structure selection leads to a number of
interference scenarios that need to be assessed and managed.
Semi-synchronised operation is defined in ECC Report 281 as the operating mode which "corresponds to
the case where part of the frame is consistent with synchronised operation as described above, while the
remaining portion of the frame is consistent with unsynchronised operation as described above. This requires
the adoption of a frame structure for all TDD networks involved, including slots where the UL/DL direction is
not specified, as well as synchronising the beginning of the frame across all networks”. This mode allows
simultaneous UL/DL transmissions between networks in some defined parts of the frame. This leads to a
degree of frame structure flexibility at the expense of some additional interference that can be controlled to
some extent. A common phase clock reference (e.g. UTC) and accuracy/performance constraints that
depend on the underlining technology (e.g. +/- 1.5 μs for LTE-TDD and 5G-NR) is required.
This Report has identified the following items that should be agreed at the national level to enable the three
operating modes.
For synchronised operation, a common framework or a multilateral agreement has to be defined at the
43
national level so that all MFCN licensees in the same band use:
 A common phase clock reference (e.g. UTC), with proper accuracy/performance constraints and
permanent monitoring and agreed remedies in case of accuracy loss;
 A compatible frame structure to avoid simultaneous UL/DL transmissions;
The feasibility and performance impacts of synchronised operation between different radio technologies have
44
to be assessed on a case-by-case basis depending on the specific technologies . The synchronised
operation of 5G-NR and LTE-TDD may imply a cost in terms of user plane latency and performance,
especially with regards to 5G URLLC latency targets. Operators may have the option to reduce the user
42
For example: the size of the guard periods between DL and UL transmissions will have an impact on maximum cell radius. Increasing
the number of UL transmissions has an impact on the UL coverage performance.
43
44
ECC report 216 has assessed LTE-TDD/WiMAX cross-technology synchronisation feasibility. This report has assessed 5G-NR/LTE-
TDD cross technology synchronisation. The case for 5G-NR/WiMAX cross-technology synchronisation has not been assessed but it
is understood that every LTE-TDD configuration has at least one 5G-NR equivalent configuration, making it therefore theoretically
feasible to align a 5G-NR carrier with an adjacent channel WiMAX carrier as assessed in ECC Report 216.
plane latency and RTT, under some circumstances, by using lower frequencies (e.g. 700, 800, 900, 1800
MHz) in combination with the 3400-3800 MHz band (e.g. through Carrier Aggregation or Supplemental
Uplink schemes).
In case of semi-synchronised operation, a common framework or a multilateral agreement has to be

45
defined at the national level so that all MFCN licensees in the same band use:
 A common phase clock reference, as for synchronised operation;
 Partial frame alignment: the agreement shall define a default frame structure as for synchronised
operation (for which UL/DL directions are defined across the whole frame) and at the same time the part
of the frame where each operator is allowed to reverse the default transmission direction (flexible part);
 The terms and conditions under which the ECC baseline out of block power limit can be applied to the
semi-synchronised operation.
Licensees can select the most appropriate frame structure. Licensees will need to agree on the terms and
conditions under which the ECC baseline out of block power limit can be applied to the unsynchronised
operation. Meaning that unsynchronised operation could be allowed at national level in specific cases where
sufficient isolation between interferer and victim base stations exists (e.g. sufficient separation distance or
adoption of low power indoor BSs) and in such cases the ECC baseline out of block power limit can be
applied to the unsynchronised operation.
According to ECC Decision (11)06 (October 2018), in case of unsynchronised and semi-synchronised
46
operation the EC restricted baseline limit applies by default. In this case: the interfering BS transmitter
requires custom filters and guard band.
Operator-specific filters would likely be necessary to meet the ECC restricted baseline emissions limit. Based
on currently available technology, it is assumed that it will be challenging to implement operator-specific
filters cost effectively in AAS BSs. Therefore, this Report provides an analysis on whether and under which
conditions the unsynchronised and semi-synchronised operating modes can be used when base stations
47
implement the ECC baseline out of block power limits . Here follows a concise summary for the main
options that have been identified noting that more details are provided in the "Toolbox Section" (Section 5).
With respect to unsynchronised operation with the ECC baseline out of block power limit, the following
options are identified:
Unsynchronised Macro-cellular networks in the same area

 Without operator-specific filters, it may not be possible to rely on guard bands alone to enable
unsynchronised operation between operators;
 Separation distances are therefore needed but a specific recommendation or single set of trigger
48
values cannot be provided due to the dependency from various factors . The studies show minimum
distances required between unsynchronised Macro-cellular networks could be up to 60 km when co-
channel operation and up to 14 km when operating in the adjacent channel.
Unsynchronised Micro BS networks and Macro-cellular networks in the same area

 The studies show that, in general, adjacent channel unsynchronised operation of Macro-cellular networks
and Micro BS networks might not be feasible in the same area. Separation distances have not been
assessed in this Report.
45
46
See ECC Decision (11)06 (October 2018) Table 4.
47
ECC baseline out of block power limit is defined in Table 3 of ECC Decision (11)06 (October 2018) with reference to the
48
Network technologies and topologies (LTE/5G-NR, non-AAS/AAS BS, BS antenna height), propagation environment and propagation
model, frequency assignments, protection criteria (I/N or network throughput loss at x%, etc.….).
 If there is no Macro-cellular network, adjacent channel unsynchronised operation between two Micro BS
networks might be feasible with careful planning avoiding line of sight between Micro BS.
Unsynchronised Indoor BS networks and Macro-cellular networks in the same area

 Under specific assumptions in the adjacent channel case, unsynchronised operation should be possible
49
with careful installation of the indoor BSs.
 Synchronised operation of indoor BS may be difficult in practice because of the challenges involved in
distributing the common clock signal to indoor BS;
 In case of co-channel operation of Macro BSs and indoor BSs, the lack of out of block filtering on the
Macro BS and on the indoor BS transmitters' sides will need to be considered.
With respect to semi-synchronised operation with the ECC baseline out of block power limit, it is useful
to distinguish DL-to-UL and UL-to-DL modifications compared to the reference frame:
 "DL to UL modifications": the default DL transmission direction in the flexible part is modified into UL:
 From BS-BS interference perspective, the network that modifies the default DL into UL will not
interfere the other network while it will receive additional interference from the other network;
intermittently and many will be mobile so any interference would be transient. It is expected that some
50
5G use cases will imply the deployment of MSs that are in fixed positions and close to each other .
No specific studies were performed on MS-MS interference. Therefore, in case of MSs that are in fixed
positions and close to each other, no conclusion can be derived. In any case, MS RF requirements are
handled by SDOs and associated harmonised standards.
 "UL to DL modifications": the default UL transmission direction in the flexible part is modified into DL:
 From BS-BS interference perspective, the network that modifies the default UL transmission direction
into DL will interfere the other network while it will not receive additional interference from the other
network;
 Coexistence is facilitated if semi-synchronised operation is applied to Micro and indoor BS but it could
be technically challenging for indoor BS to be semi-synchronised with outdoor networks;
efficient interference cancellation algorithms have been developed and implemented.
The actual coexistence feasibility for the different scenarios will depend on the specific circumstances and
assumptions that can only be clarified at national level.
A general framework could be defined at the national level by Administrations wishing to do so specifying:
 The technical parameters for synchronised operation, and for semi-synchronised operation if appropriate
(including reference clock and reference frame structure);
areas and type of cells (e.g. whether indoor cells may operate in unsynchronised operation, and in which
scenarios downlink slots may be unilaterally converted to uplink slots).
Administrations may facilitate the process to ensure fair and reasonable agreements.
Administrations could establish mechanisms through which the parameters characterising the
synchronisation framework are periodically updated. This process could be triggered by the Administrations
or by the licensees.
Administrations might want to consider consolidating similar systems together in specific portions of the
3400-3800 MHz band. Such measures will facilitate unsynchronised operation between 5G networks and
49
For example ”careful installation” would include measures like ceiling-mounted installation, placement of indoor BS away from
BS). Such measure seems to be feasible in case of industrial – type of use case (e.g. smart factory indoor coverage).
50
E.g. crowded stadiums, trains, buses, (home) CPEs in fixed wireless access (FWA) systems.
existing MFCN networks by reducing the number of geographic and spectrum “boundaries” (see ECC Report
287 [9]).
In the case of regional/local assignments, Administrations might consider defining stable borders and
coexistence conditions at those borders. Such measures will facilitate unsynchronised operation between
networks not in the same area.
ANNEX 1: SYNCHRONISATION FRAMEWORKS IN RECENT C-BAND AWARD PROCEDURES
As stated in ECC Report 281: “Several LTE-TDD networks are currently providing services to millions of end
users with hundreds of thousands of BSs deployed in the field adopting synchronisation and alignment of
UL/DL transmissions between operators using adjacent frequency blocks. Such networks provide proven
experience in the field that should be considered as the starting point for the definition of the regulatory
framework for 5G-NR.
In Europe, the majority of legacy TDD networks deployment can be grouped in two categories:
 Based on synchronised operation when operators run their networks without relying on sufficient isolation
(e.g. this is the case of LTE-TDD networks, comprising thousands of BS, in Italy operating in the 3400-
3600 MHz band;
 Based on unsynchronised operation when there is sufficient isolation between operators running their
networks on adjacent frequency blocks (e.g. one operator per region is often assumed).
Going forward, recent advances for newer TDD systems in a multi-operator context encourage
synchronisation more strongly, therefore this situation is expected to evolve in the coming years.
ECC Report 216 section 3.3 describes some potential situations where inter-operator agreement relying
solely on the market may be challenging in a multi-operator context (either at the time of auction, or later in
time). Therefore regulators may get involved at some point in the process in order to ensure an efficient
spectrum usage. This has already been done in the past, and ECC Report 216 ANNEX 3 already describes
a few of them. Since then, some new auctions have happened:
Austria
The Austrian Administration is planning to start the assignment procedure for the 3410-3800 MHz range in
Q1 ’19. The following provisions are described in the tender document from the Telekom-Control-
Kommission [10].
The “LTE compatible” NR frame structure (DSUDDDSUDD) is defined as the “default frame structure” for
which the ECC baseline out of block power limit applies. "Licence holders are responsible for ensuring that
frames are based on a uniform reference time (+/- 1.5 μs), so that all of any licence holder’s frames are
aligned equally and transmissions are consequently synchronised". "…Small cells inside buildings are
51
exempt from synchronisation. The default BEM can be used for such small cells in buildings, provided that
no damaging interference occurs to other licence holders".
According to the tender document: "… the synchronisation frame specified here can be altered by the TKK to
reflect technical and economic conditions when 5G reaches market maturity, in accordance with Art. 57 TKG
2003. If such modifications are indeed made, consideration will nonetheless have to be given in each case to
the proportionality of the measure and the economic impact on the parties affected. Even if any such change
is made, the spectrum holders will have the option of stipulating under private law a synchronisation frame".
The tender document also provides conditions associated with the use of the restrictive BEM when "other
frame structures" are adopted.
Ireland
In its June 2017 Spectrum 3600 MHz band spectrum award [1] Ireland mandated the LTE-TDD frame
configuration #2 with special sub-frame configuration #6 (or equivalent frame structures whose transmit and
receive periods are aligned with this configuration) as the default frame structure which an operator must
comply with in order to be allowed to comply with the "permissive Block Edge Mask". The operator must also
ensure compliance with a common reference time of +/- 1.5µs.
Italy
51
A small cell is defined as "a base station with a maximum EIRP of 24 dBm per 20 MHz of carrier".
In September 2018, a multiband 5G spectrum in Italy followed a light touch approach with respect to the
definition of the synchronisation framework for the 3600-3800 MHz band. The auction rules [12] did not
include specific provisions in relation to the synchronisation framework, leaving to operators the task to agree
on the most suitable framework. Soon after the frequency assignment, the Italian Ministry has announced
that it will facilitate the process by setting up a specific working group among operators that acquired
licenses in the band.
United Kingdom
In its May 2015 statement on the award of the 3.4 GHz spectrum band [14], the UK decided on the LTE-TDD
frame configuration #2 with special sub-frame configuration #6 (or equivalent frame structures whose
transmit and receive periods are aligned with this configuration) as the "preferred frame structure" (see
Figure 26) which an operator must comply with in order to be allowed to comply with the "permissive
transmission mask. An operator unwilling to adopt the "preferred frame structure" must comply with the
"restrictive transmission mask" and the "compatible frame structure" (see Figure 27), i.e. must comply with
semi-synchronised operation.
Licensees are required to synchronise their networks in order to avoid interference to one another, so traffic
alignment and the “preferred frame structure” for transmission with the limits of the "permissive transmission
mask" are mandated to implement the synchronisation. Timeslots must have a duration of 1ms. LTE-TDD
frame configuration #2 (3:1) is compatible with this frame structure.
Figure 26: Preferred frame structure in UK award for the 3400-3600 MHz range in April 2018
Figure 27: Compatible frame structure in UK award for the 3400-3600 MHz range in April 2018
Indoor base stations with a transmit power level below 24 dBm are exempt from synchronisation
requirements unless they cause interference to the macro-cellular network, in which case they are required
to synchronise.
In April 2018, Ofcom conducted the auction, and the 3.4 GHz band plan based on final auction results as
below, as announced by Ofcom [15]. Ofcom will auction 3600-3800 MHz in second half of 2019.
Figure 28: Outcome from UK award for the 3400-3600 MHz range in April 2018
China
For the first 2600 MHz TDD network in the world, China operators have agreed to coordinate their network.
Finally, under the guideline of MIIT (Ministry of Industry and Information Technology) of China, the
synchronisation among operators’ networks at the band 2600 MHz was implemented based on the same
frame structure and the same DL/UL traffic ratio. China MIIT has been actively organising MNOs and
relevant stakeholders to negotiate a single frame structure for synchronisation of 5G networks in the 3500
MHz band.
Japan
A public open hearing of potential operators for Japan 3400-3600 MHz band was held by MIC (Ministry of
Internal Affairs and Communications) in January 2014. All operators presented a clear position in favour of
the TDD duplex scheme, and advocated the necessity of operator’s consensus for collaboration for realising
TDD synchronised operation including the DL/UL configuration, ideally, in order to achieve no guard band for
efficient usage of spectrum resources. All the operators have the same opinion of DL heavy frame
configuration, by referring to the heavy data traffic in downlink side.
The MIC issued the draft guideline for the introduction of 4G for comments in July 2014 taking into account
the opinions expressed at public hearing held in January 2014. The guidelines proposed that 3480-3600
MHz should be assigned for 3 operators (40MHz per operator) for TDD use, and proposed an obligation for
licensees to agree with each other in advance about the matters for TDD synchronised operation, such as
transmission time and frame structure. The draft was approved at the Council in September 2014, and MIC
started to receive applications for the operation of the bands.
Three operators have submitted applications. As first step, regarding the synchronisation operation, NTT
DoCoMo applied for DL/UL ratio of 8:1 or 3:1, while KDDI and Softbank both applied DL/UL ratio of 3:1. The
three operators agreed to hold operator meeting after the band was granted to get mutual agreement on the
UL/DL configuration and frame synchronisation.
In December 2014, the MIC issued the licenses to three applicants as follows based on the discussion
results of Radio Regulatory Council.
NTT DOCOMO: 3480-3520 MHz
KDDI: 3520-3560 MHz
SOFTBANK MOBILE: 3560--3600 MHz
Finally, the three operators agreed with DL/UL 3:1 ratio and implemented the TDD synchronised operation
based on a common 3:1 frame structure. All operators will cover around 50% population by the end of 2018,
with around 57,000 base stations, as announced during the spectrum application procedure.
In April 2018, the MIC allocated the remaining spectrum in the 3400-3600 MHz band to 2 operators, and the
two licenses are to synchronise with the existing networks in the same band. By the year 2022, around 60%
population will be covered with around 33,000 base stations, as planned in the application materials.
SOFTBANK MOBILE: 3400-3440 MHz
NTT DOCOMO: 3440-3480 MHz
52
The MIC issued the draft guideline for the introduction of 5G for comments in November 18. The guidelines
proposed 3600-4100MHz，4500-4600MHz and 27-28.2GHz，29.1-29.5GHz for for TDD use, and proposed
an obligation for licensees to agree with each other in advance about the matters for TDD synchronised
operation such as transmission time and frame structure.
Korea
Korea completed an auction of the 3.5 GHz and 28 GHz bands.

LGU+: 3420-3500 MHz
KT: 3500-3600 MHz
SKT: 3600-3700 MHz
52 Official MIC weblink: http://www.soumu.go.jp/menu_news/s-news/01kiban14_02000358.html

ANNEX 2: TECHNICAL SOLUTIONS FOR NETWORK SYNCHRONISATION
Section 2.1.1 has provided the frequency and phase synchronisation requirements for LTE-TDD and 5G-NR.
This annex provides a simple overview on the mainstream solutions to implement synchronisation.
Currently, the main solution for 5G-NR time synchronisation includes the following two major categories:
 Type 1: distributed synchronisation scheme based on satellite;
 Type 2: centralised synchronisation scheme based on 1588v2 system.
The main principles of these two types of synchronous solutions, advantages and disadvantages and
applicable scenarios are described below.
Distributed synchronisation scheme based on satellite:
Figure 29: Distributed synchronisation scheme based on satellite
As shown in the Figure 29, GNSS signal receivers are directly deployed on base stations, each base station
acquires the available satellite time signals (GPS, Beidou, GLONASS, etc.) to achieve the time
synchronisation between different base stations and to ensure the maximum deviation of any two of the base
stations.
Table 10: Applicability of the distributed synchronisation scheme based on satellite
Applicable scenarios Inapplicable scenarios Pros Cons
 Single stations
can be activated
 The node of very efficiently;  Newly-installed GPS
transmission  The base station  Sites that need is difficult to
network does not location is time construct, leading to
support PTP surrounded by tall synchronisation high installation and
(Precision Time buildings that easily can be directly maintenance costs;
Protocol); block GPS signals; deployed without  High failure rate of a
 Base stations  Indoor base the cooperation single GPS;
located in open stations; with the  Poor maintainability
area;  Difficult to install the transmission and high installation
 Easy to install the GPS antenna network; and maintenance
GPS antenna.  The impact of a costs.
fault in a single
station is small;
Centralised synchronisation scheme based on 1588v2 system:
Figure 30: Centralised synchronisation scheme based on the 1588v2 system
The IEEE standards development organisation has proposed the IEEE 1588v2 accurate time transfer
protocol, which can achieve sub-microsecond precision time synchronisation like the current GPS currency.
As shown in the Figure 30, the clock synchronisation information of the main time source is transmitted
through the 1588v2 protocol packet on the transmission network. The base station can obtain time
information from the transmission network through the 1588v2 interface to achieve synchronisation with the
time source. The accuracy can reach ns level.
Table 11: Applicability of the centralised synchronisation scheme based on 1588v2 system
Applicable scenarios Inapplicable scenarios Pros Cons
 Single site
without  Requires all nodes
of the bearer
 Difficult to obtain the  The transmission additional
antenna network to support
satellite signal; network nodes cannot
engineering; PTP;
 All transmission support PTP;
  Clock;
network nodes support  The transport network High reliability;
synchronisation
PTP protocol. QoS is poor.  Low quality is affected
maintenance by network QoS.
costs.
The two synchronisation methods described above have been widely used by operators around the world.
Operators will take decisions depending on the country and the network situation. For example, operators in
Japan and other regions mainly use distributed synchronisation scheme based on satellite (GPS), and some
operators in Europe choose the centralised synchronisation scheme (IEEE 1588v2). Some other operators
will also consider adopting a combination of two synchronised approaches to improve reliability (e.g. China
Mobile).
ANNEX 3: 5G-NR AND LTE-TDD FRAME STRUCTURES, OPTIONS AND ASSESSMENTS - STUDY #1
AND #2
A3.1 LTE TDD FRAME STRUCTURES AND OPTIONS
LTE-TDD frame structures are defined in [16].
Each radio frame of length 10ms consists of two half-frames of length 5ms each. Each half-frame consists of
five sub-frames of length 1ms. The supported uplink / downlink configurations are listed in Table 12 where:
 “D” denotes the sub-frame which is reserved for downlink transmissions;
 “U” denotes the sub-frame which is reserved for uplink transmissions;
 “S” denotes a special "Subframe" with the three fields: Downlink Pilot Time Slot (DwPTS), Guard Period
(GP) and Uplink Pilot Time Slot (UpPTS). The length of DwPTS and UpPTS is given by
 Table 13 subject to the total length of DwPTS, GP and UpPTS being equal to 1ms.
Uplink / downlink configurations with both 5ms and 10ms DL to UL switch-point periodicity are supported. In
case of 5ms DL to UL switch-point periodicity, the special sub-frame exists in both half-frames. In case of
10ms DL to UL switch-point periodicity, the special sub-frame exists in the first half-frame only.
Sub-frames 0 and 5 and DwPTS are always reserved for downlink transmissions. UpPTS and the sub-frame
immediately following the special sub-frame are always reserved for uplink transmissions.
One radio frame, Tf = 307200Ts = 10 ms
One half-frame, 153600Ts = 5 ms
One slot,
Tslot=15360Ts 30720Ts
Subframe #0 Subframe #2 Subframe #3 Subframe #4 Subframe #5 Subframe #7 Subframe #8 Subframe #9

One
subframe,
30720Ts
UpPT UpPT
DwPTS GP DwPTS GP
S S
Figure 31: LTE-TDD frame structure (for 5ms switch-point periodicity)
Table 12: LTE-TDD uplink / downlink configurations
DL to UL Sub-frame number
UL/DL
configuration 0 1 2 3 4 5 6 7 8 9
Switch-point periodicity
0 5ms D S U U U D S U U U
1 5ms D S U U D D S U U D
2 5ms D S U D D D S U D D
3 10ms D S U U D D D D D D
4 10ms D S U U D D D D D D
5 10ms D S U D D D D D D D
6 5ms D S U U U D S U U D
Table 13: Configuration of special sub-frame (lengths of DwPTS / GP / UpPTS)
Normal Cyclic prefix in downlink Extended cyclic prefix in downlink
UpPTS UpPTS
Special sub-
frame
configuration DwPTS Normal Extended DwPTS Normal Extended
Cyclic Cyclic Cyclic Cyclic
prefix in prefix in prefix in prefix in
uplink uplink uplink uplink
0 6592.Ts 7680.Ts
1 19760.Ts 20480.Ts
2 21952.Ts 2192.Ts 2560.Ts 23040.Ts
3 24144.Ts 25600.Ts
4 26336.Ts 7680.Ts
5 6592.Ts 20480.Ts 4384.Ts 5120.Ts
6 19760.Ts 23040.Ts
4384.Ts 5120.Ts
7 21952.Ts - - -
8 24144.Ts - - -
Ts=1/(15000x2048) seconds.
Selecting a synchronisation option for LTE-TDD requires:

 Selection of a timing reference (beginning of the frame);
 Selection of a frame structure;
 Selection of special sub-frame configuration.
A3.2 5G-NR FRAME STRUCTURE AND OPTIONS
5G-NR frame structures are defined in [17].
Downlink and uplink transmissions are organised into frames with 10 ms duration, each consisting of ten
sub-frames of 1 ms duration. The number of consecutive OFDM symbols per sub-frame is given by:
subframe, µ subframe, µ
N symb = N symb
slot
N slot
.
slot
N symb
is the number of OFDM symbols in a slot
frame,µ
N slot
is the number of slots in a frame
subframe,µ
N slot
is the number of slots in a sub-frame
Each frame is divided into two equally-sized half-frames of five sub-frames each with half-frame 0 consisting
of sub-frames 0 – 4 and half-frame 1 consisting of sub-frames 5 – 9.
Slots are defined within a frame depending on the subcarrier spacing configuration µ, according to Table 14
and Table 15.
Table 14: 5G-NR number of OFDM symbols per slot, slots per frame, and slots per sub-frame
for normal cyclic prefix.
slot frame,µ subframe,µ

Subcarrier spacing configuration µ N symb N slot N slot
0 14 10 1
1 14 20 2
2 14 40 4
3 14 80 8
4 14 160 16
Table 15: 5G-NR number of OFDM symbols per slot, slots per frame, and slots per sub-frame
for extended cyclic prefix
slot frame,µ subframe,µ

Subcarrier spacing configuration µ N symb N slot N slot
2 12 40 4
As shown in Table 16, the UL or DL transmissions are configured within each slot. OFDM symbols in a slot
can be classified as:
 “D” denoting the OFDM symbol which is reserved for downlink transmissions;
 “U” denotes the sub-frame which is reserved for uplink transmissions;
 “X” denoting a flexible symbol for which transmission can either be un downlink or uplink
Table 16: 5G-NR slot formats for normal cyclic prefix
Symbol number in a slot

Format
0 1 2 3 4 5 6 7 8 9 10 11 12 13
0 D D D D D D D D D D D D D D
1 U U U U U U U U U U U U U U
2 X X X X X X X X X X X X X X
3 D D D D D D D D D D D D D X
4 D D D D D D D D D D D D X X
5 D D D D D D D D D D D X X X
6 D D D D D D D D D D X X X X
7 D D D D D D D D D X X X X X
8 X X X X X X X X X X X X X U
9 X X X X X X X X X X X X U U
10 X U U U U U U U U U U U U U
Format Symbol number in a slot
11 X X U U U U U U U U U U U U
12 X X X U U U U U U U U U U U
13 X X X X U U U U U U U U U U
14 X X X X X U U U U U U U U U
15 X X X X X X U U U U U U U U
16 D X X X X X X X X X X X X X
17 D D X X X X X X X X X X X X
18 D D D X X X X X X X X X X X
19 D X X X X X X X X X X X X U
20 D D X X X X X X X X X X X U
21 D D D X X X X X X X X X X U
22 D X X X X X X X X X X X U U
23 D D X X X X X X X X X X U U
24 D D D X X X X X X X X X U U
25 D X X X X X X X X X X U U U
26 D D X X X X X X X X X U U U
27 D D D X X X X X X X X U U U
28 D D D D D D D D D D D D X U
29 D D D D D D D D D D D X X U
30 D D D D D D D D D D X X X U
31 D D D D D D D D D D D X U U
32 D D D D D D D D D D X X U U
33 D D D D D D D D D X X X U U
34 D X U U U U U U U U U U U U
35 D D X U U U U U U U U U U U
36 D D D X U U U U U U U U U U
37 D X X U U U U U U U U U U U
38 D D X X U U U U U U U U U U
39 D D D X X U U U U U U U U U
40 D X X X U U U U U U U U U U
41 D D X X X U U U U U U U U U
42 D D D X X X U U U U U U U U
43 D D D D D D D D D X X X X U
44 D D D D D D X X X X X X U U
45 D D D D D D X X U U U U U U
46 D D D D D D X D D D D D D X
47 D D D D D X X D D D D D X X
Format Symbol number in a slot
48 D D X X X X X D D X X X X X
49 D X X X X X X D X X X X X X
50 X U U U U U U X U U U U U U
51 X X U U U U U X X U U U U U
52 X X X U U U U X X X U U U U
53 X X X X U U U X X X X U U U
54 D D D D D X U D D D D D X U
55 D D X U U U U U U X U U U U
56 D X U U U U U D X U U U U U
57 D D D D X X U D D D D X X U
58 D D X X U U U D D X X U U U
59 D X U U U U U D X U U U U U
60 D X X X X X U D X X X X X U
61 D D X X X X U D D X X X X U
62-255 reserved
It can be noted that, apart from slot formats 0 - 2, all slot formats contain a mix of D, U and X symbols.
Selecting a synchronisation / semi-synchronisation option for 5G-NR requests:

 Selection of a timing reference (beginning of the frame);
 Selection of normal or extended prefix;
 Selection of a subcarrier spacing configuration;
 Selection of a slot configuration.
A3.3 STUDIES ASSESSING THE IMPACT OF THE 5G-NR FRAME STRUCTURE ON 5G

PERFORMANCE
A3.3.1 STUDY #1
ECC Report 281 states: “Although complete alignment of UL/DL transmissions between LTE-TDD and 5G-
NR can be achieved…, this would have implications on the minimum latency achievable by 5G-NR. Full
synchronisation of the 5G-NR slot structure and LTE-TDD configuration brings significant drawbacks to the
5G-NR implementation. Many of the benefits of 5G-NR are linked precisely to the frame structure. Reverting
to the LTE-TDD structure would imply higher latency, higher MS memory cost, TCP performance loss,
mobility performance loss and spectral efficiency loss, although networks could be designed to overcome
some of these drawbacks.”
This Section provides additional elements to qualify the statement from ECC Report 281.
A3.3.1.1 Impact on latency
When 5G-NR frame structure is aligned with LTE-TDD, the UL occurrences are spaced out, matching the
LTE-TDD format. The MS must wait for the next UL opportunity to send the HARQ response, which may be
several slots (TTIs) later. Assuming 30 kHz subcarrier spacing, the timeline of 5G-NR is aligned with LTE-
TDD Configuration 2 as shown in Figure 32.
LTE TDD
Config 2 SSF7 D D GU U D D D D GU U D D
NR
30 kHz SCS
1st ACK opportunity in

8 slots
Figure 32: HARQ timeline when 5G-NR frame structure is aligned with LTE-TDD frame configuration
#2
As shown, there are seven consecutive DL-only slots without any UL control block. The earliest opportunity
to send the HARQ response is the UL symbols at the end of the 8th slot.
Similarly, for UL traffic, the consecutive DL slots will also block scheduling requests (SR) and UL data
transmission.
Considering scheduling and MS/network processing latency, this frame structure will lead to L1 latency >
4ms. Note that IMT-2020 eMBB latency requirement is 4ms and URLLC latency requirement is 0.5ms. When
5G-NR deployments follow the LTE-TDD frame structure for coexistence purposes, they would not be able to
meet the IMT-2020 requirements and most importantly deployment of innovative services such as URLLC
would not be possible.
Simulations were performed on the impact of the 5G-NR frame structure on latency. The 5G-NR
DDDDDDDSUU multi slot structure (30 kHz SCS, start shifted to the first DL slot) is with respect to timing
and UL/DL configuration an identical match with LTE-TDD frame configuration #2 also known as
DSUDDDSUDD.
D S U D D D S U D D
LTE TDD
Config 2
D D D S U U D D D D D D D S U U D D D D
NR
(30 kHz)
Figure 33: LTE-TDD – 5G-NR slot matching assumption
Figure 34 depicts the possible HARQ timeline and the transmission opportunities of the two multi-slot
structures. Figure 34 summarises the timing relations and constants for each sub-process.
Figure 34: HARQ timeline and transmission opportunities comparisons

Table 17: Timing relations and constants
DSDU DDDDDDDSUU (*)

Delay in slots Delay in slots
K0: DL Grant to DL Tx 0 0
K1: DL Tx to DL ACK 1 1
K2: UL Grant to UL Tx 1 1
K3: DL NACK to DL re-Tx

3 3
grant
K4: UL Tx to UL re-Tx grant 3 3
K5: SR to UL grant 2 2
DL HARQ processes 4 8
SR periodicity 1ms 5ms

(*) LTE-TDD-like TDD configuration still assumes typical 5G-NR processing delays
With the given assumptions, the DSDU configuration shows significant benefits over DDDDDDDSUU with respect to HARQ RTT
and UL scheduling delay.
Table 18: Simulation analysis – latency
DSDU DDDDDDDSUU
# required HARQ processes 4 8
DL HARQ RTT 2-3 ms 5ms
UL HARQ RTT 2ms 5ms
UL scheduling delay 1-2 ms 4.5-9.5 ms
The HARQ Round Trip Time (RTT) gives the minimum periodicity for the transmission of a new data packet
(transport block) including ACK in one HARQ process.
Due to fewer UL ACK or data transmission opportunities in DDDDDDDSUU, the simulations result in more
than twice the time (5ms) that is needed to complete one HARQ round trip as compared to DSDU (2-3 ms).
In the case of HARQ retransmissions this delay would multiply.
The UL scheduling delay determines the time needed to send the first high priority/low latency data packet
when no UL grant is given. It consists of the following steps:
1 Packet arrives at transmit buffer;
2 Wait for opportunity to send scheduling request (SR);
3 Send scheduling request;
4 Wait for UL grant;
5 Send UL data using granted resources.

The considerations show the significant benefits of the DSDU scheduling delay (1-2 ms) over the time
required in the case of LTE-TDD synchronisation (4.5-9.5 ms). This is achieved by more frequent transmit
opportunities for UL SR and UL data, and suitable to multiplex URLLC services with existing eMBB traffic.
A3.3.1.2 Impact on capacity
The increased flexibility of 5G-NR frame structure also has a direct impact on the overall capacity of the
network.
The more frequent UL opportunities, for instance, also allows a higher spectral efficiency due to fast channel
feedback. The UL symbols every 1ms allows MS to send sounding reference signals (SRS) and channel
quality information (CQI), allowing the gNB to have an up-to-date estimate of the channel conditions. More
accurate channel estimation allows for a more efficient usage of beamforming and better rate control through
more accurate modulation and coding scheme (MCS) selection. The result is improved cell capacity.
As the cell coverage is typically defined by PDSCH SE (spectral efficiency) at the cell edge, improved
beamforming efficiency implies a coverage improvement too – as the same cell edge SE can be reached at
higher path loss conditions.
On the other hand, when 5G-NR and LTE-TDD transmission direction is aligned, UL symbols are available
less frequently, leading to less accurate channel state information and reduction in capacity.
Simulations compare the latency, the spectral efficiency and the user-perceived throughput for a 5G-NR
DDDDDDDSU vs. 5G-NR DSDU slot configuration.
System simulations have been conducted to determine the benefits of the short 1ms switching of DSDU vs.
5ms of DDDDDDDSU with respect to spectral efficiency and user-perceived throughput.
The assumptions in Table 19 were used for the simulations.

Table 19: Simulation inputs
Parameter Value
Scenario UMi, ISD = 200 m

Layout 57 cells w/wraparound, 10 MSs/cell
# Antenna elements, # TXRU BS: (M, N, Mg, Ng, P) = (8, 8, 1, 1, 2) with 64 TXRU
Outdoor MSs (2 V elements combined) (X-pol) 4 Rx at MS (X-pol)
Carrier frequency 20%
Bandwidth 3.5 GHz
Numerology Sys bandwidth = 100 MHz, Sim bandwidth = 10 MHz
BS antenna down tilt 30 kHz SCS, 0.5ms slot
BS antenna spacing 14 degrees
Antenna gain dH = 0.5 λ, dV = 0.8 λ
Max Tx power (over 100 MHz) BS: 44 dBm; MS: 26 dBm
Noise figure BS: 5 dB; MS: 9 dB
Processing gain for SRS Based on Link-Sim
MS antenna spacing 0.5λ linear array (X-pol)
Doppler 3 kmph (Indoor), {10,30,60,90,120} kmph (Outdoor)
Penetration loss According to 38.901 (low-loss model)
Guard band overhead 10% (Useful bandwidth: 25 RBs of 360 kHz each over 10 MHz)
UL SRS power control Targets -5 dB UL SNR
Traffic model Full buffer and Bursty FTP model 3, 0.5 MB file size
Scheduler MU MIMO + sub-band p-fair scheduling (5 sub-bands)
Channel estimation Realistic
To analyse the impact to the spectral efficiency, an outdoor user running a full buffer DL traffic pattern has
been simulated with different moving speeds. More frequent opportunities to transmit sounding reference
signals (SRS) lead to better spectral efficiency over the PDSCH symbols in a fast fading channel. Faster
sounding allows better tracking of channel fluctuations, thus allowing improved demodulation performance.
Figure 35 compares the simulated spectral efficiency at 5ms and 1ms SRS transmission opportunities.
Figure 35: Spectral efficiency gains vs. speed

It can be seen that the fast switching of DSDU achieves a better spectral efficiency across all speeds as
compared to LTE-TDD compatible DDDDDDDSUU frame structure. While the median gain is 30% to 40%,
the gain at the lower percentile (e.g., cell edge conditions) rises to 70%.
To simulate the effect of the slot structure on user perceived throughput in a realistic scenario, a bursty traffic
pattern (Bursty FTP model 3, 0.5 MB file size, variable file arrival time) was simulated. The shorter DL/UL
switching periodicity of DSDU creates more transmission opportunities. The improved spectral efficiency
enables the use of larger transport blocks. With these advantages, the gain of the median throughput can be
as high as 50% (593 Mbps DSDU vs. 394 Mbps DDDDDDDSUU). Even in cell edge conditions, a 23% gain
can still be achieved.
Figure 36: Bursty traffic – perceived throughput vs. file arrival rate
The simulations show that latency and spectral efficiency significantly benefit from the DSDU configuration.
These advantages justify the additional DL/UL guard period overhead incurred by the more frequent DL/UL
switching.
A3.3.2 STUDY #2
The following Sections provide an assessment in terms of latency and capacity performance for the three
frame structures provided in Figure 37.
Among the three frames that are addressed in this study, the DDDDDDDSUU frame structure is the only one
to be LTE-TDD compatible.
DDDSU frame D D D S U
DDDDDDDSUU frame D D D D D D D S U U
DSDU frame D S D U
Figure 37: 5G-NR frame structures for evaluation – slot level
Differently from LTE-TDD, 5G-NR allows for the assignment of DL and UL transmission directions at OFDM
symbol level (in LTE-TDD the UL/DL assignment is done at sub-frame level), the assessment therefore
depends on the specific choices at OFDM symbol level which are illustrated in Figure 38.
D D D S (10:2:2) U
DL PDCCH
DL PDCCH
DL PDCCH
DL PDCCH
DL PDCCH
DL PDCCH
DL PDCCH
DL PDCCH
GP
GP
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
D D D D D D D S (10:2:2) U U
DL PDCCH
DL PDCCH
DL PDCCH
DL PDCCH
DL PDCCH
DL PDCCH
DL PDCCH
DL PDCCH
DL PDCCH
DL PDCCH
DL PDCCH
DL PDCCH
DL PDCCH
DL PDCCH
DL PDCCH
DL PDCCH
GP
GP
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
D S (10:2:2) D U (1:2:11)
DL PDCCH
DL PDCCH
DL PDCCH
DL PDCCH
DL PDCCH
DL PDCCH
DL PDCCH
GP
GP
GP
GP
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
UL
UL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
DL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
UL
Figure 38: 5G-NR frame structures for evaluation – OFDM symbol level
A3.3.2.1 Latency assessment (grant-free UL transmissions)

53
The latency assessment results in Table 4 account for the “grant-free” UL transmissions feature (also
known as “configured grant”) which is an optional feature for 5G-NR in 3GPP. The “grant-free” UL is
beneficial for low latency since it avoids the need to first transmit a scheduling request on UL followed by a
scheduling grant on downlink before UL data transmission can take place
It is to be noted that decisions to mandate features for Rel-15 5G-NR MSs were made in consideration of
eMBB services, which are the first targets on the 5G market. Most of the features related to low latency
and/or reliability are optional. This includes not only the “grant-free” feature, but also other features such as
the mini-slots (frequency control monitoring and short transmission durations), MS processing capability #2
(necessary for the "self-contained" slot operation), dynamic signalling of slot format (see Table 16), etc.
Simulations results with grant-free based UL transmission are valuable in deriving the lowest possible user
plane latency performance.
According to Report ITU-R M.2410 [18], the user plane latency is the contribution of the radio network to the
time from when the source sends a packet to when the destination receives it (in ms). It is defined as the
one-way time it takes to successfully deliver an application layer packet/message from the radio protocol
layer 2/3 SDU ingress point to the radio protocol layer 2/3 SDU egress point of the radio interface in either
uplink or downlink in the network for a given service in unloaded conditions, assuming the mobile station is in
the active state.
Based on the definition and the evaluation method provided in Report ITU-R M.2412 [19], the components of
user plane latency are listed in Table 20.
53
In case of grant-based transmission, the latency is expected to be higher.
Table 20: Components of user plane latency
ID Component Notations
tUE = tUE,rx + tUE,tx

For DL:
tUE,rx is the time interval between when the PDSCH is received and when the
data is decoded;
tUE,tx is the time interval between when the data is decoded, and when the ACK /
1 MS processing delay NACK packet is generated.
For UL:
tUE,tx is the time interval between when the data arrived, and when the packet is
generated;
tUE,rx is the time interval between when the ACK is received and when the ACK is
decoded.
Frame alignment tFA,DL or tFA,UL
2 (transmission It includes the waiting time (e.g. in TDD, data transmission needs to wait for the
alignment) next available DL/UL non-slot/slot)
TTI for data packet
3 tdata_duration
transmission
4 HARQ retransmission tHARQ
tBS = tBS,rx + tBS,tx
For UL:
tBS,rx is the time interval between when the PUSCH is received and when the
data is decoded;
tBS,tx is the time interval between when the data is decoded, and when the
5 BS processing delay ACK/NACK packet is generated.
For DL:
tBS,tx is the time interval between when the data arrived, and when the packet is
generated;
tBS,rx is the time interval between when the ACK is received and when the ACK is
decoded.
tUP = (tBS,tx + tFA,DL) + tdata_duration + tUE,rx + n×tHARQ
Total one way user where:
-
plane latency for DL tHARQ = (tUE,tx + tFA,UL) + tdata_duration + tBS,rx + (tBS,tx + tFA,DL) + tdata_duration + tUE,rx,
n is the number of re-transmissions (n≥0)
tUP = (tUE,tx + tFA,UL) + tdata_duration + tBS,rx + n×tHARQ
Total one way user where:
plane latency for UL tHARQ = (tBS,tx + tFA,DL) + tdata_duration + tUE,rx + (tUE,tx + tFA,UL) + tdata_duration + tBS,rx,
n is the number of re-transmissions (n≥0)
The role of the described components in a BS-to-MS data transmission procedure is illustrated in Figure 39.
Figure 39: illustration of User Plane latency components
It is noted that the values of the above components depend on the frame structure and numerology, MS
capability on processing, as well as PDSCH / PUSCH mapping type. These impact factors are further subject
to duplexing schemes like FDD, TDD and TDD+SUL (Supplemental UpLink).
The MS processing time, tUE, has been agreed in 3GPP RAN1#90bis, known as “MS capability 1” (PDSCH
processing capability 1). For different numerologies, the value of tUE is expressed in terms of OFDM symbols
(OS), say K OS, as listed in Table 21 (see Section 5.3 in [20]).
Table 21: MS processing time (PDSCH processing capability 1)
PDSCH decoding time N1 [symbols]

Sub-carrier
µDL
spacing No additional PDSCH DM-RS Additional PDSCH DM-RS
configured configured
0 15 kHz 8 13
1 30 kHz 10 13
2 60 kHz 17 20
3 120 kHz 20 24
The BS processing time (tBS) is assumed to be the same as that of MS’s. For this evaluation, it is further
assumed that the two parts of tMS and tBS, that is, tX,rx and tX,tx, are equal. Based on these assumptions, one
has
tBS,rx = tBS,tx = tUE,rx = tUE,tx = K/2
If the number of OFDM symbols per TTI is M, then tdata_duration = M. Taking FDD as an example, the average
user plane latency is calculated as below,
DUP [symbol] = K + M + p × 2 × ( K + M )
Where p is the error probability of the first HARQ retransmission.
For TDD, the extra frame alignment delay for both DL and UL, tFA,DL and tFA,UL, might be larger than 0 due to
the DL/UL configuration (the time needed to wait for the next available DL/UL TTI).
The following assumptions were made as basis for this evaluation:

 The MS processing time is determined with the assumption of MS capability 1 and no additional DMRS
configured;
 PDSCH / PUSCH mapping Type B is employed, which is more flexible to support non-slot based
scheduling;
 Additionally, the packet is assumed to be arrived randomly in any symbol in any slot.
Based on the assumptions above, the average latencies for DL and UL for different frame structures are
illustrated in Table 22and Table 23.
Due to the shorter DL/UL switching period, the user plane latency and round-trip time (RTT) of DSDU frame
is lower in most cases.
The user plane latency and RTT associated with DDDSU and DDDDDDDSUU frames can be further
reduced by using lower frequencies (e.g. 700, 800, 900, 1800 MHz) in combination with the 3400-3800 MHz
band (e.g. through Carrier Aggregation or Supplemental Uplink schemes). The resulting RTT will meet the
most stringent latency requirement for URLLC and eMBB simultaneously.
Table 22: DL latency evaluation results for 5G-NR frame structures (SCS = 30 kHz)
Slot / non- Frame structure (GP: 2 OFDM symbols)

slot based Latency
scheduling DDDSU DDDDDDDSUU DSDU
User plane p=0 0.53 0.57 0.52

2OS non- latency
slot based (ms) p=0.1 0.71 0.88 0.63
scheduling
RTT (ms) 1.77 3.02 1.12
User plane p=0 0.87 0.93 0.90
Slot-based latency
(ms) p=0.1 1.07 1.25 1.04
scheduling
RTT (ms) 1.98 3.24 1.42
Table 23: UL latency evaluation results for 5G-NR frame structures (SCS = 30 kHz)
Slot / non- Frame structure ( GP: 2 OFDM symbols)

slot based Latency
scheduling DDDSU DDDDDDDSUU DSDU
User plane p=0 1.22 2.01 0.78

Non-slot latency
based (ms) p=0.1 1.39 2.30 0.88
scheduling
RTT (ms) 1.71 2.95 1.05
User plane p=0 1.61 2.40 1.46
Slot-based latency
(ms) p=0.1 1.80 2.72 1.62
scheduling
RTT (ms) 1.90 3.14 1.69
A3.3.2.2 Latency assessment in case of support from lower frequency bands (grant-free UL transmissions)
If the lower frequency bands are used in combination (e.g. Carrier Aggregation or Supplemental Uplink
schemes) with the 3400-3800 MHz band with DDDDDDDSUU, the RTT delay will be significantly reduced
based on the analysis in Section A3.3.2.1, and the performance will be improved for 5% UPT in case of SUL.
In this evaluation, it was assumed that one MS always uses either the SUL for all its UL transmissions, or
always uses the 3400-3800 MHz UL for all its UL transmissions, based on measured RSRP. NR
specifications allow configuring the MS to use both uplinks in a TDM manner, which provides more flexibility
in the operation than simulated here. With this respect, the simulation results can be taken as a best case.
Accounting for partial resources availability for NR users in the SUL band (or in the uplink portion of the band
in case of CA) would reduce the gain in latency to some extent.
It is to be noted, that licensees for the 3400-3800 MHz frequencies do not necessarily have access to lower
frequency bands.
While the 1800 MHz SUL band and the 3400-3800 MHz frequencies may be related by second-order inter-
modulation products, such interference can be always avoided by ensuring the MS always uses one of the
two uplinks only. When configured with the two uplinks, the MS is configured with timing patterns that ensure
uplink transmissions only in time-division multiplexing manner, as specified in 3GPP. Other choices of SUL
frequencies are also available in bands specified by 3GPP (e.g. 700, 800 or 900 MHz), where second-order
inter-modulation products would not appear in the 3400-3800 MHz band. TDM patterns to avoid the same
inter-modulation products are also specified in 3GPP for LTE-NR dual connectivity with or without the use of
a SUL.
Table 24: DL latency evaluation results for 5G-NR frame structures (SCS = 30 kHz)
Frame structure (GP: 2 OFDM symbols)

Slot / non-slot
Latency DDDSU DDDDDDDSUU
based scheduling
+ SUL + SUL
User plane p=0 0.53 0.57

2OS non-slot based latency (ms) p=0.1 0.60 0.66
scheduling
RTT (ms) 0.78 0.82

Slot-based scheduling latency (ms) p=0.1 0.99 1.04
RTT (ms) 1.12 1.17
Table 25: UL latency evaluation results for 5G-NR frame structures (SCS = 30 kHz)
Slot / non- Frame structure ( GP: 2 OFDM symbols)

slot based Latency
scheduling
DDDSU + SUL DDDDDDDSUU + SUL
Non-slot latency
based (ms) p=0.1 0.58 0.58
scheduling
RTT (ms) 0.82 0.86
Slot-based latency
(ms) p=0.1 1.06 1.08
scheduling
RTT (ms) 1.43 1.59
A3.3.2.3 DL Capacity assessment (grant-based UL transmissions)
The following aspects have an impact on the 5G-NR frame structure performance in terms of system
capacity:
Guard Period (GP) overhead:
GP is introduced at the DL/UL switching point. Frequent DL/UL switching will introduce larger GP overhead,
which determines system capacity reduction.
UL slot availability:
The UL slot availability affects the Channel State Information (CSI) feedback and ACK / NACK feedback
delay. The more frequent availability of UL sub-frames will reduce the CSI and ACK / NACK feedback
delays. The reduced CSI feedback delay is beneficial for system capacity for fast varying channels. The
reduced ACK / NACK feedback delay is beneficial for reducing RTT delay, and increasing user perceived
throughput in some cases.
DL and UL ratio:
The DL and UL ratio associated with a certain frame structure should be consistent with the DL and UL traffic
pattern. Otherwise, the DL or UL system capacity will be degraded.
Considering the above aspects, it is observed that the DSDU frame structure performance benefits from fast
CSI measurement and feedback, however the frequent DL/UL switching that characterise this frame
structure leads to extra overhead.
On the other hand, the DDDSU and DDDDDDDSUU frame structures may suffer from a relatively slower CSI
feedback, yet benefiting from reduced GP transmission overhead.
Taking into account the channel varying nature that depends on the device moving speed distribution
(adopted assumptions: 80% indoor users with 3km/h and 20% outdoor users with larger moving speeds), the
trade-off of the CSI feedback and the overhead introduced by DL/UL switching point needs to be carefully
evaluated for different candidate frame structures.
In this study, the spectral efficiency and the user-perceived throughput (UPT) are evaluated for the
DDDDDDDSUU, DDDSU, and DSDU frame structures. The detailed assumptions for this study are provided
in Section A3.3.2.6.
Based on the evaluation assumptions listed in Table 26, the total overhead for the different frame structures
are provided in Table 27. In Table 27, DSDU is associated with the highest overhead due to the increased
CSI-RS and GP overhead for the fast CSI measurement and DL/UL switching. As the length of GP increases
from 2 to 4 OFDM symbols, the difference in the total overhead for the DSDU frame and for the DDDSU
frame will be further increased. The DDDSU frame structure provides good balance for overhead and CSI
acquisition.
Table 26: Assumptions for overhead calculations for different frame structures in DL
Overhead
DDDSU DDDDD DDSUU DSDU
assumption
2 complete symbols in the

2 complete symbols in the 2 complete symbols in the downlink dominant slot; 1
PDCCH
downlink dominant slot downlink dominant slot complete symbol in the
uplink dominant slot
Type II DMRS, dynamic Type II DMRS, dynamic Type II DMRS, dynamic
DMRS calculation according to calculation according to calculation according to
paired layer number paired layer number paired layer number
4 ports per MS with 5 slots 4 ports per MS with 10 4 ports per MS with 4 slots
CSI-RS period; slots period; period;
40 REs/PRB for 10 users 40 REs/PRB for 10 MSs 40 REs/PRB for 10 MSs
SSB 8 SSBs per 20 ms 8 SSBs per 20 ms 8 SSBs per 20 ms
Overhead
DDDSU DDDDD DDSUU DSDU
assumption
2 burst consecutive slots 2 burst consecutive slots 2 burst consecutive slots

TRS per 20ms, bandwidth with per 20ms, bandwidth with per 20ms, bandwidth with
51 PRBs 51 PRBs 51 PRBs
GP 2/4 symbols 2/4 symbols 2/4 symbols
Table 27: Overhead calculation for different frame structures in DL
GP (2 symbols) GP (4 symbols)
Overhead DDDDD DDDDD

DDDSU DSDU DDDSU DSDU
DDSUU DDSUU
PDCCH 0.15 0.15 0.16 0.15 0.15 0.16

DMRS 0.10 0.10 0.09 0.10 0.10 0.09
CSI-RS 0.06 0.03 0.08 0.06 0.03 0.08
SSB 0.03 0.03 0.03 0.03 0.03 0.03
TRS 0.002 0.002 0.002 0.002 0.002 0.002
GP 0.04 0.02 0.09 0.07 0.04 0.19
Total overhead 0.38 0.33 0.46 0.41 0.35 0.55
For spectrum efficiency evaluation, the performance under downlink full buffer traffic with different moving
speeds is illustrated in Figure 40 and Figure 41. It should be noted that the moving speeds indicated in
Figure 40 and Figure 41 apply to 20% outdoor users, and the moving speed of 80% indoor users is kept to 3
km/h.
According to Figure 40 and Figure 41, the cell average spectrum efficiency for DDDSU and DDDDDDDSUU
are comparable. The DDDDDDDSUU frame structure can attain higher cell average spectrum efficiency and
cell-edge spectrum efficiency due to the lower overhead, for the low speed scenario in particular.
When the speed of outdoor users is 10 km/h, the cell average and cell-edge spectrum efficiency can achieve
15% and 23% gain for DDDDDDSUU vs. DSDU, respectively. This is determined by the slow channel
variation due to the low speed users and in this case, the fast CSI measurement and feedback cannot bring
significant gain.
Figure 40: DL spectrum efficiency with different speeds - cell average spectrum efficiency
Figure 41: DL spectrum efficiency with different speeds - cell-edge spectrum efficiency
The assessment of the User Perceived Throughput (UPT) is provided for downlink burst traffic. It is noted
that UPT in this evaluation is defined on the basis of a data packet, i.e., UPT = (packet size) / (time to
complete the transmission of this packet). Therefore the UPT performance statistic is conducted per data
packet basis, which offers the assessment on data packet throughput distribution.
In Figure 42, Figure 43, Figure 44 and Figure 45 the UPT performance with different arrival rates is illustrated
to evaluate the performance under different traffic loads.
In the evaluation, the speed of the 20% outdoor users is assumed to be 30 km/h and the 80% indoor users
again keep the moving speed of 3km/h.
With the good balance of overhead and feedback delay, DDDSU frame structure has better performance in
most cases and the gain compared to DSDU can be achieved by more than 10%. Due to the lower
overhead, the frame structure DDDDDDDSUU also can obtain gain for average UPT and 95% UPT (up to
20% gain), where re-transmission is less needed.
For 5% UPT, it is observed that DSDU has around 5% gain over DDDDDDDSUU. This is due to the fact that
these poor packets are usually transmitted via re-transmissions. In this case, the RTT delay is a major factor
that impacts UPT, and DSDU can outperform by reduced RTT delay. However, the gain is not significant
after the trade-off with the GP overhead is taken into account.
Besides, if the lower frequency bands are used in combination (e.g. Carrier Aggregation and Supplemental
Uplink schemes) with the 3400-3800 MHz band with DDDDDDDSUU, the RTT delay will be significantly
reduced based on the analysis in section A3.3.2.1, and the performance will be improved for 5% UPT in case
of SUL.
Figure 42: DL UPT with different arrival rates in burst traffic - user average UPT
Figure 43: DL UPT with different arrival rates in burst traffic - 5% UPT
Based on the above evaluation for latency and capacity, the study demonstrates that the frame structures of
DDDSU and DDDDDDDSUU can introduce good system capacity and high user perceived throughput. The
study also shows that, in case of DDDDDDDSUU frame structure, the combined use of the 3400-3800 MHz
band with low frequencies with SUL scheme can further reduce the latency and boost the lower part (i.e. cell
edge) of user perceived throughput.
A3.3.2.4 UL Capacity assessment (grant-based UL transmissions)
Similarly to the capacity assessment in DL, here follows an assessment for the UL overhead, spectral
efficiency, and UPT for the different 5G-NR frame structures.
Based on the evaluation assumptions listed in Table 28, the total UL overhead for the different frame
structures are provided in Table 29 and Table 27. For DMRS overhead assumption in Table 28, 2 complete
OFDM symbols are assumed to acquire good performance of channel estimation. In Table 29, DSDU frame
structure has the highest overhead resulted by the frequent SRS and DMRS transmission. As the length of
GP increases, the valid UL resource for DSDU will be further reduced and the overhead increases. On the
contrary, the DDDSU frame structure provides good balance for overhead, UL CSI acquisition, and RTT
delay.
Table 28: Overhead assumption for different frame structures in UL
Overhead
DDDSU DDDDDDDSUU DSDU
assumption
7 OFDM symbols; 7 OFDM symbols; 7 OFDM symbols;

PUCCH
2 PRBs/symbol 2 PRBs/symbol 2 PRB/symbol
2 complete OFDM 2 complete OFDM 2 complete OFDM

DMRS
symbols symbols symbols
2 OFDM symbols per 5 2 OFDM symbols per 10 2 OFDM symbols per

SRS
slots slots 4 slots
Table 29: Overhead calculation for different frame structures in UL
GP (2 symbols) GP (4 symbols)
Overhead DDDDD DDDDD

DDDSU DSDU DDDSU DSDU
DDSUU DDSUU
PUCCH 0.017 0.018 0.021 0.017 0.018 0.025
DMRS 0.125 0.133 0.154 0.125 0.133 0.182
SRS 0.125 0.067 0.154 0.125 0.067 0.182
Total overhead 0.267 0.218 0.329 0.267 0.218 0.389

For UL
spectrum efficiency evaluation, the performance under full buffer traffic with different moving speeds is
illustrated in Figure 46 and Figure 47. Similarly to what was observed in DL, the cell average and cell-edge
spectrum efficiency for DDDSU and DDDDDDDSUU are better than that of DSDU due to the combined
effect from overhead and the fast CSI acquisition. The cell average and cell-edge spectrum efficiency can
achieve 14% and 15% gain for DDDDDDSUU vs. DSDU, respectively, when the speed of outdoor users is
10 km/h.
Figure 46: UL spectrum efficiency with different speeds - cell average spectrum efficiency
Figure 47: UL spectrum efficiency with different speeds – cell-edge spectrum efficiency
In Figure 48 and Figure 51, the UL UPT performance with different arrival rates is presented under different
traffic loads. In the evaluation, the speed of the 20% outdoor users is assumed to be 30 km/h and the 80%
indoor users still keep the moving speed of 3km/h.
Under different traffic loads, DDDSU frame structure has the best performance in most cases. Compared to
DSDU with 3 files/s arrival rate, the frame structure DDDDDDDSUU can obtain 14% and 12% gain for
average UPT and 95% UPT, respectively, where re-transmission is less needed.
For 5% UPT, it is observed that DSDU has around 2% gain over DDDDDDDSUU. This is similar to the
observation in DL 5% UPT, which resulted by the higher RTT delay for DDDDDDDSUU. However, the gain is
not significant after the trade-off with the GP overhead is taken into account.
Figure 48: UL UPT with different arrival rates in burst traffic - user average UPT
Figure 49: UL UPT with different arrival rates in burst traffic – 5% UPT
A3.3.2.5 UL Capacity assessment in case of support from lower frequency bands (grant-based UL
transmissions)
The combined use of lower frequency bands with the 3400-3800 MHz band (e.g. via Carrier Aggregation and
Supplemental Uplink schemes) allows to enhance the 5G-NR uplink coverage providing good UL
performance under the same inter-site distance used for the lower frequency bands.
It is to be noted, that licensees for the 3400-3800 MHz frequencies do not necessarily have access to lower
frequency bands.
In general terms, the evaluation procedure for the 5G-NR multi-band operation with TDD band + SUL is
summarised hereafter.
a) Step 1: users are dropped across the network coverage area, each user selects its serving cell and
frequency band fi (i=1 or 2) based on RSRP. gNB configures MS’s frequency band fi (i=1 or 2);
b) Step 2: Data transmission simulation on each frequency band;
c) Step 3: Collect each user’s simulated throughput on specific band;
d) Step 4: Generate the CDF of user throughput from all users on frequency band f1 and f2, and take the
th
5 percentile point of the user throughput CDF as the cell-edge data rate.
For frequency bands f1 and f2, the carrier frequency 3500 MHz (TDD band) and 1800 MHz (SUL band) are
used, respectively. For SUL band, the duplexing mode FDD is used.
Since the channel bandwidth in the TDD band (100 MHz) and SUL band (20 MHz) are different in the
evaluation, the spectrum efficiency metric cannot intuitively present the performance difference. Hence the
cell average throughput and cell-edge throughput metrics are selected to evaluate the SUL capacity.
In Figure 52 and Figure 53, the UL throughput for TDD band only and TDD+SUL band is provided. In the
evaluation, the speed of the 20% outdoor users is assumed to be 30 km/h and the 80% indoor users keep
the moving speed of 3km/h.
For cell average throughput, using the SUL band allows for about 40% gain due to the bandwidth increase.
More spectrum resources can be allocated to the 70% users in the TDD band when the 30% users are
offloaded to the SUL band. Up to 5-times gain can be reached for cell-edge users exploiting the SUL band
due to the fact that such users benefit from the lower path loss and the sufficient bandwidth,
Figure 52: UL throughput with TDD+SUL band - cell average throughput
Figure 53: UL throughput with TDD+SUL band – cell-edge throughput
A3.3.2.6 Assumptions
The system configuration parameters for the Macro BS Urban test environment are illustrated in Table 30
and the technical parameters are illustrated in Table 31 and Table 32.
Table 30: Simulation parameters for Macro BS Urban test environment
System configuration parameters Values
Test environment Macro BS Urban

3.5GHz (1 Macro BS layer)
Carrier frequency
Lower frequency band: 1.8GHz (when applicable)
BS antenna height 20m
Total transmit power per TRxP 46 dBm
System configuration parameters Values
MS power class 23 dBm

Inter-site distance 500 m
Indoor: 80%
User distribution
Outdoor: 20%
Indoor users: 3 km/h
MS speeds of interest
Outdoor users: 10/30/60/90/120 km/h
Inter-site interference modelling Explicitly modelled
BS noise figure 5 dB
MS noise figure 9 dB
BS antenna element gain 8 dBi
MS antenna element gain 0 dBi
Thermal noise level -174 dBm/Hz
Full buffer;
Traffic model Burst buffer: file size 0.5 Mbytes, arrival rate
0.5/1/2/3
MS density 10 MSs per TRxP
Outdoor MSs: 1.5 m
Indoor UTs: 3(nfl – 1) + 1.5;
MS antenna height
nfl ~ uniform(1,Nfl) where
Nfl ~ uniform(4,8)
Channel model variant Channel model A (follow TR36.873)
TRP number per site 3
Mechanic tilt 90 degree in GCS (pointing to horizontal direction)
Electronic tilt 100 degree
Handover margin (dB) 1
UT attachment Based on RSRP from port 0
Wrapping around method Geographical distance based wrapping
Minimum distance of TRxP and MS Follow corresponding channel model variant
Polarised antenna model Model-2 in TR 36.873
frequency reuse 1
Antenna pattern Follow TR 36.873
Below rooftop BS antenna deployment 0%
Indoor user terminal penetration loss Follow channel model A in TR 36.873
Table 31: Technical configuration parameters in DL
Technical configuration Parameters Values
Waveform OFDM-based
Multiple access OFDMA
Duplexing TDD
Network synchronisation Synchronised
Modulation Up to 256 QAM

Numerology 30kHz SCS
Guard band ratio on simulation bandwidth 8.2% (for 20 MHz)
Simulation bandwidth 20 MHz (51 PRBs)
Frame structure DDDSU, DDDDDDDSUU, DSDU
Transmission scheme Closed SU/MU-MIMO adaptation
DL CSI measurement Precoded CSI-RS based, non-PMI
DL codebook N/A
MU dimension Up to 12 layers
SU dimension Up to 4 layers
Non-precoded SRS, 4 Tx ports,
8 PRBs per symbol,
SRS transmission 2 symbols per 10 slots for DDDDDDDSUU,
2 symbols per 5 slots for DDDSU,
2 symbols per 4 slots for DSDU
CQI/RI feedback every 10 slots for DDDDDDDSUU,
CQI/RI feedback every 5 slots for DDDSU,
CSI feedback
CQI/RI feedback every 4 slots for DSDU,
Non-PMI feedback, Sub-band based
Interference measurement SU-CQI
Max CBG number 1
ACK/NACK delay N+1
Re-transmission delay the next available DL slot after receiving NACK
32 TxRU, (8,8,2,1,1;2,8),
Antenna configuration at TRxP (dH, dV)=(0.5, 0.8)λ,
Vertical 1 to 4.
4 TXRU, (1,2,2,1,1; 1,2)
Antenna configuration at MS
(dH, dV)=(0.5, N/A)λ
Scheduling PF
Receiver MMSE-IRC
Channel estimation Non-ideal
PDCCH 2 symbols
SSB 8 SSB / 20ms
DDDDDDDSUU: 4 ports per MS with 10 slots
period (MS-specific beamformed CSI-RS);
Overhead DDDSU: 4 ports per MS with 5 slots period (MS-
CSI-RS specific beamformed CSI-RS);
DSDU: 4 ports per MS with 4 slots period (MS-
specific beamformed CSI-RS);
40 REs/PRB for 10 users per cell
DMRS Type II, up to 12 ports
2 burst consecutive slots per 20ms, bandwidth 51

TRS
PRBs
GP 2 symbols
Table 32: Technical configuration parameters in UL
Waveform OFDM-based
Multiple access OFDMA
Duplexing TDD, FDD
Network synchronisation Synchronised
Modulation Up to 256 QAM
Numerology TDD: 30kHz SCS; FDD: 15Hz SCS
TDD: 8.2% (for 20 MHz), 1.8% (for 20 MHz);
Guard band ratio on simulation bandwidth
FDD: 4.6% (for 20 MHz)
TDD: 20 MHz, 100 MHz
Simulation bandwidth
FDD: 20MHz
Frame structure DDDSU, DDDDDDDSUU, DSDU, UUUUU
Transmission scheme SU adaptation
UL CSI measurement Non-precoded SRS and wideband PMI
UL codebook Codebook based
SU dimension Up to 2 layers
Non-precoded SRS, 2 Tx ports, 8 PRBs per symbol,
2 symbols per 10 slots for DDDDDDDSUU,
SRS transmission 2 symbols per 5 slots for DDDSU,
2 symbols per 4 slots for DSDU,
2 symbols per 5 slots for FDD
Max CBG number 1
32 TxRU, (8,8,2,1,1;2,8),
Antenna configuration at TRxP (dH, dV)=(0.5, 0.8)λ,
Vertical 1 to 4.
2 TXRU, (1,1,2,1,1; 1,2)
Antenna configuration at MS
(dH, dV)=(0.5, N/A)λ
Power control parameters P0 = -60 dBm, alpha = 0.6
Continuous PRB allocation model: follow TS 38.101;
Power backoff
Non-continuous PRB allocation: additional 2 dB reduction
Scheduling PF
Receiver MMSE-IRC
Channel estimation Non-ideal
PUCCH 7 symbols, 2 PRBs per symbol
Overhead DMRS 2 complete symbols
SRS 2 symbols per SRS transmission period
ANNEX 4: MFCN PARAMETER VALUES AND ASSUMPTIONS FOR SIMULATIONS
The following tables list the parameters that have been agreed for the coexistence studies at the basis of this
Report.
Table 33: 5G-NR BS and User Equipment parameters
Parameter Value Remarks
Carrier Frequency 3.5 GHz

60 MHz
Channel bandwidth
(for Macro, Micro and indoor BS deployments)
Handover margin 3 dB
MACRO-CELLULAR NETWORK PARAMETERS
Layout Hexagonal grid
Number of sites 19
Number of cells per site 3 cells/site
Inter-Site Distance (ISD) 500 m (3GPP Case 1)
Network shift (Macro BS Baseline: 70m
↔ Macro BS case) Additional for reference: 288 m
BS antenna height 25 m
BS antenna mechanical
10 degrees ECC Report 281
down tilt
BS antenna horizontal 3 dB
65°
beam-width
BS antenna vertical 3 dB
80°
beam-width
BS antenna array 8x8 array
BS Element spacing (X, Y) 0.5 λ , 0.5 λ
BS antenna beamforming Beamforming towards MSs with (8x8) array
BS max transmitted power 51 dBm
Inter-site fading correlation 0.5
Maximum Coupling Gain -70 dB (i.e. ~= 30m from BS)
Scheduler (# active MSs / OPTION 1: Proportional Fairness
cell) OPTION 2: Round robin
OPTION 1: 3 active MSs / cell For reference: 10 active MSs /
# active MSs / BS
OPTION 2: 1 active MSs / cell Cell (Macro BS to Macro BS)
MICRO DEPLOYMENT PARAMETERS

BS antenna type 3D direction antenna
# sites 19
Option 1: 1 cell / site

# sectors per site
Option 2: 3 cells / site
Inter-Site Distance (ISD) 166 m
BS antenna mechanical Option 1: 10 degrees
ECC Report 281
down tilt Option 2: 0 degrees
BS antenna horizontal 3 dB
65°
beam-width
BS antenna vertical 3 dB
80°
beam-width
BS antenna array 8x8 array
BS antenna beamforming Beamforming towards MSs with (8x8) array
3GPP TS 38.104 Table 6.2.1-1:
defines the rated output power
for Medium Area BS as <=
BS max transmitted power 40 dBm 38dBm. 40dBm value is kept
since no significant difference
is expected in the simulations
results.
Maximum Coupling Gain -53 dB (i.e. @ 3m from BS)
Min. separation distance
between Macro BS and 30m
Micro BS
INDOOR BS NETWORK PARAMETERS
One floor (50x120 m)
3GPP TR 36.873
(Related to the small cell
transmission power)
# floors / building
Recommendation ITU-R M. 2101-0 (02/2017) -

Figure 5
Also addressed in From 3GPP
"Modelling and simulation of IMT networks and
TR 38.901 ("Study on channel
Floor layout systems for use in sharing and compatibility
model for frequencies from 0.5
studies)"
to 100 GHz") Table 7.2-2
From 3GPP TR 36.873 ("Study on 3D channel
model for LTE-TDD): 50m x 120m
# INDOOR BS / floor 2 3GPP TR 36.873
Other configurations could be
% ceiling mounted INDOOR
100% considered with clear setup
BS in the building
description
Floor height (antenna

height for ceiling mounted 3m
INDOOR BS)
Min. distance between
Macro BS and building 70 m
(outer wall)
OPTION 1 (challenging case: small distance, BS antenna panel azimuth pointing
towards the building)
# buildings / Macro BS 1
Recommendation ITU-R P.2109 ("Prediction of

building entry loss")
Indoor penetration loss
Traditional building: 100%
Thermally efficient building: 0%
OPTION 2 (statistical with random drop of buildings in a cell)

# buildings / sector 3
Minimum distance between
70 m
buildings
# sectors / INDOOR BS 1 panel / site
Inter Site Distance (ISD) 60 m
INDOOR BS antenna
3m 3GPP TR 36.873
height
INDOOR BS antenna
90 deg (ceil mount)
mechanical tilt
INDOOR BS antenna
5 dBi
element gain
INDOOR BS antenna array 4x4
INDOOR BS Element
0.5 λ , 0.5 λ
spacing (X, Y)
INDOOR BS antenna
Beamforming towards MSs with (4x4) array
beamforming
Ohmic loss 3 dB
INDOOR BS noise figure 5 dB
3GPP TS 38.104 Table 6.2.1-1:
defines the rated output power
INDOOR BS max TRP 24 dBm
for Local Area BS as <=
24dBm.
MS PARAMETERS
MS max transmitted power 23 dBm
MS antenna type Omni directional (3 dimensional)
MS antenna gain -4 dBi
1.5 m (if connected to Macro and Micro)
MS antenna height
1 m (if connected to indoor BS)
UL Power Control NOTE: parameters are chosen to maximise the network performance: maximise the
parameters 5% and average t-put network performance (for a single operator)
OPTION 1: (Set 1 in 3GPP TR 36.942)

CLxile = 94 dB, ɣ = 1
MSs connected to
5% MSs transmitting at full power
Macro BS
OPTION 2:
P0 = -92 dBm, ɣ = 0.8
Clxile = 82 dB, ɣ = 1
MSs connected to
Micro BS
OPTION 2:
P0 = -86dBm, ɣ = 0.8
MSs connected to
CLxile = 94 dB, ɣ = 1
Indoor BS
MS distribution
Uniform MS distribution (excluding the area of the

building with indoor BS)
MSs connected to OPTION 1:
Macro BS 80% indoor – see 3GPP 38.802 table 8.2.1-1
OPTION 2: 0% indoor
Uniform MS distribution
OPTION 1:
MSs connected to
70% indoor (urban) - see ECC Rep. 281, ITU-R
Micro BS M.2292
OPTION 2: 7% indoor
Low power Indoor BS

MSs connected to Uniform MS distribution over the floor
assumed, targeting indoor
indoor BS - Indoor MSs: 100 %
coverage only.
Table 34: Radio channel parameters
Channel parameters
Propagation models
Macro BS -> Macro BS Free space path loss
3D UMa
Macro BS -> Macro MS
Indoor penetration (to reach indoor MSs) according
(same network)
to Table 7.4.3-3
Macro BS -> Micro BS 3D Uma 3GPP TR 38.901 ("Study on
Micro BS -> Micro BS channel model for frequencies
from 0.5 to 100 GHz")
3D UMI
Micro BS -> Micro MS
(same network)
Macro BS -> INDOOR BS 3D UMa
Channel parameters
MS connected to Macro BS
<->
3D UMi
MS connected to INDOOR
BS
Macro BS -> MS (indoor
InH – Office 3GPP TR 38.901
propagation)
For MS connected to Macro BS: according to

3GPP TR 38.901 – Clause 7.4.3.1 “O2I building
Outdoor to indoor penetration loss”, parameter settings in Table 3GPP TR 38.901
penetration loss 7.4.3-3. “O2I building penetration loss model for
single-frequency simulations <6 GHz”
LoS probability model Indoor - Open office 3GPP TR 38.901
Table 35: 5G MFCN Antenna element and array parameters
Parameter Values
  θ ′′ − 90°  2 
′′ (θ ′′, φ ′′ = 0°) = − min 12
AdB  , SLAV 
  θ 3dB 
Vertical cut of the radiation
pattern (dB) 
with θ 3dB = 65°, SLAV = 30 dB and θ ′′ ∈ [0°, 180°]
  φ ′′  2 
′′ (θ ′′ = 90°, φ ′′) = − min 12
AdB  , Amax 
  φ3dB 
Horizontal cut of the radiation
pattern (dB) 
with φ3dB = 65°, Amax = 30 dB and φ ′′ ∈ [- 180°, 180°]
3D radiation pattern (dB) AdB ′′ (θ ′′, φ ′′ = 0°) + AdB

′′ (θ ′′, φ ′′) = − min{− ( AdB ′′ (θ ′′ = 90°, φ ′′)), Amax }
Maximum directional gain of an

8 dBi
antenna element GE dBi
Number of beamforming
(NV , NH ) = (8,8)
antenna elements
0.5λ horizontal separation
Element spacing
0.5λ vertical separation
ANNEX 5: UNSYNCHRONISED OPERATION OF TWO MACRO BS MFCN NETWORKS NOT IN THE

SAME AREA- STUDY #3 AND #4
A5.1 STUDY #3
Coexistence between unsynchronised Macro BSs at 3400-3800 MHz is assessed when additional isolation is
available.
More specifically the objective is to derive the minimum isolation, expressed in terms of separation distance,
required between two unsynchronised networks when all deployed BSs meet the ECC baseline out of block
power limits as defined in ECC Report 281.
The following two cases have been considered for the deployments of the interfering and victim BSs, see
Figure 54 and Figure 55:
 The two networks operating in adjacent frequency channels;
 The two networks operating in the same frequency channel.
Operator
Distance
2
Operator
1 f
3400 3800
Interferer
Victim
Figure 54: Unsynchronised operation of two MFCN networks not in the same area –
adjacent channel
Operator
Distance
2
Operator
Interferer 1 f
Victim 3400 3800
Figure 55: Unsynchronised operation of two MFCN networks not in the same area –
co-channel
The impact of BS-BS interference due to simultaneous UL and DL transmissions is taken into account.
The impact of interference is assessed by evaluating the degradation in the mean uplink throughput of the
victim MFCN.
For each of the two cases, the two scenarios are addressed according to whether the interferer and victim
BSs use AAS technology or not, namely:
 Interference from AAS BSs to non-AAS BSs;
 Interference from AAS BSs to AAS BSs.
Accounting for the propagation loss as a function of distance, this analysis shows how the mean UL
throughput loss decreases with the distance increase between the victim and interferer networks.
Since this analysis assumes fixed ACIR (given by 3GPP SEM), the required guard band remains unchanged
regardless of the distance between victim and interferer networks.
Note that an AAS BS is considered to form a beam towards a MS (assumed to be uniformly distributed within
a cell), whereas a non-AAS BS is assumed to have a fixed antenna directional pattern.
A5.1.1 SIMULATION PARAMETERS
Table 36 and Table 37 show parameters used in simulating the various scenarios.
The antenna directional pattern for non-AAS BSs is modelled as per described in ECC Report 203. Table 3
shows the antenna array characteristics modelled for AAS BSs, aligned with earlier agreements in the PT1
correspondence group.
Table 36: Parameters for “AAS to non-AAS” scenario
Interferer Victim
Beamforming towards MSs with (8×8) array. Fixed directional pattern

MSs uniformly distributed in each hexagonal cell. (effectively single antenna)
Network Hexagonal cells Network Hexagonal cells
Deployment ISD = 500m. Deployment ISD = 500m.
Maximum antenna
Element gain 8 dBi 18 dBi
gain
Channel bandwidth 60 MHz Channel bandwidth 20 MHz
Effective channel Effective channel
90% 90%
bandwidth bandwidth
Tx (conducted) power 51 dBm/(60 MHz) Noise figure 5 dB
45 dB 46 dB
ACLR (from 3GPP for 60MHz ch. ACS (from 3GPP for 20MHz
BWBWBWBWBWBWBW.W) channel bandwidth)
Table 37: Parameters for “AAS to AAS” scenario
Interferer Victim
Beamforming towards MSs with (8×8) array. Beamforming towards MSs with (8×8) array.
MSs uniformly distributed in each hexagonal cell. MSs uniformly distributed in each hexagonal cell.
Network Hexagonal cells Network Hexagonal cells
deployment ISD = 500m. deployment ISD = 500m.
Element gain 8 dBi Element gain 8 dBi
Channel bandwidth 60 MHz Channel bandwidth 60 MHz
Effective channel Effective channel
90% 90%
bandwidth bandwidth
Tx (conducted) power 51 dBm/(60 MHz) Noise figure 5 dB
ACLR 45 dB ACS 46 dB
Interferer Victim
(from 3GPP for 60MHz (from 3GPP for 60MHz

channel bandwidth) channel bandwidth)
Table 38: Antenna radiation pattern
Antenna radiation pattern
According to 3GPP TR 37.840 (Section 5.4.4.2):

a E dB (θ, φ) = −min�−�AE,V dB (θ) + AE,H dB (φ)�, Am dB �,
φ 2
AE,H dB (φ) = −min �12 � � , Am dB �,
φ3dB
θ − 90° 2
AE,V dB (θ) = −min �12 � � , SLAV dB �,
θ3dB
Antenna element where
directional pattern a E dB (θ, φ) 3 dB elevation beam width θ3dB = 65°,
3 dB azimuth beam width ϕ3dB = 80°,
Front-to-back ratio Am = 30 dB,
Side-lobe ratio SLAV = 30 dB.
NOTE: a E (θ, φ) ≤ 1.
NOTE: Each antenna element is larger in size in the vertical
direction, and so θ3dB < ϕ3dB . See 3GPP TR 37.840.
Antenna element gain GE dB 8 dBi
Number of BS beamforming
elements (8,8)
(NV, NH)
0.9λ vertical separation.
0.6λ horizontal separation.
Element spacing
NOTE: Larger vertical spacing provides narrower array beam
width in elevation. See 3GPP TR 37.840 (Table 5.4.4.2.1-1).
The following propagation model is used from an interfering BS to a victim BS: Recommendation ITU-R
P.452 20% time percentage, smooth earth path loss – for both co-channel and adjacent channel case.
Figure 56: ITU-R P.452 P=20% propagation loss (smooth earth, Tx/Rx antenna height = 25 m)
A5.1.2 SIMULATION RESULTS
The studies have characterised the impact of BS-BS interference between MFCNs with simultaneous UL/DL
transmission in terms of the resulting degradation in UL throughput of the victim MFCN. Specifically, the co-
channel and the adjacent channel operation of the interfering and victim networks was considered. For each
of the two cases, the “AAS to non-AAS” and “AAS to AAS” interferer to victim scenarios have been
addressed.
A5.1.2.1 Unsynchronised operation of two MFCN networks not in the same area – adjacent channel
AAS to AAS scenario
Assuming all BSs meet the baseline limit defined in ECC Report 281, the minimum required separation
distance of ca. 10.5km is required to ensure mean UL throughput degradation of ca. 5%
The following figure shows the relationship between the mean uplink throughput loss and distance.
Figure 57: Mean UL throughput loss (%) vs. separation distance for the AAS to AAS case
– adjacent channel
AAS to non-AAS scenario

 Assuming all BSs meet the baseline limit defined in ECC Report 281.
The minimum required separation distance of ca. 14km is required to ensure mean UL throughput
degradation of ca. 5%.
Figure 58: Mean UL throughput loss (%) vs. separation distance for the AAS to non-AAS case
– adjacent channel
A5.1.2.2 Unsynchronised operation of two MFCN networks not in the same area – co-channel
When the two networks operate on the same channel, compared with adjacent channel results, additional
42.5 dB loss is required to guarantee uplink throughput loss below 5%. Larger separation distances will
therefore be required in this case.
With larger separation distances, different terrain environments will significantly affect the propagation loss,
impacting the actual separation distance to a significant extent. As described in Section 2 of, the ITU-R
P.452, the specified propagation model with 20 % time percentage considers smooth-earth path loss.
AAS to AAS scenario
Assuming all BSs meet the baseline limit defined in ECC Report 281.
The minimum required separation distance of ca. 60 km is required to ensure mean UL throughput
degradation of ca. 5%.
Figure 59: Mean UL throughput loss (%) vs. separation distance for the AAS to AAS case
– co-channel
AAS to non-AAS scenario
Assuming all BSs meet the baseline limit defined in ECC Report 281.
The minimum required separation distance of ca. 60km is required to ensure mean UL throughput
degradation of ca. 5%
Figure 60: Mean UL throughput loss (%) vs. separation distance for the AAS to non-AAS
– co-channel
A5.1.3 CONCLUSIONS FROM RESULTS
The following conclusion can therefore be derived from the results above for the unsynchronised operation of
Macro BSs meeting the baseline ECC out of block power limits as defined in ECC Report 281, belonging to
different networks in different areas:
 AAS to AAS scenario:
 Adjacent channel case: assuming all BSs meet the baseline limit defined in ECC Report 281, the
minimum required separation distance of ca. 10.5km is required to ensure mean UL throughput
degradation of ca. 5%;
 Co-channel case: assuming all BSs meet the baseline limit defined in ECC Report 281, the minimum
required separation distance of ca. 60km is required to ensure mean UL throughput degradation of
ca. 5%.
 AAS to non-AAS scenario:
 Adjacent channel case: assuming all BSs meet the baseline limit defined in ECC Report 281, the
minimum required separation distance of ca. 14 km is required to ensure mean UL throughput
degradation of ca. 5%;
 Co-channel case: assuming all BSs meet the baseline limit defined in ECC Report 281, the minimum
required separation distance of ca. 60km is required to ensure mean UL throughput degradation of
ca. 5%.
A5.2 STUDY #4
A5.2.1 SIMULATION SCENARIOS AND ASSUMPTIONS
Simulation scenario is illustrated in Figure 61. Network A and Network B are two unsynchronised macro-
cellular networks separated at a distance D which is the distance between the two nearest sites of the
network A and B.
Figure 61: Simulation scenario between two networks (Network A and Network B)
Three possible cases are
1 Non-AAS Network A to non-AAS Network B
This case represents two LTE-TDD FWA networks in the C-band (3400-3800 MHz)
2 AAS Network A to non-AAS Network B

This case represents the situation of 5G-NR AAS network to a non-AAS LTE-TDD FWA network in the
C-band (3400-3800 MHz)
3 AAS Network A to AAS Network B
This case represents the situation of two 5G-NR AAS networks in the C-band (3400-3800 MHz)
The system parameters and network assumptions are summarised in Table 39below.
Table 39: IMT System parameters and assumptions
Channel bandwidth (MHz) 20

BS antenna height (m) 25
Cell Range (m) 300

Non-AAS BS Tx Power (dBm) 49
Non-AAS BS antenna gain (dBi) 18
AAS BS Tx Power (dBm) 46,2
AAS antenna element 8x8
A5.2.2 SIMULATION RESULTS FOR THE CASE NON-AAS TO NON-AAS WITH PROPAGATION
MODEL ITU-R P.2001-2.
The simulation results of co-channel interference and adjacent channel interference for the case of Non-AAS
Network A to non-AAS Network B are given in Table 40 and Table 41.
Table 40: Co-channel interference simulation results (non-AAS to non-AAS)
Throughput Loss (%)

D (km)
Ref_Cell Network
60 2,163 0,985
55 10,5 5
52 25,1 12,4
50 40,2 21,3
49 48,2 26,3
48 58,4 32,1
45 91,5 55,6
30 100 100
10 100 100
For the co-channel interference case where two networks A and B are in urban area, and the separation
zone between the two networks are in rural environment, based on reference cell 5% throughput loss
protection threshold, the required separation distance is about 58 km. Based on reference cell 50%
throughput loss protection threshold, the required separation distance is about 49 km.
Table 41: Adjacent interference simulation results (non-AAS to non-AAS)

Throughput Loss (%)

D (km)
Ref_Cell Network
33 2,9 1,4
32 4,1 2
31 5,7 2,7
30 7,7 3,7
20 27,3 15,6
15 39 23,1
12 50,2 30,2
10 60,1 36,4
For the adjacent channel interference case where two networks A and B are in urban area, and the
separation zone between the two networks are in rural environment, based on reference cell 5% throughput
loss protection threshold, the required separation distance is about 31 km. Based on reference cell 50%
throughput loss protection threshold, the required separation distance is about 12 km.
A5.2.3 SIMULATION RESULTS FOR THE CASE NON-AAS TO NON-AAS WITH PROPAGATION
MODEL ITU-R P.452-16
The simulation results for the case of non-AAS to non-AAS in co-channel case are presented in Table 42.
The results for adjacent channel are presented in Table 43. These results were obtained with the
propagation model Recommendation ITU-R P.452-16 [21] at 50% time.
Table 42: Co-channel simulation results with P.452-16
Network capacity
D (km) iRSS_unwanted
loss
50 4,96% -92,1
Table 43: Adjacent channel simulation results with Recommendation ITU-R P.452-16
D (km) Network capacity loss iRSS_unwanted iRSS_Blocking
15 4,90% -94,2 -99,5

10 9,58% -90,6 -95,9
A5.2.4 SIMULATION RESULTS FOR THE CASE AAS TO AAS WITH PROPAGATION MODEL ITU-R
P.452-16
The simulation results for the case of AAS to AAS in co-channel case are presented in Table 44. The results
for adjacent channel are presented in Table 44 and Table 45. These results were obtained with the
propagation model P.452-16 [21] at 50% time.
Table 44: Co-channel simulation results with Recommendation ITU-R P.452-16

D (km) Network capacity loss iRSS_unwanted
52 2,73% -91,4
50 4,99% -88,6
49 6,30% -87,1
40 34,90% -72,5
30 64,40% -60
20 77,10% -53,2
Table 45: Adjacent channel simulation results with Recommendation ITU-R P.452-16
D (km) Network capacity loss iRSS_unwanted iRSS_Blocking
15 4,51% -89,9 -95

12 6,10% -88,6 -93,9
10 7,37% -87,3 -92,6
5 16,20% -80,2 -87,3
A5.2.5 SEPARATION DISTANCE CALCULATION BASED ON I/N PROTECTION RATIO
The required separation distances based on I/N protection ratio using the propagation model ITU-R P.2001-2
are given in Table 46. Based on I/N=-6 dB, the required separation distance is 44 km. This approach is valid
only for the case non-AAS to non-AAS. For the case AAS to AAS or AAS to non-AAS, due to the dynamic
moving of AAS antenna radiation pattern, the antenna gain in the direction of the victim BS is not constant.
Table 46: Required separation distance based on I/N protection ratio for non-AAS to non-AAS
I/N Ptx G1 G2 BW NF N (dBm/18 I (dBm/18 PL

D (km)
(dB) (dBm) (dB) (dB) (MHz) (dB) MHz) MHz)) (dB)
-6 49 2,4 2,4 18 5 -96,4 -102,4 156,2 44
A5.2.6 APPLICATION OF TRIGGER VALUES
Between two neighbouring countries, there exists a physical borderline, the field strength trigger value is
defined at borderline, the C-Band trigger values are given in the ECC Recommendation (15)01.[4].
Between two unsynchronised networks within a given country the physical borderline does not exist, ECC
Recommendation (15)01 cannot be applied directly to deal with the case between two unsynchronised TDD
networks within a country. There are two possible ways to deal with the two unsynchronised TDD networks
within a country:
1 Define a separation distance;
2 Define a trigger value at the victim network side.
These two approaches should be equivalent; there is no need to define both together.
The trigger value can be defined at the victim network edge towards the interfering network, such as the
reference cell shown in figure 1. It can be defined at the BS antenna height, e.g. 25 m, or at 3 m similar to
the receiving antenna height used in the ECC Recommendation (15)01.
A5.2.7 ANTENNA HEIGHT CONVERSION FACTOR
There is no analytical formula for converting the field strength level between 25 m and 3 m, the differences of
path losses calculated with different receiving antenna heights of 25 m and 3 m with the propagation model
ITU-R P.2001-2 [22] at the frequency 3600 MHz and transmitting antenna height at 25 m in the rural
environment are given in Table 47.
Table 47: Differences of Path losses at 25 m and 3m (ITU-R P.2001-2)
ITU-R P.2001-2 f=3600 MHz

Diff
D (km) Hrx=25 m Hrx=3 m
10 123.64 124.22 0.58

20 129.75 141.56 11.81
30 135.79 157.4 21.61
40 150.08 173.1 23.02
50 165.4 188.24 22.84
60 180.63 195.2 14.57
Propagation model ITU-R P.1546-5 (Land) is valid for frequency range until 3 GHz. The differences of path
losses calculated with different receiving antenna heights of 25 m and 3m with the propagation model
Recommendation ITU-R P.1546-5 (Land) at the frequency 3000 MHz and transmitting antenna height at 25
m in the rural environment are given in Table 48.
Table 48: Differences of Path losses at 25 m and 3m (ITU-R P.1546-5 Land)
ITU-R P.1546-5 (Land) f=3000 MHz

Diff
D (km) Hrx=25m Hrx=3m
10 142.5 165.61 23.11

20 159.1 182.31 23.21
30 169.6 192.02 22.42
40 176.8 199.09 22.29
50 180.63 203.28 22.65
60 183.06 205.43 22.37
Based on the calculation results in Table A5-9 and A5-10, it is proposed to use a conversion factor of 22 dB
between 25 m and 3 m.
A5.2.8 CALCULATION OF TRIGGER VALUES
The relation between field strength E (dBuV/m) and power level Pr (dBm) can be expressed as:
E=Pr+20*log10(F)+77,2 (3)
Where F is the frequency in MHz.
Pr=Ptx+G1 - PL (4)
Pr is the received power level at the receiving BS antenna (before antenna), Ptx is the transmit power before
antenna, G1 is interfering BS antenna gain including feeder loss in the direction of the receiving antenna, PL
is the path loss at the distance D.
Using equation (3) and (4), the calculated trigger value at 25 m and 3 m are summarised in Table 49 and
Table 50 for co-channel case and in Table 51 for adjacent channel case.
Table 49: Calculated trigger value at 25 m and 3 m for co-channel case with ITU-RP.2001-2
Rx Antenna Height (m) Ptx (dBm) G1 (dB) D (km) PL (dB) Pr (dBm) E (dBuV/m/20 MHz)
25 49 2.4 49 163.8 -112.4 35.9

3 49 2.4 49 186.8 -135.4 12.9
Table 50: Calculated trigger value at 25 m and 3 m for co-channel case with ITU-R P.2001-2
25 49 2.4 44 156.2 -104.8 43.5

3 49 2.4 44 179.2 -127.8 20.5
The results in Table 49 correspond the protection ratio of 50% reference cell throughput loss. The results in
Table 50 correspond the protection ratio of I/N=-6 dB at reference cell BS.
Table 51: Calculated trigger value at 25 m and 3 m for adjacent channel case with ITU-R P.2001-2
25 49 2.4 12 125.2 -73.8 74.5

3 49 2.4 12 128.2 -76.8 71.5
In Table 51, the trigger value for 3 meters receiving antenna height is calculated with the propagation model
ITU-R P.2001-2. When using the conversion factor of 22 dB, the field strength level at 3m height is 74.5-22 =
52.5 dBuV/m/20 MHz.
A5.2.9 Conclusions from results
 ECC Recommendation (15)01 may be used to deal with the case of two unsynchronised macro-cellular
networks within a given country, when the physical borderline is defined between two networks within a
country.
 The required separation distance and trigger values calculation depend many elements:
o Cellular network topology (LTE-TDD or 5G-NR, non-AAS or AAS, BS antenna height,
environment or cell range);
o Propagation environment and propagation model;
o Frequencies and overlap of the channels, e.g. full overlap as co-channel case, or partial
overlap or adjacent channel;
 Protection ratio, e.g. I/N, or network throughput loss at x%, etc. As an example, the simulated results
show that in a co-channel case the required separation distance is about 50 km and in an adjacent
channel case the required separation distance is between 12 and 15 km;
 A conversion factor of 22 dB can be used for the field strength conversion between 25 m and 3 m. For
different antenna height the conversion factor is different.
ANNEX 6: COEXISTENCE STUDIES BETWEEN UNSYNCHRONISED MICRO BSS AND MACRO BSS -
STUDY #5 AND #6
A6.1 STUDY #5
A6.1.1 MACRO BS VS. MICRO BS NETWORKS
A6.1.1.1 Assumptions
The Macro BS vs. Micro BS scenario models the interference between one building and a hexagonal macro-
cellular network.
Macro-cellular network assumptions
The assumptions used for the macro-cellular network are identical to the Macro BS - Indoor BS study in
section 2.
The only difference is that in this study there is no building area that should be avoided when dropping the
Macro users.
Micro BS network assumptions
Table 52: BS parameters for the Micro BS network
BS parameters Value Source / reference
BS Element gain 8 dBi
BS Antenna array 8x8 array
BS Element spacing (Horizontal, Vertical) 0.5 λ , 0.5 λ
Beamforming towards MSs

BS antenna beamforming
with (8´8) array
C-band synch toolbox –

BS noise figure 8 dB studies and assumptions
v0.11
C-band synch toolbox –

BS transmit power 40 dBm studies and assumptions
v0.11
BS mechanical down tilt 10 deg. ECC Report 281 [2]

Table 53: MS parameters for the Micro BS network
MS parameters Value Source / reference

MS antenna type Isotropic, -4 dBi ITU-R M.2292-0 [23]
Correspondence group
Power control: CLxile 94 dB
discussions
Correspondence group
Power control: gamma 1
discussions
Table 54: System related parameters for the Micro BS network
e Value
Channel bandwidth 60 MHz
Number of sites 19
1 sector/site (Random
Number sectors per site
orientation)
MS antenna height 1.5 m
MS distribution Uniform over the area
Fraction of MSs that are indoor 70%
Table 55: Propagation parameters for propagation in the Micro BS network
Propagatione Value Source / reference
Propagation model (BS-MS within the

UMi 3GPP TR 38.901 [24]
system)
Indoor penetration model parameters

According to Table 7.4.3-3 3GPP TR 38.901 [24]
for Macro MS
Macro-cellular network vs. Micro BS network parameters
Table 56: Parameters related to inter-system deployment
System parameters Value
Macro BS vs. Micro BS distance 30m (min. distance)

Table 57: Propagation parameters for inter-system propagation
Propagation Value Source / reference
Propagation model Macro BS vs.

UMa 3GPP TR 38.901 [24]
Micro BS
Propagation model Macro MS vs.
ITU-UMi
indoor MS
Figure 62shows the one snapshot of the deployment of the micro and macro-cellular networks. The centre
BS in the macro-cellular network is located 30m from the centre BS in the Micro BS network.
Figure 62: Deployment of micro and macro-cellular networks
A6.1.1.2 Simulation results – Impact on micro BS network
The impact on the throughput loss in the Micro BS network is shown in Figure 63 to Figure 65. In all figures
the results are averaged over many different snapshots of MS locations.
In Figure 63, the results are averaged over all Micro BSs for several realisations of the Macro BS - Micro BS
propagations. In Figure 64Figure 63: Mean uplink throughput loss for the Micro BS network. , separate
curves are shown for each realisation of Micro BS - Macro BS propagation. Finally, in Figure 65 only the
results from the centre BS are shown.
100
90
80
70
Throughput Loss (%)
60
50
40
30
20
10
5
0
0 20 40 60 80 100
ACIR in dB
Figure 63: Mean uplink throughput loss for the Micro BS network. Throughput loss averaged over
different Macro BS – Micro BS propagation realisations and the interfering Macro BS serving
different users
100
90
80
70
Throughput Loss (%)
60
50
40
30
20
10
5
0
0 20 40 60 80 100
ACIR in dB
Figure 64: Mean uplink throughput loss for the Micro BS network. Throughput loss averaged over
many realisations of Macro BS serving different users
100
90
80
70
Throughput Loss (%)

60
50
40
30
20
10
5
0
0 20 40 60 80 100
ACIR in dB
Figure 65: Mean uplink throughput loss for the Micro BS in the centre (worst case). Throughput loss
averaged over many realisations of Macro BS serving different users
A6.1.1.3 Simulation results - Impact on macro-cellular networks
Figure 66 shows the throughput loss vs. ACIR for the BS in the centre of the Micro BS network. I.e. the BS
most impacted by the Micro BS network for different realisations of Micro BS to macro-cellular networks
propagation.
Figure 67 depicts the averages the throughput loss vs. ACIR for all the Macro BSs and realisations of Micro
BS to Macro BS propagation, while Figure 68 shows the throughput loss for all Macro BSs for each
realisation of Micro BS to Micro propagation.
100
90
80
70
Throughput Loss (%)
60
50
40
30
20
10
5
0
0 20 40 60 80 100
ACIR in dB
Figure 66: Average uplink throughput loss for the macro-cellular network. Throughput loss averaged
over different Macro BS - Micro BS propagation realisations and the interfering Micro BS serving
different users
100
90
80
Throughput Loss (%) 70
60
50
40
30
20
10
5
0
0 20 40 60 80 100
ACIR in dB
Figure 67: Average uplink throughput loss for the macro-cellular network. Throughput loss averaged
over many realisations of Micro BS serving different users
100
90
80
70
Throughput Loss (%)
60
50
40
30
20
10
5
0
0 20 40 60 80 100
ACIR in dB
Figure 68: Average uplink throughput loss for the Macro BS in the centre (worst case). Throughput
loss averaged over many realisations of Micro BS serving different users
A6.2 STUDY #6
This study considers the impact of BS-BS interference between MFCNs with simultaneous UL/DL
transmission in terms of the resulting degradation in the mean UL throughput of the victim MFCN. The
MFCNs consist of Macro and Micro BSs.
The study addresses two scenarios according to the specific class of base stations, namely:
 Macro-cellular network (hexagonal grid placed outdoors) operates as the interferer and the Micro BS
network (hexagonal grid placed outdoors) is interfered;
 Micro BS network(hexagonal grid placed outdoors) operates as the interferer and the Macro BS (placed
outdoors) is interfered;
 Interference from one Micro BS to another Micro BS (both base stations are placed outdoors)
The two interfering deployments operate in the same geographic area on adjacent frequency channels.
All Micro and Macro BSs are assumed to be AAS base stations forming a beam towards a MS (MSs are
assumed to be uniformly distributed within a cell).
The Macro BSs have 25m high antennas and comprise three sectors per site; the Micro BSs are placed 6
meters above ground, comprising one sector per site with random boresight.
See Section A6.2.3 for the full list of assumptions and parameters.
A6.2.1 INTERFERENCE BETWEEN MACRO-CELLULAR NETWORKS AND MICRO BS NETWORKS
Figure 69 provides the topology used for the coexistence studies in case of a macro-cellular network (placed
outdoors) operating as the interferer towards a Micro BS network (hexagonal grid placed outdoors).
Figure 69: Topology for the Macro BS to Micro BS interference scenario
Figure 70 provides the topology used to support coexistence studies in case of a Micro BS network
(hexagonal grid placed outdoors) operating as the interferer towards the macro-cellular network (hexagonal
54
grid placed outdoors). In line with ECC Report 203 , the simulations address one Macro BS that is
completely surrounded by the Micro BSs grid.
Figure 70: Topology for the Micro BS to Macro BS interference scenario
A6.2.2 INTERFERENCE FROM AAS MICRO BSS TO AAS MICRO BSS
Approach 1: In this analysis, the separation distance between the Micro BSs is an input parameter, the UMi
path loss model determines the associated Line-of-sight probability.
The following two settings have been considered:

 Case 1a: 30m separation distance between the two Micro BSs leading to 80% LoS probability based on
the UMi path loss model (the smaller the distance, the greater the probability the two Micro BSs will be
along the same street),
Figure 71: Micro BS to Micro BS interference scenario topology - Case 1a:

 Case 1b: 100m separation distance between the two Micro BSs leading to 25% LoS probability based on
the UMi path loss model (the larger the distance, the greater the probability the two Micro BSs will be
located in different streets),
54
ECC Report 203 [5] page 26: “One important thing to note here is that the results contained in Table 17 are for one reference cell in
the macro-cellular network, which is overlapped completely by the Micro BS network (Manhattan) grid (see Figure 19).”
Figure 72: Micro BS to Micro BS interference scenario topology - Case 1b:

Approach 2: in this analysis the separation distance between the Micro BSs is an input parameter as
well as the line-of-sight probability.
This approach accounts for the fact that it is difficult to carry out meaningful simulations to assess
the interference between two Micro BS networks in the same urban area since the interference
scenario will be strongly impacted by the LoS/NLoS conditions which radically change as the Micro
BSs change their locations with respect to buildings.
The study therefore considers two specific set of cases for the deployments of the interfering and
victim base stations:
 Cases 2a, 2b and 2c: two Micro BSs located in different streets at 30m, 50m and 75m separation
distance with 0% LoS probability
Figure 73: Micro BS to Micro BS interference scenario topology. Case 2a, 2b and 2c:
30m separation distance and 0% LoS probability (different streets)
 Case 2d: two Micro BSs located along the same street (100% LoS probability) at 100 m separation
distance.
Figure 74: Micro BS to Micro BS interference scenario topology. Case 2d: 100m separation
distance and 100% LoS probability (same street)
A6.2.3 SIMULATIONS PARAMETERS
The assumptions used the basis for the coexistence studies are as in Table 33 in ANNEX 4, with the
following options used:
Table 58: Simulation Parameters
Parameter Value
MACRO-CELLULAR NETWORK DEPLOYMENT PARAMETERS
Scheduler (# active MSs / cell) Proportional Fairness scheduling Option 1 in Table 33
# active MSs / BS 3 active MSs / CELL Option 1 in Table 33
MICRO DEPLOYMENT PARAMETERS

#sectors per site 1 cell/site Option 1 in Table 33
BS antenna tilt 10 degrees Option 1 in Table 33
MS PARAMETERS
NOTE: parameters are chosen to maximise the network performance: maximise
the 5% and average t-put network performance (for a single operator)
UL Power Control
Macro MS: P0 = -92dBm, ɣ = 0.8
Option 2 in Table 33
Micro MS: P0 = -86dBm, ɣ = 0.8
MS distribution
80% indoor
Macro ↔ Macro (aligned with RAN1 3GPP 38.802 table Option 1 in Table 33
8.2.1-1)
Macro MSs:
Macro ↔ Micro Option 1 in Table 33
- Indoor: 70 %, outdoor: 30 %
A6.2.3.1 MS Power Control parameters selection
As shown in Table 58, the following MS power control parameters are chosen for the MSs connected to the
Micro BS network:
P0 = -86 dBm , ɣ = 0.8
This section shows how the selected parameters maximise the network performance in terms of the average
uplink throughput and on the 5% edge throughput loss (for a single operator).
Impact on the average uplink throughput:
7
Figure 75: Relationship between Po and average uplink throughput, ɣ = 0.8
Impact on the cell edge uplink throughput loss:

Figure 76: Relationship between Po and edge UL average throughput, ɣ = 0.8
The 5G MFCN antenna element and array parameters were as in Table 38 in ANNEX 5.
A6.2.4 SIMULATIONS RESULTS
The following sections present the simulations results expressed in terms of degradation of the mean uplink
throughput of the victim MFCN due to base station to base station interference from the interfering MFCN,
presented as a function of ACIR. In general terms, as expected, the impact of interference on network
performance diminishes with increasing values of ACIR.
Note that the required ACLR is assumed to be nominally equal to the required ACIR, with the understanding
that interference is not dominated by the adjacent channel selectivity (ACS) of the victim base station.
Note that both victim BS and interferer base stations are assumed to operate with 60 MHz channel
bandwidth.
It is important to highlight that the following were not accounted for:

 blocking effect on the victim BS receiver;
 MS – MS interference.
A6.2.4.1 Macro-cellular network interferes with Micro BS network
The results presented in this section refer to the topology proposed in Figure 69.
Figure 77 shows how an ACIR greater than 68 dB is required to ensure a mean uplink throughput
Figure 77: Impact of the macro-cellular network interference to Micro BS network

In terms of mean uplink throughput
A6.2.5 MICRO BS NETWORK INTERFERES WITH MACRO-CELLULAR NETWORK
The results presented in this section refer to the topology proposed in Figure 21.
Figure 78: Impact of the Micro BS network interference to macro-cellular network

In terms of mean uplink throughput
A6.2.5.1 Micro BS interferes with another micro BS
The results presented in Figure 79 refer to the topology proposed in Figure 71.
Figure 79 shows how an ACIR greater than 63dB is required to ensure a mean uplink throughput
Figure 79: Micro BS to Micro BS interference scenario simulation results - Case 1a:
Figure 80: shows how an ACIR greater than 54dB is required to ensure a mean uplink throughput
Figure 80: Micro BS to Micro BS interference simulations results - Case 1b:

Figure 81 shows how an ACIR greater than 49 dB is required to ensure mean uplink throughput degradation
smaller than 5% for 30 m separation distance. If separation distance is 50 m, 45 dB ACIR can satisfy the
requirement of 3GPP.
Figure 81: Micro BS to Micro BS interference scenario simulations results – Case 2a, 2b and 2c:
30 m, 50 m and 75 m separation distance and 0% LoS probability (different streets)
Figure 82: Micro BS to Micro BS interference scenario simulations results – Case 2b:
100 m separation distance and 100% LoS probability (same street)
ANNEX 7: COEXISTENCE STUDY BETWEEN UNSYNCHRONISED INDOOR BS AND MACRO BASE

STATIONS - STUDY #7
A7.1 ASSUMPTIONS
A7.1.1 MACRO BS VS. INDOOR BS
The Macro BS vs. Indoor BS scenario models the interference between one building and a hexagonal
macro-cellular network.
Macro-cellular network assumptions
Table 59: BS parameters for the macro-cellular network
BS parameter Value Source / reference

BS Element spacing (Horizontal, vertical) 0.5 λ , 0.5 λ
with (8´8) array
BS transmit power 51 dBm
BS mechanical down tilt 10 deg. ECC Report 281 [2]
Table 60: MS parameters for the macro-cellular network
MS parameter Value Source / reference

Power control: CLxile 94 dB
Power control: gamma 1
Table 61: System related parameters for the macro-cellular network
System parameters Value Source / reference
Channel bandwidth 60 MHz

Number of sites 19
Number sectors per site 3 sectors/site
BS antenna height 25 m
MS antenna height 1.5 m
Uniform over the area (excluding the

area of the building).
MS distribution NOTE: Uniform distribution is commonly
used, but how the area occupied by the
building is handled should be discussed.
Fraction of MSMS that are indoor 80%
Table 62: Propagation parameters for propagation in the macro-cellular network
e Value Source / reference
Propagation model
UMa 3GPP TR 38.901 [24]
(BS-MS within the system)
Indoor penetration model parameters
According to Table 7.4.3-3 3GPP TR 38.901 [24]
for Macro MS
Indoor system assumptions
Table 63: BS parameters for the Indoor system
BS parametere Value Source / reference

with (4x4) array
BS transmit power 24 dBm 3GPP TR 36.873 [26]
BS mechanical down tilt 90 deg. (ceil mount) 3GPP TR 36.873 [26]
Table 64: BS parameters for the Indoor system
MS parameters Value Source / reference

Power control: CLxile 94 dB Correspondence group discussions
Power control: gamma 1 Correspondence group discussions
Table 65: System related parameters for the Macro BS network
Building layout 1 floor 50x120 m building 3GPP TR 36.873 [26]

Number of sites 2, ceiling mount 3GPP TR 36.873 [26]

Number sectors per site 1 panel/site 3GPP TR 36.873 [26]
Inter-Site Distance (ISD) 60 m 3GPP TR 36.873 [26]
BS antenna height 3m 3GPP TR 36.873 [26]
MS antenna height 1m 3GPP TR 36.873 [26]
MS distribution Uniform over the floor
Fraction of MS that are indoor 100%
50
40
30
m
20
10
0
0 20 40 60 80 100 120
m
Figure 83: Building and the Indoor BS layout
Table 66: Propagation parameters for propagation in the macro-cellular network
Propagation model indoor BS-MS InH - Office 3GPP TR 38.901 [24]

LoS probability model Indoor - Open office 3GPP TR 38.901 [24]
Macro-cellular network vs. indoor BS network parameters
Table 67: Parameters related to inter-system deployment
Macro BS vs. building distance (outer wall) 70 m

Case 1: Short wall facing BS
Building orientation
Case 2: Long wall facing BS
Table 68: Propagation parameters for inter-system propagation
Propagation model Macro BS vs. indoor BS UMa 3GPP TR 38.901 [24]

Propagation model Macro MS vs. indoor MS UMa 3GPP TR 38.901 [24]
Two cases of building orientation are studied. One when the building has the short wall toward the BS and
one when the BS has the long wall toward the BS.
400
200
-200
-400
-600
-600 -400 -200 0 200 400 600 800
Figure 84: Case 1: Short wall facing the BS
400
300
200
100
-100
-200
-300
-400
-500
-600
-400 -200 0 200 400 600
Figure 85: Case 2: Long wall facing the BS
A7.1.2 SIMULATION RESULTS – IMPACT ON INDOOR NETWORKS
The results for Case 1 where the short edge of the building is 70 m away from the Macro BS and oriented
such that the boresight of the antenna beam is towards the short edge of the building, is shown in Figure 84
and Figure 85. In both figures the throughput loss is averaged over many realisations of MS locations and
consequently the direction of the interfering BS beam.
In Figure 86 the results are also averaged over several realisations of the outdoor-to-indoor channel model,
while in Figure 87 each realisation is plotted individually.
The corresponding results for Case 2 when the long edge of the building is facing the outdoor Macro BS are
shown in Figure 88 and Figure 89 respectively. The performance is slightly worse in this case. The reason is
that in this case there are two BS relatively close to the victim, while in case 1 the other BS is farther away.
100
90
80
70
Throughput Loss (%)
60
50
40
30
20
10
5
0
0 20 40 60 80 100
ACIR in dB
Figure 86: Average uplink throughput loss for the Indoor network in Case 1. Throughput loss
averaged over different O2I channel realisations and the interfering Macro BS serving different users
100
90
80
70
Throughput Loss (%)
60
50
40
30
20
10
5
0
0 20 40 60 80 100
ACIR in dB
Figure 87: Uplink throughput loss for the Indoor network in Case 2. Throughput loss in each curve
averaged over many realisations of the Macro BS serving different users
100
90
80
70
Throughput Loss (%)
60
50
40
30
20
10
5
0
0 20 40 60 80 100
ACIR in dB
Figure 88: Average uplink throughput loss for the Indoor network in Case 2. Throughput loss
averaged over different O2I channel realisations and the interfering Macro BS serving different users
100
90
80
70
Throughput Loss (%)
60
50
40
30
20
10
5
0
0 20 40 60 80 100
ACIR in dB
Figure 89: Uplink throughput loss for the Indoor network in Case 2. Throughput loss in each curve
averaged over many realisations of the Macro BS serving different users
A7.1.3 IMPACT ON MACRO BS NETWORK
The reverse case where the macro-cellular network is the victim has not been simulated. However it can be
observed that the indoor system has lower output power, which means that there should be lower impact
from the indoor system. On the other hand if there are several buildings with indoor systems deployed there
is a need to consider the effect of the aggregate interference.
ANNEX 8: COEXISTENCE BETWEEN SEMI-SYNCHRONISED MFCN NETWORKS - STUDY #8
Simulation results are presented, focusing on the following scenarios:

 Macro BS to macro-cellular networks:
 Fully unsynchronised operation: Operator A and Operator B have simultaneous UL/DL transmissions
for the whole frame duration. This is the worst-case scenario that becomes realistic only when the
operators choose to have unsynchronised duplex directions for the whole frame duration;
 Semi-synchronised operation in case only 50% of the frame is designated for flexible operation and
operators have unsynchronised duplex directions in all flexible slots;
operators have unsynchronised duplex directions in all flexible slots;
operators have unsynchronised duplex directions in 50% of the flexible slots - in the following such a
case will be referred to as 10% unsynchronised operation.
 Micro BS to Micro BS networks:
 Same semi-synchronised operation cases as for Macro BS to macro-cellular networks.
A8.1 SIMULATION ASSUMPTIONS
Simulations were performed based on the set of assumptions in Table 33 with the exception of the
parameters specified in Table 69 for macro-cellular network to macro-cellular network and Table 70 for Micro
BS network to Micro BS network.
They are consistent with several coexistence studies in 3400-3800 MHz band (e.g.3GPP TR 36.942 [27],
ECC Report 203 [5]).
Table 69: Macro-cellular network to macro-cellular network simulation assumptions
Parameter Value
Deployment
Networks shift 288 m

BS / MS parameters
PC Set 1 in 3GPP TR 36.942
Uplink Power Control
CLxile = 94 dB, Gamma = 1
Table 70: Micro BS network to Micro BS network simulation assumptions
Parameter Value
BS / MS parameters
PC Set 1 in 3GPP TR 36.942
Uplink Power Control
CLxile = 82 dB, Gamma = 1
Uplink power control settings are derived from 3GPP TR 36.942 and scaled to account for a different Carrier
Frequency, Channel bandwidth and deployment scenario.
With reference to the assumptions above;

 As shown in Figure 15, the 288 m network shift assumption between macro-cellular networks represents
a best case assumption. Similarly, the 96 m network shift assumption between micro-cellular networks
represents best-case assumptions.
 Differently from the approach followed in this study, the recommended approach is to use the separation
distance and the line-of-sight probability as input parameter during the coexistence studies between the
Macro-cellular network and the Micro BSs network. This approach accounts for the fact that it is difficult
to carry out meaningful simulations to assess the interference between two Micro BS networks in the
same urban area since the interference scenario will be strongly impacted by the LoS/NLoS conditions
which radically change as the Micro BSs change their locations with respect to buildings.
 Coexistence between the macro-cellular network and the Micro BS network was not assessed by this
study.
Antenna radiation pattern is as in 3GPP TR 38.901 [24] and is shown in Table 35 of ANNEX 4.
A8.2 SIMULATION RESULTS
This Section presents simulation results for the semi-synchronised operation scenarios listed above for the
macro-cellular network to macro-cellular network case and for Micro BS network to Micro BS network
deployment case.
A8.2.1 MACRO BS TO MACRO-CELLULAR NETWORKS
Figure 90 below shows the impact of Adjacent Channel Interference (ACI) on victim network performance in
terms of average throughput loss for macro-cellular network to macro-cellular network. As expected, the
impact diminishes when the operators have unsynchronised duplex directions for smaller portion of the
frame.
Results show that with the baseline requirement for synchronised MFCNs in ECC Report 281 [2]
performance degradation is ~9% for 10% unsynchronised operation among operators. It is important to
notice that results are preliminary and do not consider any interference mitigation technique that would likely
bring degradation down.
ACI average throughput loss - Macro-to-macro

100
Fully unsynchronized
90
50% unsynchronized
20% unsynchronized
80 10% unsynchronized
70
60
Loss [%]
50
40
30
20
10
0
35 40 45 50 55 60 65 70 75 80 85 90
ACIR [dB]
Figure 90: ACI impact on network performance - average throughput loss for Macro BS to Macro BS
deployment and different semi-synchronised operation cases
A8.2.2 MICRO BS TO MICRO BS NETWORK DEPLOYMENT
Figure 91 shows the impact of Adjacent Channel Interference (ACI) on victim network performance in terms
of average throughput loss for Micro BS to Micro BS network deployment. As expected, the impact
diminishes when the operators have unsynchronised duplex directions for smaller portion of the frame.
Results show that it is possible to achieve 5% average throughput loss with 38 dB ACIR in the case the two
operators have unsynchronised duplex directions for 20% of the frame. In this case it will be possible to use
the current baseline requirement in ECC Report 281 [2] for synchronised BSs without additional throughput
degradation.
It is again important to notice that results are preliminary and do not consider any interference mitigation
technique that would likely bring degradation down.
ACI average throughput loss - Micro-to-micro

100
Fully unsynchronized
90 50% unsynchronized
20% unsynchronized
80
70
60
Loss [%]
50
40
30
20
10
0
25 30 35 40 45 50 55 60 65 70
ACIR [dB]
Figure 91: ACI impact on network performance - average throughput loss for Micro BS to Micro BS
deployment and different semi-synchronised operation cases
ANNEX 9: STATUS OF UNSYNCHRONISED AND SEMI-SYNCHRONISED OPERATION IN 3GPP
The 5G-NR core RF requirements in 3GPP Release 15 were derived with the assumption of synchronised
operation between two TDD systems. This follows the same approach adopted during LTE-TDD
requirements specification.
During the course of Rel-14 5G-NR SI and Rel-15 5G-NR WI, 3GPP TSG RAN4 has not studied the 5G-NR
RF requirements that can be applied for unsynchronised or semi-synchronised operation between two TDD
systems on adjacent channels, where simultaneous UL/DL transmissions occur. However, RAN4
specifications allow the possibility to support unsynchronised or semi-synchronised operation across
adjacent channels and/or operators, e.g. when either the two TDD systems are sufficiently isolated or some
interference mitigation schemes are deployed. Furthermore, 5G-NR has the design flexibility to operate with
full bandwidth in some time instances and narrower bandwidth in other time instances. Therefore, this may
allow simultaneous transmissions to be operated instantaneously with a narrower bandwidth to provide
isolation in frequency from adjacent channels. This type of operation has some restriction however, and will
lead to spectrum efficiency impacts during those simultaneous UL/DL feedback on the amount of support in
Release 15 5G-NR specifications for the unsynchronised or semi-synchronised operations described in the
ECC Report 281. In the following, synchronisation between two networks or two BSs means the same
beginning of the slot and the alignment of transmission directions (DL, UL), as in the definitions from CEPT.
3GPP TSG RAN1 has taken a general approach when specifying 5G-NR, enabling support for synchronised,
semi- and unsynchronised deployment. It is RAN1 understanding that there is no coexistence issue due to
adjacent channel interference among TDD networks in case of synchronised operation of multiple TDD
networks on adjacent channels.
Unsynchronised or semi-synchronised 5G-NR operation may occur in the context of two scenarios:
Operators choose to use semi-statically configured DL / UL partitioning but use different DL / UL patterns
One or more operators choose to use dynamic DL / UL partitioning
The air-interface specifications developed by RAN1 support that each slot can be dynamically scheduled to
transmit on either uplink or downlink, or the slot could include both a DL part and an UL part. Slots (and
symbols within a slot) can be semi-statically configured to be UL or DL or an ‘undefined’ state that can be
dynamically allocated to UL or DL. In the presence of the semi-static configuration, the dynamic allocation
applies only to the ‘undefined’ part, while the slots and symbols indicated as DL or UL can only be used in
the indicated duplex direction.
The air-interface specifications developed by RAN1 allow adjusting the bandwidth occupied by the
modulated waveform (note that LTE-TDD is not capable of this since there are always-on wideband common
reference signals in LTE-TDD). More specifically, 5G-NR can adapt by scheduling the DL and UL
bandwidths occupied by physical channels and signals, and those bandwidths can be different in different
symbols and between DL and UL.
RAN1 will not determine what would be the exact conditions for 5G-NR to allow synchronised,
unsynchronised and semi-synchronised operation. The current available tools that RAN1 sees to mitigate
BS-BS and MS-MS interference are sufficient guard bands, sufficient geographical separation, sufficient
physical isolation (such as outdoor to indoor propagation isolation), or applicable transmission power.
3GPP in the past has studied inter-operator coexistence for LTE-TDD on adjacent channels [28][29]. The
study in [28] concluded that significant BS-BS coexistence challenges have been observed to apply different
TDD UL-DL configurations in different cells without any interference mitigation mechanisms for scenario 7,
which represents inter-operator coexistence on adjacent channels in macro-cellular networks. It was also
noted that in Macro BS to Macro BS coexistence (scenario 7 in [28] ), only the uplink exhibited degradation.
There was no study in [28] or conclusion on the coexistence feasibility with interference mitigation
mechanisms for scenario 7. The study in [28]was completed at RAN#56 (June 2012), and did not include
other techniques specified subsequently, e.g., AAS. There has been no further studies on coexistence
feasibility of inter-operator MBS Macro BS deployments with unsynchronised or semi-synchronised operation
on adjacent channels and using the techniques specified since June 2012, e.g., AAS.
When LTE-TDD is present and used in the same band as 5G-NR then, depending on the scenario such as
large cell vs. small cell deployments or geographical separation, sufficient physical isolation, restrictions on
the transmission directions may be necessary between the LTE-TDD and 5G-NR carriers to avoid the use of
fixed guard bands. Likewise, similar restrictions on the transmission directions of 5G-NR Macro BSs would
be necessary between neighbour 5G-NR networks in adjacent frequencies, although the constraints in terms
of DL and UL patterns might be different than for coexistence with LTE-TDD.
However, RAN1 believes that such restrictions on the transmission directions are only required for certain
types of deployments such as Macro BSs, and may be relaxed in more deployments with appropriate cross-
link interference mitigation techniques. RAN1 believes that enabling 5G-NR deployments in the above three
types of operation (either in earlier or later deployments) is important for achieving the full potential of 5G-NR
TDD in areas/scenarios where sufficient conditions are met to mitigate 5G-NR BS-BS and 5G-NR MS-MS
interference.
3GPP TSG RAN has approved a Release 16 work item [30] on cross-link interference handling and Remote
Interference Management (RIM) for 5G-NR, which is relevant to deployments where interference occurs
between uplink and downlink of two networks. The target completion date of the work item is June 2019. The
work item should specify cross-link interference mitigation techniques to support flexible resource adaptation.
The detailed objectives for cross-link interference mitigation to support flexible resource adaptation for
unpaired 5G-NR cells are:
 Specify cross-link interference measurements at a MS (e.g., CLI-RSSI and/or CLI-RSRP) (RAN1 and
RAN4);
 Identify when cross-link interference mitigation techniques based on such measurement(s) provide
benefits with practical RF performance (RAN4);
 Specify network coordination mechanism(s) including at least exchange of intended DL/UL configuration
(RAN and RAN3);
 Perform coexistence study to identify conditions of coexistence among different operators in adjacent
channels (RAN4);
 Target no or very minimal impact on RF requirement.
ANNEX 10: LIST OF REFERENCES
[1] ECC Report 216: "Practical guidance for TDD networks synchronisation", August 2014
[2] ECC Report 281: “Analysis of the suitability of the regulatory technical conditions for 5G MFCN
operation in the 3400-3800 MHz band”, July 2018
[3] ECC Decision (11)06: "Harmonised frequency arrangements and least restrictive technical conditions
(LRTC) for mobile/fixed communications networks (MFCN) operating in the band 3400-3800 MHz",
amended October 2018
[4] ECC Recommendation (15)01: “Cross-border coordination for mobile / fixed communications networks
(MFCN) in the frequency bands: 694-790 MHz, 1452-1492 MHz, 3400-3600 MHz and 3600-3800 MHz,
March 2014
[5] ECC Report 203: "Least Restrictive Technical Conditions suitable for Mobile/Fixed Communication
Networks (MFCN), including IMT, in the frequency bands 3400-3600 MHz and 3600-3800 MHz",
November 2013
[6] ECC Report 131: "Derivation of a Block Edge Mask (BEM) for terminal stations in the 2.6 GHz frequency
band (2500-2690 MHz)", February 2009
[7] ITU-R M.2374: “Coexistence of two time division duplex networks in the 2 300-2 400 MHz band”
[8] 3GPP TS 38.104
[9] ECC Report 287: “Guidance on defragmentation of the frequency band 3400-3800 MHz”, October 2018
[10] Tender Document F 7/16: “Procedure for Spectrum Award in the 3410 to 3800 MHz range”, Telekom-
Control-Kommission Austria, September 2018: https://www.rtr.at/en/tk/5G-Auction-Tender-
Documents/Tender_Documents_3_4_-_3_8_GHz_EN.pdf
[11] ComReg 17/46: “Results of the 3.6 GHz Band Spectrum Award – Information Notice”, June 2017:
https://www.comreg.ie/publication/results-3-6-ghz-band-spectrum-award-2/
[12] AGCOM (Autorità per le Garanzie nelle Comunicazioni) Resolution No. 231/18/CONS – May ‘18
“Assignment procedures and rules for the use of the frequencies available in the 694-790 MHz, 3600-
3800 MHz and 26.5-27.5 GHz bands for terrestrial systems of electronic communications in order to
favor the transition to 5G technology, under the Law 27 December 2017, n. 205”;
[13] void
[14] Ofcom “Public Sector Spectrum Release: Award of the 2.3 and 3.4 GHz spectrum bands – Statement
and consultation”, May 2015:
https://www.ofcom.org.uk/__data/assets/pdf_file/0027/68337/Public_Sector_Spectrum_Release_statem
ent.pdf
[15] Ofcom “Award of 2.3 and 3.4 GHz spectrum bands - Publication under regulation 111 of the Wireless
Telegraphy (Licence Award) Regulations 2018 of results of auction” ”, April ‘2018
[16] 3GPP TS 36.211: “3rd Generation Partnership Project; Technical Specification Group Radio Access
Network; Evolved Universal Terrestrial Radio Access (E-UTRA); Physical Channels and Modulation”
[17] 3GPP TS 38.213 “3rd Generation Partnership Project; Technical Specification Group Radio Access
Network; NR; Physical layer procedures for control (Release 15)” section 11.1
[18] Report ITU-R M.2410-0 (11/2017): “Minimum requirements related to technical performance for IMT-
2020 radio interface(s)”
[19] Report ITU-R M.2412-0 (10/2017): “Guidelines for evaluation of radio interface technologies for IMT-
2020”
[20] 3GPP TS 38.214 V15.0.0, “NR; Physical layer procedures for data (Release 15)”, December, 2017.
[21] Recommendation ITU-R P.452-16: "Prediction procedure for the evaluation of interference between
stations on the surface of the Earth at frequencies above about 0.1 GHz"
[22] Recommendation ITU-R P.2001-2: "A general purpose wide-range terrestrial propagation model in the
frequency range 30 MHz to 50 GHz"
[23] Report ITU-R M.2292-0: "Characteristics of terrestrial IMT-Advanced systems for frequency
sharing/interference analyses"
[24] 3GPP TR 38.901: "Study on channel model for frequencies from 0.5 to 100 GHz"
[25] 3GPP TR 38.802: "Study on new radio access technology Physical layer aspects"
[26] 3GPP TR 36.873: "Study on 3D channel model for LTE"
[27] 3GPP TR 36.942: "Evolved Universal Terrestrial Radio Access (E-UTRA); Radio Frequency (RF)
system scenarios"
[28] 3GPP TR36.828: “Further enhancements to LTE Time Division Duplex (TDD) for Downlink-Uplink (DL-
UL) interference management and traffic adaptation, Rel-11”
[29] 3GPP TR36.825: “Feasibility study on possible additional configuration for LTE Time Division Duplex
(TDD), Rel-13”
[30] 3GPP RP-181431: “New WID on Cross Link Interference (CLI) handling and Remote Interference
Management (RIM) for NR”, LG Electronics, June 2018

ECC Report 296

Uploaded by

Copyright:

Available Formats

ECC Report 296

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

ECC Report 296

Uploaded by

Copyright:

Available Formats

ECC Report 296

National synchronization regulatory framework options in

Approved 8 March 2019

Section 2 provides a general introduction about synchronised, semi-synchronised and unsynchronised

As a consequence, and considering that:

0 Executive summary ................................................................................................................................. 2

2 Synchronisation modes for TDD networks coexistence ................................................................... 12

3 Performance impacts from frame structure selection ....................................................................... 22

5 Toolbox: options for administrations to support the design of synchronisation frameworks at

ANNEX 1: Synchronisation frameworks in recent C-band award procedures........................................ 59

ANNEX 2: Technical solutions for network synchronisation.................................................................... 62

ANNEX 4: MFCN parameter values and assumptions for simulations .................................................... 91

ANNEX 8: Coexistence between semi-synchronised MFCN networks - Study #8................................ 131

ANNEX 9: Status of unsynchronised and semi-synchronised operation in 3GPP ............................... 134

ANNEX 10: List of References .................................................................................................................... 136

3GPP 3rd Generation Partnership Project

This Report provides:

2 SYNCHRONISATION MODES FOR TDD NETWORKS COEXISTENCE

Figure 1: Interference scenarios in case of simultaneous UL/DL transmissions in MFCN TDD

2.1 SYNCHRONISED OPERATION

Figure 2: Frequency, phase and time synchronisation concepts

Technology Frequency accuracy

GSM, UMTS, WCDMA, LTE-FDD 50 ppb N/A

50 ppb ±1.5 μs (for cell radius ≤ 3 km)

5G-NR 50 ppb ±1.5 μs

Figure 3: Examples of simultaneous UL/DL transmissions in TDD networks

2.1.2 Benefits and challenges of synchronised operation

2.2 UNSYNCHRONISED OPERATION

2.2.2 Benefits and challenges of unsynchronised operation

Spectral leakage from the interfering BS transmitter side:

Blocking of the victim BS receiver:

Figure 4: BS-BS interference mechanisms in case of simultaneous UL/DL transmissions

Figure 5: AAS and non-AAS base stations architecture

2.3 SEMI-SYNCHRONISED OPERATION

Fixed DL Flexible DL Flexible UL Fixed UL

Fixed DL Flexible DL Fixed UL

Figure 6: Example of semi-synchronised operation

2.3.2 Benefits and challenges of semi-synchronised operation

2.4 OPERATING MODES SUMMARY

Table 2: Options associated with the synchronised, unsynchronised

Options Synchronised Semi-synchronised Unsynchronised

Mandatory (with proper SLA)

Options Synchronised Semi-synchronised Unsynchronised

The ECC restricted baseline limit (as defined in ECC Decision

Both requirements are difficult to be cost effectively implemented

3 PERFORMANCE IMPACTS FROM FRAME STRUCTURE SELECTION

3.1 GENERAL INFORMATION ABOUT 5G-NR FRAME STRUCTURE

3.2 PERFORMANCE IMPACTS OF 5G-NR FRAME STRUCTURES

 DL/UL traffic ratio;

17 As defined in Section A3.1.7.

3.3 COMPATIBILITY OF 5G-NR FRAME WITH LTE-TDD FRAME STRUCTURE

Figure 8: Variant 1 - Synchronised operation of 5G-NR ("DDDSUUDDDD" frame)

Figure 9: Variant 2 - Synchronised operation of 5G-NR (“DDDDDDDSUU” frame)

3.3.1 Summary from Study #1 (grant-based UL transmissions)

The analysis carried out in this study assumes grant-based UL transmissions.

A summary of the latency analysis results is reported in Table 3.

Table 3: Simulation analysis – latency

# required HARQ processes 4 8

DL HARQ RTT 2-3 ms 5 ms

UL scheduling delay 1-2 ms 4.5-9.5 ms