Scf251 5g NR Fr1 Reference Design DEC-2021
Scf251 5g NR Fr1 Reference Design DEC-2021
Scf251 5g NR Fr1 Reference Design DEC-2021
Reference Design
The case for a common, modular
5G FAPI: RF architecture for 5G NR FR1
small cell distributed radio units
and Digital
DECEMBER 2021
Frontend
Control API
Month Year
Broad roll-out of small cells will make high-grade mobile connectivity accessible
and affordable for industries, enterprises and for rural and urban communities.
That, in turn, will drive new business opportunities for a widening ecosystem of
service providers.
Our focus and work program reflects two key areas of diversification in the
small cell ecosystem – the emergence of alternative deployment models such
as neutral hosting and private networks, and Open RAN specifications enabling
disaggregation of small cells at both component and network level.
All content in this document including links and references are for informational
purposes only and is provided ‘as is’ with no warranties whatsoever including
any warranty of merchantability, fitness for any particular purpose, or any
warranty otherwise arising out of any proposal, specification, or sample.
The Small Cell Forum Release Program website can be found here: www.scf.io
If you would like more information about Small Cell Forum or would
like to be included on our mailing list, please contact:
Email [email protected]
Acknowledgements
Author Company
Husbands, Ryan BT
Suggett, Dan Colt
Roy, Banaj JIO
Lingasamy, V STL
Chapman, Chuck Alpha Technologies
Small Cell Forum members believe that, to reduce complexity, a certain level of
standardization is required. This will help speed deployment, lower costs and enable
takeup of small cells to scale more rapidly.
Underpinning such a design should be the capacity to deliver flexibility and scalability
and the ability to swop out regionally specific RFFE modules.
This paper will be a useful reference for the entire wireless infrastructure ecosystem,
but is especially targeted at:
These interfaces, and reference designs based upon them, will enable the industry to
move away from products that are proprietary to one vendor or specific to just one
use case. Currently, the interfaces between different vendors’ small cells, or between
different functional splits, are not uniform. This increases cost, complexity and
integration time and reduces scale.
Common interfaces address the challenge that not every component can ever be
uniform. The biggest issue, as this paper describes, is that the RF front end will be
different, depending on spectrum frequency, output power and other factors. Neutral
host and enterprise deployments, in particular, will need to support as many
frequencies as possible to accommodate many users, which increases the bandwidth
requirement. Some scenarios outlined in this paper would benefit from dual-band
small cells with two different RFFE units.
The balance between uniformity and diversity is particularly delicate, then, when it
comes to the RFFE. The solution is a modular approach to small cell network design, in
which a variety of RFFE designs can be interfaced in a standardised way to the same,
common baseband and TRx units. The RFFE can be swapped in and out to support
different markets without changing the rest of the platform.
This paper provides the foundational framework for understanding the components
and interfaces that make up a 5G NR FR1 small cell distributed radio unit. Reference
designs can then be built within this framework to address different markets and
scenarios with different RFFE requirements. That will enable designers to create
products that can be easily configured and reconfigured for different applications. The
paper also provides an overview of available solutions, as vendors and deployers
increasingly look to have a second or third source of each component.
Of course, the market for 5G NR small cells is in its early stages, and it is important
that frameworks and reference designs are defined with future trends in mind that will
drive future component requirements. Several important trends will shape component
specs over the next few years as 5G small cell networks mature and address many
deployment environments.
These include:
All these changes in market requirements and technology capabilities will drive new
performance into small cells. If reference designs and interfaces fully reflect these
evolving requirements, they will support products that are not only more powerful and
efficient, but can be designed and deployed in a modular, flexible way all the way to
the RFFE.
Having laid the foundations for a common, modular architecture for 5G NR FR1 small
cells at component and reference design levels, Small Cell Forum will continue to
evolve the platform in line with new requirements and business models, and work with
members to apply it to real-world use cases and products.
In this way, the industry can achieve true scale – and adoption, by all types of
deployer, can be accelerated.
Figures
Figure 1-1 New small cell deployments & upgrades by spectrum SCF050 [1] .....1
Figure 1-2 5G scenarios & diverse solutions ...................................................2
Figure 1-3 Small cell forecast deployment for vRAN architectures SCF050[1] .....2
Figure 1-4 New deployments and upgrades of indoor small cells SCF050[1] ......3
Figure 1-5 Smal cell forecast adoption splits in open RAN by 2023 SCF050 [1] ..4
Figure 1-6 Small cells supporting open RAN architectures SCF050 [1] ...............5
Figure 2-1 FR1 TX/RX per deployment scenarios SCF238 [2] ...........................7
Figure 2-2 Small cell disaggregated RAN .......................................................8
Figure 2-3 Functional composition of a typical small cell radio unit ................. 10
Figure 3-1 Main functional blocks of a small cell radio unit ............................ 11
Figure 3-2 Example of a RF transceiver architecture ..................................... 15
Figure 3-3 Small cell RFFE components in TDD mode .................................... 17
Figure 3-4 Small cell RF line-up in TDD mode (simplified illustration) .............. 18
Figure 3-5 2T2R Indoor Small Cell TX power budget ..................................... 18
Figure 3-6 Small cell RFFE components in FDD mode .................................... 20
Figure 4-1 Example Small Cell Radio Unit Interfaces ..................................... 22
Figure 4-2 Key layers of the JESD204B/C standard ....................................... 24
Figure 5-1 Taiwan 5G NR Mid-band spectrum assigments [16] ....................... 32
Figure 5-2 Small cell DRU 2T2R example with one operator ........................... 33
Figure 5-3 Dual-band Small cell DRU 2T2R example ..................................... 33
Figure 5-4 UK 5G NR Mid-band spectrum assigments [16] ............................. 34
Figure 5-5 Small cell DRU 2T2R example with one operator ........................... 34
Figure 5-6 Dual-band Small cell DRU 2T2R example ..................................... 35
Figure 5-7 Dual-band Small cell DRU 2T2R less optimal example.................... 35
Figure 5-8 Unified Small cell DRU 2T2R ....................................................... 36
Figure 5-9 Unified Dual-band Small cell DRU 2T2R ....................................... 36
Figure 5-10 Small cell 2T2R basic configuration ............................................. 37
Tables
Small cells are low-power base stations used to improve network coverage and
capacity, delivering higher-quality wireless connectivity to the end user. As the 5G
roadmap shifts networks to higher frequencies to provide the increased data rates
demanded by both consumer and enterprise markets, radio signals will suffer from
higher propagation losses, reducing individual small cell coverage. To compensate for
this loss, network infrastructure requires more densification. How networks are
densified – and the architectures that will be used to make that happen – will play a
major role in the successful deployment of the new 5G wireless infrastructure network.
The latest SCF SCF050 market forecast report from July 2021 [1] shows that the vast
majority of the current and short term deployments for small cells are for 5G FR1 or
Sub-6GHz. An excerpt of this forecast can be seen in Figure 1-1.
The forecast, and expectations from market, indicate that, in the coming years, as 5G
is deployed, the adoption of small cell networks for 5G FR2 (mmWave) is expected to
ramp up and, by 2026, finally overtake the FR1 (‘sub6’) deployments. Operators are
expected to massively deploy FR2 small cell solutions for multiple verticals and end-
to-end applications, including the Internet of Things (IoT), machine-to-Mmachine
(M2M) communications and Industry 4.0 among many others.
However, for the next few years, we can expect the backbone of 5G to be the mid-
band spectrum of FR1. This is critical to the utility, versatility and deployment of 5G
because it bridges the speed, capacity, coverage and penetration gaps between the
low bands of FR1 and higher bands of FR2.
Figure 1-1 New small cell deployments & upgrades by spectrum SCF050 [1]
For this reason, this paper is focused on the current 5G deployment situation and
therefore will be looking at reference design for 5G NR FR1 small cell.
The SCF238 5G Small Cell Architecture and Product Definitions report from July 2020
[2] notes three important trends in the industry that drive the diversification of small
cells.
As 5G small cells are deployed, the trend will shift towards disaggregated and
virtualized RAN (vRAN) architectures in which radio units (RU) are separated from
virtualized basebands, ultimately going to the cloud. Figure 1-3 shows progress
towards this new implementation and how deployers will focus on different
architectures over time. This in turn clearly shows the diversity this will bring about
and the need for flexible solutions.
Figure 1-3 Small cell forecast deployment for vRAN architectures SCF050[1]
Figure 1-4 shows current and forecast small cell deployments categorized by
deployment type. It is expected that the private network deployment will be the major
deployment scenario, with more than 50% of the market post 2023. After that, the
fastest growing scenario is neutral hosts followed by traditional public MNO networks.
4,500
4,000
Figure 1-4 New deployments and upgrades of indoor small cells SCF050[1]
With these three trends working in parallel, it is not surprising that one size will no
longer fit all in the small cell landscape. Diversification will play an important role.
This trend towards diversification highlights the need for a clear understanding of the
requirements for each of the many options that will be needed. Just as no single
design works for all the diverse requirements in 4G/LTE and 5G, so there will be a
need for several split options to accommodate the wide variety of network
architectures and user environments that will be adopted SCF238 [2]. Many operators
plan to support more than one split, depending on scenario. For instance, as can be
seen in Figure 1-5, split 7.x (O7), which also spans macro cells, has its strongest
support in public urban and rural networks, where there will be cells of all sizes in use.
On the other hand, split 6.x (O6) will be mostly used for private networks, campuses
and indoors.
The key is not rigid uniformity. The industry needs to converge around a small number
of splits and interfaces, to provide flexibility for operators without sacrificing scale and
interoperability. There are strong signs that the industry is indeed narrowing
in on just two or three split choices in small cells. However, once the common
frameworks are in place, all stakeholders need to understand the capabilities that will
be required in each environment. That will enable them to launch and deploy small
cells which can achieve scale rapidly because they are clearly aligned to the practical
requirements of each sector and use case. That is why this study is so important.
In a traditional RAN system, the radio, hardware and software are proprietary – nearly
all the equipment comes from one supplier. Operators are unable to, for example,
deploy a network using radios from one vendor alongside hardware and software from
another vendor. In contrast, open RAN allows networks to be built using
subcomponents from a variety of vendors.
Figure 1-6 shows that in 2024 more than half of new small cell deployments or
upgrades will be based on open RAN interfaces. This figure will rise to 77% in 2026.
From a supply chain perspective, open RAN encompasses the idea of chopping up the
5G supply chain into smaller pieces and imposing standards on hardware and software
so all the different products can work together.
However, in order to achieve this aim, all different modules and parts from the system
need to be able to operate together. This requires a measure of harmonization, not
only in the higher software level but also in the component hardware level.
1.3.1 Harmonization
A good example of this comes from the recently created Open RF Association
(OpenRF). This industry consortium is dedicated to improving the 5G ecosystem. Its
focus is on the RF front end, its functionality and the interoperability between
hardware and software. The aim of this association is to standardize on some common
block elements and in non-competitive areas. This can speed up time to market,
ensure compatibilities between component generations and platforms, and improve
economies of scale.
For example, a list of reference designs allows users to quickly find the designs that
best fit their requirements and require minimal adaptations for their specific project.
This reduces product time to market.
Reference designs are built upon recommended and proven best practices. RF and
electronic component manufacturers often provide reference designs for their
products. This helps their customers to be more efficient and effective when applying
them. It also allows customers to adapt their products to customer solutions more
quickly.
However, this process makes a few demands of its own. For example, before an ODM
or system integrator can even begin to compare and contrast reference designs, they
need to develop and agree on a list of basic requirements. An appropriate reference
design can then be selected from a list of different possible options. Only then can
designs be compared and trade-offs analyzed.
As we have seen, there are multiple deployment scenarios for 5G. Therefore the
market will have multiple product requirements and need multiple solutions.
Because of the increasing diversity of the services small cell networks must support,
the environments they are deployed in, and the organizations that build and manage
them, the requirements for design and performance of small cells are also becoming
very diversified.
The July 2020 SCF SCF238 5G Small Cell Architecture and Product Definitions report
[2], looked at product requirements for different deployment scenarios. The table in
Figure 2-1 shows 4T4R is a common configuration. However, for low-cost indoor and
enterprise 2T2R is also required, while for outdoor 8T8R has become another popular
configuration.
80%
60%
40%
20%
0%
Residential Indoor Private Campus Dense Rural
enterprise industrial urban public
In the next chapter we will look in more detail at how these small cell requirements
and drivers translate to product architecture and design.
In this paper we will take as reference two of the most typical small cell configurations
(see Table 2-1) – one for indoor and one for outdoor.
Disaggregation of RAN and small cell networks is presently being pursued in the
industry because it can potentially bring several advantages. These include efficient
operation of remote radio units via centralized units and gains from pooling centralized
resources. There is also a potential for cost reduction due to low-complexity radio
units and shared centralized units.
Figure 2-2 shows the functional entities of a 5G small cell disaggregated RAN. These
comprise:
• Central unit (CU), which implements RRC and PDPC functions and
interfaces to the core network over the 3GPP F1 interface.
• Distributed unit (DU), which implements RLC, MAC and optionally PHY
functions (complete PHY for Split-Option-8, Upper-PHY for Split-Option-7.2).
• Radio unit (RU), which implements some PHY functions (Lower-PHY for
Split-Option-7.2, complete PHY for Split-Option-6), and integrates physical
radios and antennas. The RU interfaces to the DU using either nFAPI as
specified in SCF225[3] or eCPRI as used in the Open Fronthaul specified by
the O-RAN Alliance [4].
The scope of the small cell radio unit may be restricted to RU functionality in a fully
disaggrated RAN, or RU+DU when those functions are integrated into a single physical
unit.
The small cell radio interface split options as described in SCF238[2] define the
following functions residing in the small cell radio unit:
The relative merits of these split options are discussed in detail in SCF238 [2] and
work is currently in progress on a fronthaul study and calculator that will be published
as SCF247.
[Note: in the case of split options 6 and 7.2, the small cell radio network interface is
classed as fronthaul, while for split 2 it is classed midhaul by ITU-T G.8300.]
A small cell is defined in SCF238 [2] as a cellular base station that transmits and
receives 3GPP-defined RF signals with small power and small form factor. In most
cases, it services a small coverage area. A small cell radio unit is provided with power
and a fronthaul/midhaul data connection by a host network and radiates RF, either
using integral antennas or via external antennas. Synchronization may be derived
from the fronthaul/midhaul network, from the macro network, or, if located outdoors
via a GNSS receiver, through an integral or external antenna.
A typical small cell radio unit can be decomposed into the functions shown in Figure
2-3.
Figure 3-1 illustrates the main hardware subsystems and components inside a small
cell radio unit, and can be used as a reference when reading the more detailed
descriptions of the sub-systems in this section.
The baseband part of the small cell radio unit contains the digital electronics required
to run 3GPP protocol and physical layer stacks. It is also responsible for time
synchronization, security functions, and configuring and monitoring the other functions
in the small cell.
The baseband part may comprise a single chip, in the case of a low layer fronthaul
split (for example Split 7.2 and Split 6), or separate NPU and PHY chips for higher
level splits where more of the 3GPP protocol functionality is implemented in the small
cell (for example Split 2).
The baseband may serve one or more carriers and/or operators, with similar or
different radio access technologies TR37.900[7].
3.1.1 NPU
3.1.2 PHY
The PHY implements 3GPP physical layer processing functions and can be split into
high-PHY and low-PHY functions as follows:
• Scrambling
• Channel coding
• Modulation
• Layer mapping
• Equalization
• Forward error correction
• Precoding
• Digital beamforming
• Fast Fourier Transform
• PRACH filtering
• Digital front end
In addition the PHY provides real-time control to the RFFE, (for example TDD
switching).
3.1.3 Synchronization
Synchronization may be conveyed to the small cell using IEEE1588 PTP + SyncE over
the fronthaul/midhaul interface – or from GNSS in the case of an outdoor small cell. In
either case, the baseband implements a servo-loop in which small cell time and
frequency are accurately aligned to the primary synchronization source.
In cases where neither PTP nor GNSS are available, a third option is to monitor over-
the-air macro network transmissions and align to those. Where small cells are
deployed indoors, lower frequency macro transmissions with lower penetration losses
may be preferred.
The PHY and RFIC clocks are locked to the TXCO to ensure that the baseband symbol
timing and RF carrier frequency meet 3GPP requirements.
3.1.4 Security
Small cells have particular security vulnerabilities as they are often deployed at
physically insecure sites and may even be a target for abuse by users trying to
circumvent financial or technical restrictions. SCF has commissioned its own study of
small cell security issues and published a condensed report SCF171 [5].
The O-RAN Alliance has published a comprehensive threat model analysis of the 5G
RAN segment [6] and is working towards a complete set of implementation
requirements to mitigate the security risks.
Table 3-1 summarizes the hardware feature requirements that help create a secure
small cell platform.
3.2 RF transceiver
The radio frequency transceiver is a device used in the small cell and the macro base
station equipment of advancedcommunication systems – for example massive MIMO.
It provides the high radio performance demanded by cellular infrastructure
applications.
The transceivers support both TDD and FDD applications. To fulfil the 5G FR1
specifications, the transceiver operates in a wide frequency band that can cover a
range from 650 MHz up to 6 GHz (or higher). Depending on the number of controlled
transmitters and receivers in the single-chip transceiver device (2T2R, 4T4R, etc.), the
transceiver bandwidth performance can be specified. One or two observation receiver
channels can monitor feedback from the transmitter outputs.
The feedback loop can be used to implement error correction, calibration and signal-
enhancing algorithms such as crest factor reduction (CFR) and digital pre-distortion
(DPD).
The observation receivers operate in the same frequency range as the transmitter
channels and can support a channel bandwidth up to more than two times the
transmitter operation bandwidth.
To enable transmitter dynamic range of the radio unit, the transceiver can provide a
power control range using variable attenuator(s). To achieve a high level of RF
performance, the transceiver includes integrated phase-locked loops (PLLs).
The PLLs provide low-noise RF synthesis for the transmitter and receiver signal paths,
support an independent frequency for the observation receiver, and generate clocks
for the converter and the digital circuits for signal processing and communication
interfaces.
In this section we briefly review the RF functionality of a small cell radio, along with
design choices and technology constraints.
The low noise amplifier (LNA), as the first amplifier on the Rx-path, should ideally
have a noise figure lower than 1 dB and a sufficient IIP3 (input third-order intercept
point) to meet the sensitivity and blocking requirements in the small cell. The pHEMT
(pseudomorphic high-electron mobility transistor)-based LNA is a common choice due
to its low NFmin, and sufficient high operating voltage. In the filters, the passband loss,
the rejection out of the passband, and power handling are the three key parameters in
the component selections. The acoustic wave filters, including SAW (surface acoustic
wave) and BAW (bulk acoustic wave) types, are the popular choice for the Rx-path
due to their high rejection within a small filter size. The ceramic type filter is typically
chosen as the last stage of RFFE to connect with the antenna due to its relatively
higher power handling than the acoustic type.
Based on the current 3GPP definition, there are more than 20 frequency bands for
sub-6GHz 5G. This implies that RFFE hardware will be band-selective. The component
suppliers will typically try to develop multiple products in a pin-to-pin-compatible
family with each other to cover the entire sub-6GHz 5G bands in order to minimize
hardware effort on changing the RFFE’s frequency band to another band. In addition,
the number of transmitting and receivIng paths in the RFFE needs to be scalable in the
order of 2 to meet the desired MIMO (multi-input multi-output) configuration – that is,
2T2R, 4T4R, and 8T8R.
The baseband function is almost entirely digital in nature, configured by software and
independent of RF frequency. Baseband components include digital processors,
memories and interface components.
Small cell radio units designed for high-volume markets may integrate all functions
onto a single printed circuit board for lowest cost. Alternatively, some lower-volume
markets may be addressed more economically using a modular solution which, for
example, integrates interfaces, power, baseband, and RF transceiver functions onto
one printed circuit board applicable across a range of markets, while band-specific
components are grouped into a separate market-specific RFFE module.
Figure 3-3 shows a 2TRO (transmitter receiver with 2 observation path) RFFE in TDD
(time division duplexing) mode for sub-6GHz 5G application. ‘In TDD mode’ means
both transmitter and receiver operate alternately at the same frequency band. Ideally,
in a 2T2R configuration, the two transmitting paths are identical to each other; the
same holds for the receiving and observing paths. All the RF components in the RFFE
The bandpass filter between circulator and antenna port – the so-called antenna filter
– typically helps to eliminate the harmonic signals generated by the Tx power
amplifier (PA) and protect the receiving path from interference and jamming. The filter
after the low noise amplifier (LNA), as a second filtering component on the Rx-path,
further suppresses the interference signal from the antenna (ANT) port. The circulator,
as a non-reciprocal component, directionally transmits the Tx signal and shields both
PA and switch from the reflecting signal from the antenna port due to the impedance
mismatch.
The system integration can further insert a power sensor between the switch and
circulator to monitor the input power coming from the ANT port. When the sensor
indicates the input power is higher than the expected value, the Rx components can
be set either into shutdown or bypass mode to avoid damage. Both antenna filter and
circulator need to have a power handling of at least equal to, or higher than, the
output power of the PA.
For simplicity the following block diagram in Figure 3-4 only shows single TRO RFFE
and its interfaces with a transceiver board and an external antenna.
The passive components and interconnections all contribute to loss in the system. The
RF ports of the transceiver (TRx) are all in the differential type and require the passive
baluns to connect to the RFFE board. The passive components on the RFFE board
include the coupler, circulator and bandpass filters. Additional losses occur from the
interconnection and path losses in the system.
Figure 3-5 illustrates the power amplifier output power will need to be +28dBm to
compensate for filter and antenna losses to allow +24dBm to be delivered to the
antenna port.
Typically, the TRx can deliver around 4 to 6 dBm power in CW mode. With the PAPR
(peak to average power ratio) of the NR signal, targeted DPD expansion and the
balun’s loss, the output average power from the transceiver board is about -7 dBm.
To deliver 24 dBm to the antenna, the power gain of the power amplifier (PA) needs to
be equal to or higher than 35 dB to compensate for all the rest of the loss on the
transmitting path. Due to the DPD effort, the PA’s saturated output power (Psat)
should be equal to or higher than 36.25 dBm, and its OP1 (output 1 dB compression)
RFFE Tx line-up
RFFE Rx line-up
Table 3-2 RFFE Tx and Rx Line-up requirements for indoor small cell case.
In the Rx line-up, the FEM, which integrates an LNA and SPDT switch into a single
component, is the only active component on the Rx path of the RFFE PCB. The typical
gain and loss of the extra components, including BPF2’s loss, and both FEM and TRx’s
gain and noise figures, are summarized in the table.
By using the textbook equation of the noise figure (NF) in a cascading system, the NF
of the entire Rx path based on the presented components is about 5.47 dB. The
effective receiving gain, which is a summarization of all the gain and loss on the Rx
path, is 35.9 dB. The total NF can be further improved by either using the passive
components with less loss or increasing the LNA gain.
The temperature sensors in the RFFE block diagram mainly help the system to monitor
the operating temperature of the RFFE in real operation. The multiple temperature
The purpose of the memory inside the RFFE is to store the following information to
help the small cell in system initialization, online or offline calibrations. Section 3 looks
in more detail at the memory’s interface, parameter definitions, data type and
formats.
Figure 3-6 shows the FDD (frequency division duplexing) RFFE block diagram. The
dashed line in the figure indicates the boundary between the RFFE and two duplexers
in the 2T2R configuration. The diagram on the left side of the line is identical to the
same portion of TDD mode RFFE in Section 2.5 except for the connections of the
circulators and switch. This design approach consequently unifies the RFFE interface
with antennas in the two system operations. This methodology can reduce the extra
effort when the designer wants to change the current design to another or wants to
implement a multi-mode small cell to provide TDD and FDD operations in the same
hardware.
The operating frequencies of the transmitting and receiving paths in the FDD small cell
are different. The observation path needs to be designed at the same frequency as the
3.4 Antenna
The antenna, which is the interface between the small cell and the free space, radiates
the RF power and captures the signal from the user equipment (UE). In both TDD
(time division duplexing) and FDD (frequency division duplexing) small cells, the
transmitting and receiving paths from the same RF transceiver share the same
antenna. The number of RF transceiver ports and the feature of the polarization
diversity of the small cell define the total number of antenna elements connecting with
the cell. Based on the different system requirements, the antenna can be designed to
be either omnidirectional – for the full 360 degree angle – or directive – to focus the
radiation on certain angles. The antenna can be in the planar form, which is realized
by a multilayer printed circuit board (PCB) and further integrated into the RFFE board.
The planar antenna, for example, patch, slot antenna, and the planar inverted-F
antenna (PIFA), typically has a gain of about 3 to 7 dBi. The 3D type of antenna is
also popular in small cells. These antennas, including dipole, monopole, and helical,
typically use cable or connector to interface RFFE.
Antennas may be integrated within, or separate from, small cell radio. If antennas are
integrated, they may be incorporated onto the PCB without connectors, or mounted
off-board, in which case flying leads may be soldered or connected using, for example,
UFL-style connectors. When antennas are not integrated, as may be the case for
higher-power outdoor small cells, waterproof RF connectors such as N-type or 4.3-10
may be employed.
External interfaces are required for fronthaul/midhaul connection and power. For
outdoor small cells, interfaces are typically presented using weatherproof connectors
at the base of the unit to prevent water ingress. Indoor small cells are less restrictive
and tend to utilise standard IT connectors.
Figure 4-1 provides a diagram of the small cell radio unit functional blocks and
highlights all the inferfaces – comprising the data path interfaces u1 to u5, control
interfaces c1 to c3, and test signal interface t1.
This section explains the function of each interface, analyzes the interface
requirements and variations across various small cell product requirements and
provides recommendations for harmonization.
The fronthaul/midhaul interface provides the main data connection back to S-DU or
CU. Interface speed is dependent on RF capability and functional split, and can range
from 1Gb/s (Split 2, low-band spectrum) to 25Gb/s (Split 7.2, mid-band spectrum).
Rates of 1Gb/s, 2.5Gb/s, 5Gb/s and 10Gb/s may be supported over copper
(Cat5e/Cat6a) with 10Gb/s and 25Gb/s supported over fibre (SR multi-mode or LR
single mode).
The fronthaul/midhaul interface, labelled u0 in Figure 4-1, provides the main interface
back into the operator network. Ethernet standards are preferred, with copper
recommended where PoE is feasible and fibre where >100m cable runs and/or 10G
and higher rates need to be supported.
The interface between NPU and PHY, labelled u2 in Figure 4-1, carries FAPI messages
SCF222[9], SCF223[8] and [SCF224[10] or O-RAN C/U-plane messages. Where
separate chips are used for NPU and PHY functions PCIe is a suitable interface, with
four lanes of Gen3 providing approximately 25Gb/s and Gen4 providing 50Gb/s net
throughput and reduced data path latency.
The baseband port interface, labelled u3 in Figure 4-1, carries the digital baseband IQ
samples between the PHY and the RF transceiver. Each complex IQ sample is typically
32 bits in total across real and imaginary parts. The IQ sample rate is as-per 3GPP
sub-carrier spacing as specified in Table 4.2-1 TS38.211 [11] multiplied by FFT size
B.5.2 in TS38.104 [12]. As a typical example, a single FR1 100MHz carrier directed to
one antenna has a sample rate of 30kHz * 4096 = 122.88MS/s, which equates to a
throughput requirement on this interface of 3.93216Gb/s. In the case DPD is
implemented in the PHY. This value should be multiplied by 2-4x to allow for DPD
bandwidth expansion.
JESD204 is a suitable interface between the PHY and the RF transceiver. It is designed
for transport of high-bandwidth digitized samples and can support multiple links to
separate endpoints (RFIC) with deterministic latency; this is very important in meeting
3GPP timing specifications.
The JESD204 specification defines four key layers that implement the protocol data
stream, as shown in Figure 4-2. The transport layer maps the conversion between
samples and framed, unscrambled octets. The optional scrambling layer
scrambles/descrambles the octets, spreading the spectral peaks to reduce EMI. The
datalink layer handles link synchronization, setup and maintenance, and
encodes/decodes the optionally scrambled octets to/from 10-bit characters, in the
case of JESD204B, and 66 bit characters, in the case of JESD204C. The physical layer
is responsible for the transmission and reception of characters at the bit rate.
Table 4-2 lists example SerDes lane rates and net throughputs for the JESD204
interface. The final column provides a count of the number of 100MHz antenna-
carriers (AxC) that can be supported over a single JESD204 lane.
Taking, for example, four JESD204 lanes operating at 9.8304G, capacity for up to
eight AxC could be utilised as follows:
In the above examples, the number of RFICs required may be one or two, depending
on the capabilities of the RFIC.
The baseband serves as the point of control for the RF transceiver and RFFE. Control
may be in the form of configuration (for example centre frequency, as carried over
FAPI P19 as specified in SCF223 [8]), or fast control (for example, TDD switching as
inferred from FAPI P7 slot procedures as specified in SCF222 [9]) or directly from the
SCF223 FAPI P19 TXRX_PATTERN_TLV [8]. FAPI P19 messages are interpreted and
actioned by c1 and c2 in Figure 4-1.
Serial control provides the means to configure, activate and monitor devices.
Candidate interfaces are serial peripheral interconnect (SPI), I2C and MIPI RFFE.
Common features across these include:
I2C implements a 2-wire bus-interface with address specified as part of the serial
protocol. It also supports bi-directional data transfer on a single line. As with SPI,
transaction length and data formatting are flexible. I2C is widely used to control
simple, low-pin-count devices where bus speed is not critical.
MIPI RFFE is a 2-wire interface similar to I2C but operating at much higher speed.
Unlike SPI and I2C, the MIPI specification standardises transactions and data formats.
It also includes provision for time-accurate synchronised triggering across multiple
controlled devices.
Both SPI and I2C have been widely supported across a range of industries for decades.
MIPI RFFE was initially developed for handset applications but may become more
widely applicable to infrastructure designs as well. This is explained in more detail in
section 6.2 when speaking about trends and future work.
The control interface between baseband and RF transceiver, labelled c1 in Figure 4-1,
may support functionality including:
• RF transceiver configuration
• Carrier frequency
• RF bandwidth
• Transmit gain
• Receive gain
• AGC
• Receiver observation path selection
• RF transceiver monitoring
Most functionality can be supported by reading and writing to registers with the RF
transceiver over a serial control interface. Where sub-microsecond-level speeds are
needed dedicated timed GPIO signals may be appropriate.
The control interface between baseband and RFFE, labelled c2 in Figure 4-1, may
support functionality including:
• RFFE configuration
• RFFE fast control
• RFFE monitoring
A mix of serial control for configuration/monitoring and dedicated timed-GPIO for fast
control is usually appropriate.
1
SPI 3-wire mode combines DO, DI signals into one bi-directional DIO signal
Table 4-4 below illustrates common control and monitoring signals between BB and
RFFE (interface c2) for small cells in outlined deployment scenarios.
The control interface between the RF transceiver and the RFFE, labelled c3 in Figure
4-1, is optional and may include fast control from the TRx rather than BB – for
example LNA bypass (when an AGC overload condition is detected in the RFIC
receiver), or PA shutdown (when a signal overload condition is detected (circuit
breaker) and a dedicated GPIO is appropriate for this function).
There’s also the MDIO, an implicit control interface that operates between NPU and
Ethernet PHY. The small cell radio unit places no special requirements on this
interface.
The RFFE itself has a memory, allowing it to store certain pre-defined parameters. A
proposal by the SCF working group of typical RFFE calibration memory configuration
and information is listed in Table 4-5.
Category Item
Basic Product number
Basic Hardware series number
Basic Hardware manufacturing information
Basic Timing or version on cal. data generation
Basic RFFE configuration
Tx Operating frequency range
Tx Max. voltage and current consumption
Tx Max. input CW power
Tx Power gain verses frequency, temperature, and biasing
Tx OP1 verses frequency, temperature, and biasing
Tx Psat verses frequency, temperature, and biasing
ORx Operating frequency range
ORx Max. output CW power
ORx Coupling verses frequency and temperature
Rx Operating frequency range
Rx Max. voltage and current consumption
Rx Max. input CW power
Rx Gain verses frequency, temperature, and biasing
Rx Noise Figure verses frequency, temperature, and biasing
Rx OP1 verses frequency, temperature, and biasing
Rx OIP3 verses frequency, temperature, and biasing
ANT Operating frequency range
ANT 3dB beamwdith
ANT Gain verses frequency
ANT Efficiency verses frequency
ANT Polarizaiton
Interface/Logic Tx input impedance
Interface/Logic Tx output impedance
Interface/Logic Rx input impedance
Interface/Logic Rx output impedance
Interface/Logic ORx output impedance
Interface/Logic Logic of SW control
Interface/Logic Logic of Tx EN
Interface/Logic Logic of Rx EN
Interface/Logic Logic of Rx bypass
Interface/Logic Logic/interface of power reflection
Interface/Logic Logic/interface of temperature sensor
Power
Small cell radio unit power consumption may span a range from as low as 10 watts for
a low power indoor small cell to several hundred watts for a high-power outdoor small
cell. Highly efficient low-power chipsets are a good choice, both for low power
consumption/OPEX and for ease of installation, especially indoors where small cell
radios may be wall or ceiling-mounted with no forced-air cooling and minimal visible
heat sinking. DC power is preferable to AC for safety and for support for battery
backup. Power over Ethernet is especially attractive in reducing cabling overheads
and goes hand-in-hand with 30-40W power consumptions achievable in 4T4R units.
Table 4-6 summarizess the options for powering a small cell radio unit.
The hardware of the TRx contains a power management unit, which provides the
necessary digital and analog DC supplies for the TRx. It can also provide the main DC
supply for the RFFE through wire housing DC connectors, such as those shown in
Figure 4-3. The RFFE supply voltage depends on the components used. These may
range from -60V to +60V. The low voltage connector is suitable for low-power
provision up to 0.5W RF output designs.
The outdoor small cell is typically installed on existing infrastructures, including lamp
posts, bus stops and billboards. These public infrastructures generally don’t have DC
power outlets or AC supply, and may not have a power network in some cases.
Therefore, the PoE interface is mandatory in the outdoor small cell. The input AC
voltage generally is 100-240V in 50-60 Hz, and the current is around 2.5 A. Due to
the outdoor application, the outdoor small cell is usually shielded by the metal case.
The enclosure provides an additional ground point to connect the external ground
cable for electrical safety. All the interfaces, including PoE, and AC ports can be
equipped with water-resistant connectors to connect the cables.
RF interface
The RF interface between the transceiver and the radio front end module (RFFE)
includes the transmitters, the receivers, and the observation receivers. To reduce the
number and footprint of RF connectors between the TRx and the RFFE, the RF
interface should be an AC coupled, single-ended, 50 Ω impedance using a compact
plug/jack connector. Figure 4 shows examples of RF board connectors from Samtec.
IJ5/IP5 GRF1-J/GRF1-P
As for the transmitter path, the transmit power at the TRx/RFFE interface may vary.
For example, analog device transceivers have nominal values of -16 dBm and a
maximum peak of 6 dBm. This can vary ±3 dB depending on production and
temperature. These values include the 3 dB DPD expansion, 3 dB overhead, and 13 dB
peak to average ratio. The TRx receiver expects an input p1dB point from the RFFE of
about -11 dBm in maximum gain. The TRx can provide a complete automatic gain
Table 4-7 illustrates common RF connections between the TRx and the RFFE (interface
u4) for small cells in outlined deployment scenarios. The last column shows that there
are too many implementation variants to standardise the interface across all small
cells.
For outdoor installation a GNSS receiver may be included. GNSS can be used for
timing/synchronization and for asset tracking. The GNSS antenna may be integral to
the unit or external. If external it is usually connected via a TNC connector.
A small cell radio unit may include a local access port to allow installation and
commissioning checks to be performed prior to and after equipment has entered
service. This interface may be a physical Ethernet port, but it could also be wireless
(such as Bluetooth). Whatever the interface type it is important to make sure this port
can be secured against unauthorised access.
It is also possible to run additional test signals, labelled t1 in Figure 4-1, inside the
unit at board level to aid testing of 3GPP radio performance as described in section
5.2.3 of SCF228 S-RU and S-DU Test support [14] including:
This process will take place during conformance and production testing; it should be
disabled once the equipment is delivered to the field.
In the previous chapters we have looked into different product, component and
interface specifications for small cells. Based on that, this chapter shows a couple of
very high-level examples of a potential real implementation illustrating the
architecture of the small cell distributed radio unit and a suggested list of its
components inside.
Band N78 has six different channels allocated in total. From those channels, four
operators have already purchased a combined bandwidth of 270MHz; 30MHz remains
to be allocated. Another representative 5G band selected is N38 which has two
operators with a total bandwidth of 50MHz.
From a mobile operator’s perspective, having a small cell that covers only its band of
operation in the spectrum would be sufficient. However, from an equipment vendor
and neutral host perspective, having one small cell that covers the whole band N78
from 3.3 to 3.6GHz is the best implementation choice. This guarantees an equipment
solution that allows all mobile phone users to be covered by all the country’s mobile
operators – at least all the leading ones.
Bear in mind that, in this case, the total bandwidth of operation is increasing; this
means that the system integrator will have to make sure that the components used
for BB, TRx and RFFE can support the total bandwidth required.
Figure 5-2 shows the optimum implementation for this case, with a small cell
distributed radio unit covering the whole N78 with the same RFFE.
However, as can be seen in Figure 5-1, this small cell implementation omits one
operator: GT. Including GT will benefit both equipment vendors and neutral hosts.
Therefore, we have proposed implementing a dual-band small cell (Figure 5-3) with
two different RFFE units, one operating at N78 and other at N38.
It is important to mention that in this topology two different RFFE units are needed to
serve both different bands. On the transceiver and baseband side, both components
should be able to simultaneously operate two differents bands of operation. Otherwise
two basebands and two transceivers would need to be used inside the small cell unit,
increasing component count and cost.
Compared to the previous spectrum chart for mid-band, in the UK (Figure 5-4) there
are less operators but spectrum ownership is more complex.
In this case if we want to operate the whole N78 together, the components inside the
small cell will have to operate from 3.4 to 3.8GHz. As reference, remember that in
Taiwan it was from 3.3 to 3.6GHz only.
As with Taiwan, the best option for equipment vendors and neutral hosts is to
implement a small cell that covers the whole N78 band from 3.4 to 3.8GHz
Figure 5-5 Small cell DRU 2T2R example with one operator
Figure 5-5 shows the optimum small cell distributed radio unit implementation, which,
in the UK, involves operating in the whole N78 band.
As in the previous example, in this topology two different RFFE units are likely to be
needed to serve both different frequency bands and, depending on the performance
limitations of the transceiver and baseband units, one or two of the units would be
required. This is shown in Figure 5-7 with two alternative implementation solutions;
these are less integrated solutions compared to the previous one. However, as
mentioned in 5.1.2, a final implementation and frequency plan needs to be properly
planned to assess which topology would be the best.
Figure 5-7 Dual-band Small cell DRU 2T2R less optimal example
These examples are highly relevant to system integrators, ODMs and OEMs. They
show that a system integrator could develop one small cell distributed radio unit
This means that system integrators, ODMs and OEMs could develop one small cell unit
that would serve multiple customers minimizing the need for small cell variants and
improving both economies of scale and profitability. Exactly the same concept can be
applied for dual-band implementation – as seen in Figure 5-9.
Note, however, that, as before, increasing the bandwidth of operation from the
component side has a physical limit and an associated cost. Such factors need to be
taken into consideration while designing these small cell distributed radio unit
solutions.
As we have seen, choosing the right components is key to minimizing the work
required from the multiple design houses and maximizing the opportunity for selling
the same small cell equipment to multiple end customers. Here we look at a 2T2R
small cell implementation.
Based on Figure 5-10, we can now select different commercial components available
that we could use to implement such a small cell distributed radio unit. We will do this
exercise for a single an dual band cases.
Table 5-1 provides an example (by no means exhaustive) of a part list for
implementing the architecture and interfaces defined in this paper. For each building
block, at least two different potential component vendors has been offered, showing
that is already feasible to have a rich component vendor diversification in the supply
chain for the same small cell distributed radio unit.
2
https://www.whizzsystems.com/wp-content/uploads/2021/02/ORU-Whizz-Whitepaper_Rev6_Nov05-
2020_QQS.pdf
Notice that for the selection of components we have used an 8W amplifier for the band
from 3.3 to 3.8GHz and 4W for the band from 2.5 to 2.7GHz to show that in the same
small cell you could have two different output power levels per band. This shows the
level of diversity that small cell distributed radio units could manage.
Both component lists shows that finding the right product components from different
vendors to fit the different requirements of the small cell is an engineering challenge.
However, component vendors are already working on bringing to market components
that are easy to replace, offering the same footprints and being pin-to-pin compatible
with competiting or the same product families. This would make life easier for
engineers and system integrators, as multiple small cell distributed radio unit solutions
with minimum component changes would allow a more flexible and modular
approaches.
Important note: all the commercial components, brands and part numbers
selected in both tables are used for reference only. No promotion or
recommendation for commercial use is intended. The list is an indication of
component market diversity and therefore, it is by no means exhaustive.
These are the major industry trends that we believe will influence component
specifications in the future.
More efficiency: Energy-efficient networks are becoming more and more important.
All wireless infrastructure, including small cells, will need to ensure very low energy
consumption. At the component level this means quick turn on/off times with minimal
current drawing will be important.
New spectrum and higher bandwidths: The FR1 frequency range was the first
band to be explored below 6 GHz. It is already extending to 7.125 GHz. The next
logical step for FR1 will be to add multiple frequency bands within 7-24 GHz for 5G
infrastructure. This would allow the introduction of more bands and, probably, higher
bandwidth.
There is also a growing demand for higher bandwidth support in the FR1 line-up for
hardware, allowing system integrators to develop a hardware solution where the
occupied bandwidth is equal to the frequency span of the band. This is important as it
will minimize the quantity of small cell variants that have to be developed.
It is clear that small cell solutions for FR1 have a high level of integration at the
baseband SoC and transceiver SoC. However, the RFFE is still realized in an entirely
discrete manner.
So which functionalities could be integrated? Look again at the RFFE block diagrams
from Figure 3-3 and Figure 3-6 for TDD and FDD operation.
If you omit the BPF filters, one could further integrate the RFFE for TDD and FDD. The
architecture of the integrated RFFE approach could even be shared between FDD and
TDD, with the market advantages that implies.
In a TDD implementation, a highly integrated FEM solution could look like the one
shown in Figure 6-1.
RFFE for low-cost small cell solutions could benefit in a number of ways from a higher
level of integration. Clearly, an integrated solution yields a significant cost reduction
compared to the discrete component solution. This integration will also result in
smaller PCB footprint while retaining functionality and performance. This is especially
relevant to the development of new low-cost small cells. Their smaller dimensions and
lower cost will mean they can be placed in many different locations to improve
coverage, connectivity and data throughput.
The integrated solution is also relevant to control, including such items as temperature
sensors and integrated bias DACs and ADCs for the PA, and could support more
features more easily than a discrete solution in terms of AGC and DPD.s
Finally, even the interface with the transceiver and baseband would benefit from
integration. Unlike a discrete RFFE implementation, which relies on a significant
numbers of dedicated GPIO pins for the different controls to each discrete block, the
setup and control of the highly integrated RFFE could be done with just a single digital
interface with a minimum of GPIO pins in support for critical controls.
Naturally, integration always comes with some challenges like heat dissipation. As
mentioned in section 1.3.1, there is already an association, called openRF, looking at
this point in detail; this underlines the ongoing importance of this technology trend in
the industry.
To sum up then, it could reasonably be argued that a higher level of integration in the
RFFE for FR1 mobile infrastructure would be beneficial for low-power indoor small cell
solutions.
As we have said, there are several candidate interfaces for the future control of the
RFFE solutions, especially with a higher level of integration: they include serial
peripheral interconnect (SPI), I2C and MIPI RFFE.
The MIPI Alliance is constantly upgrading the standard and supporting new features to
address some of the challenges in today’s mobile and infrastructure applications. 5G is
bringing an explosion in the number of RF bands in both uplink and downlink
combined with reduced sub-carrier spacing (SCS). In addition, 5G targets latencies in
the order of 1ms, which is one order faster than 4G. This poses a stringent
requirement on the switching timing and control between the bands. MIPI RFFE 3.0
helps to deliver tighter timing precision and reduced latency. In fact MIPI RFFE v3.0
delivers a 20x improvement in the precision of timing over the previous version.
For some of the reasons mentioned earlier, the MIPI RFFE working group points out
that the standard serial peripheral interface (SPI) is not suitable for its use: SPI slave
devices cannot share the bus lines. Without the proper slave addressing on a shared
data bus, SPI requires one slave select line and one slave-to-master data line per
slave.
MIPI RFFE clearly has some interesting features when it comes to RFFE control. The
common command codes for broadcasting operation are very useful if all slaves have
common configuration settings. The multi-master support could show flexibility in set-
ups where baseband and RF transceiver become a master in the communication and
the RFFE is the slave. Moreover, the timed triggers and their grouping, provide a lot of
flexibility. This could even allow the RFFE to omit the GPIO control pins for switching
between TX-mode and RX-mode. Finally, as the term I3C suggests, it supports I2C on
the same bus.
Even for FR2, the MIPI interface seems very attractive. But the slave count for 5G FR2
is limited to 15. This places tighter constraints on the antenna count versus IC count
in the FR2 line-up. Currently, the first generation of beam-forming (BFIC) front-end
modules in RF2 seem to support mainly SPI-like interfaces. However, most solutions in
the market have a proprietary SPI interface to support the FR2 BFIC. At the time of
A future SCF white paper could look in more detail at the interface and line-up
challenges and trade-offs for FR2 small cells distributed radio units. However, it seems
clear that this and following generations of MIPI RFFE could dominate digital interfaces
in the mobile infrastructure market in the near future.
As we enter the 5G era, many processes and approaches used for 3G and 4G are
changing to meet a new and more demanding mobile market in which small cells will
play a major part as networks densify. But how do we keep small cell prices down?
There is currently a growing volume of components and interfaces available and used
in the construction of 5G NR FR1 small cells. The problem is that cheaper, quicker
deployment implies an agile infrastructure, based on common platform components
that can adapt to evolving business models and usage pattern.
However deployment scenarios may, for now, be too diverse, and components too
specialized or proprietary to permit a single small cell harmonized reference design
that could cut development and rollout costs.
But this need not be the end of the story. While we understand that different users
need different customized solutions, SCF believes that its work to date makes a strong
case for a modular approach to reduce costs and complexity.
This, combined with a concerted effort to narrow the number of use cases, could be a
way forward that changes the economics of 5G small cell rollout for the better.
Having laid the foundations for a common, modular architecture for 5G NR FR1 small
cell distributed radio units at component and reference design levels, Small Cell Forum
will continue to evolve the platform in line with new requirements and business
models, and work with members to apply it to real-world use cases and products.
Term Description
25GAUI 25 Gigabit attachment unit interface
3GPP The 3rd Generation Partnership Project. A standards organization
developing 5G standards (TS38 series).
AGC Automatic gain control
BB Baseband
BPF Band pass filter
CU The centralized unit in a disaggregated radio access network (RAN)
DFE Digital front end
DPD Digital pre-distortion
DU The distributed unit in a disaggregated radio access network (RAN)
eCPRI enhanced Common Public Radio Interface
EIRP Effective isotropic radiated power
FAPI Functional application platform interface
FDD Frequency division duplexing
FFT Fast Fourier transform
FHI Fronthaul interface
FR1 Frequency bands within ranges below 7.225 GHz
FR2 Frequency bands within ranges above 24.250 GHz
gNodeB 5G network equipment that transmits and receives wireless
communications between a UE and a mobile network
GNSS Global navigation satellite system
GPIO General Purpose input output
HARQ Hybrid automatic repeat request
I 2C Inter-integrated circuit (serial bus)
IQ In-phase / Quadrature
LNA Low noise amplifier
MAC Media access control
MDIO Management data input output
MHI Midhaul Interface
MIMO Multiple input multiple output
MIPI Mobile industry processor interface
MNO Mobile network operator
NPU Network processing unit
NR New radio. This is a new radio access technology (RAT) developed by 3GPP
for the 5G (fifth generation) mobile network
ODM Original Design Manufacturer
Open RAN Open radio access network
O-RAN Alliance An organization focused on operator-defined next generation RAN
architecture and interfaces
PA Power amplifier
PAPR Peak to average power ratio
PCB Printed circuit board
PCIe Peripheral component interconnect express2
PDPC Packet Data Convergence Protocol
PHY Physical layer
PMIC Power management integrated circuit
PoE Power over Ethernet
[2] SCF238. 5G small cell architecture and product definitions. July 2020. Document
238.10.01
[4] O-RAN Fronthaul Control, User and Synchronization Plane Specification (O-
RAN.WG4.CUS.0)
[5] Comprehensive overview of small cell security, SCF171 release 9.0, 2017,
[6] O-RAN Security Threat Modeling and Remediation Analysis 1.0 - March 2021
https://www.o-ran.org/specifications
[7] TR 37.900. Radio Frequency (RF) requirements for Multicarrier and Multiple Radio
Access Technology (Multi-RAT) Base Station (BS)
[12] TS38.104 NR; User Equipment (UE) radio transmission and reception
[13] O-RAN White Box Hardware Working Group. Deployment Scenarios and Base
station Classes. Document O-RAN.WG7.DSC.0-v02.00.
[14] SCF228. S-RU and S-DU Test Support. May 2021. Document 228.1.0