by Skyworks Solutions, Inc.
This is part one of a two part article. The second part will run in the June issue of Microwave Product Digest.
As demand grows for ubiquitous wireless connectivity and the promise of new and previously unimagined applications— such as autonomous vehicles, artificial intelligence, telemedicine and virtual reality—so does the anticipation for 5G. 5G will be revolutionary, delivering higher data throughput, extremely low latency and speeds up to 100 times faster than 4G. As a result, 5G is moving toward commercial reality faster than many expected. With that in mind, mobile operators are implementing near-term, tactical efforts to ensure that 5G demonstration hardware would become available by late 2018 and throughout 2019.
This article explores the practical first steps of any rollout, focusing on the spectrum below 6 GHz as standards for mmWave applications have yet to be defined. Our approach is not intended to subscribe to any particular solution; rather, it is an introduction to what Skyworks believes will likely transpire over the next several years. In addition, our framework focuses primarily on the practical solutions for the 5G RF front-end (RFFE) in the sub-6 GHz arena. To help readers better understand what that practicality means, Skyworks presents its perspective on how early 5G will be implemented, with particular emphasis on enhanced mobile broadband applications, or eMBB, in 3GPP parlance. Our goal is to provide some reasonable expectations of the future and correlate it to current 4G LTE Advanced Pro to see how manufacturers will address the new requirements. We will describe how the early rollout for 5G will proceed, how the standards will be translated into networks and devices, and what we can expect to see over the next several years as 5G becomes commercialized.
With decades of experience over previous standards coupled with our systems and technology expertise, Skyworks is well-positioned to deliver the significantly more powerful and complex architecture demands associated with 5G.
3GPP Release 15 Summary: The Framework for Early 5G
Release 15 of 3GPP marks the commercial beginning of 5G. Its impact will be felt for the next several decades across multiple markets—from telecommunications to industrial, health, automotive, the connected home and smart cities as well as other emerging, yet unforeseen ones.
We expect this framework to underpin commercial 5G networks being deployed in 2020, even though additional network configurations have extended completion of the standard by approximately six months. The changes were incorporated to ensure delivery of all new radio (NR) architecture options and the finalization of option 3a (non-standalone) and option 2 (standalone). The updates will also include further development of the standalone 5G NR specifications as well as refinements to some of the earlier work. Release 16 will likely be used for application of NR to unlicensed bands, minor changes and improvements, with more substantial changes expected in Release 17.
With Release 15 in hand, mobile operators, device manufacturers, and chipset providers have the confidence and ability to move forward with concrete developments to support commercial deployments. We fully expect to see commercial products and announcements in 2019, ahead of larger scale network deployments in 2020.
In the following section, we explore some of the key takeaways from Release 15, with particular emphasis on the impact to the RF front-end.
Key RFFE Takeaways from Release 15
5G standards draw heavily upon the experiences and lessons learned from 4G LTE, including many of the concepts successfully proven to support increased data rates. This evolution and reliance upon existing technology allows several techniques from 4G to be integrated into the initial rollout of 5G, providing immediate benefits without the need to wait for future releases. The rollout is also supported by the use of E-UTRA (Evolved Universal Terrestrial Radio Access) NR Dual Connectivity (EN-DC) combinations where NR is always associated with an LTE link.
Multiple Input Multiple Output (MIMO) and Antenna Implications
A key takeaway from the early draft of Release 15 is that 4×4 downlink MIMO, particularly at frequencies above 2.5 GHz (which include n77/78/79 and B41/7/38), will be mandatory. Draftees of the specification recognize the benefit of 4×4 downlink, as well as its impact on the data rate and network capacity, and have thus made this a requirement for the first implementation phase of 5G.
The presence of four MIMO layers not only enables expanded downlink data rates, it also means there will be four separate antennas in user equipment (UE), opening up additional degrees of freedom for the RF front-end design community.
An additional feature that is strongly desired by mobile operators, though not mandatory, is the deployment of 2×2 uplink MIMO. Having 2×2 uplink MIMO in UE requires two 5G NR transmit power amplifiers (PAs) to transmit from separate antennas. This is particularly beneficial in cases where higher frequency time division duplex (TDD) spectrum is used—as is the case with n41, n77, n78, and n79 as well as other TDD bands. The effective doubling of the uplink data rate enables shorter uplink bursts and flexible use of the 5G frame timing to increase the number of downlink sub-frames, potentially increasing downlink data rates by up to 33 percent. However, when the downlink data rate becomes extremely high, the uplink is challenged by the requirement of rapid and constant CQI and ACK/NACK response from UE and will be required to support 5 to 6 percent of the downlink data rate. As a result, the uplink data rate can eventually limit the downlink data rate and, without uplink MIMO, the coverage area and maximum downlink data rates will be limited by the uplink data rate performance.
A further use of the available second transmit path is a new transmission mode called “2Tx coherent transmission.” This effectively uses the principles of diversity, which are heavily leveraged on the downlink side of the network and enable up to 1.5-2 dB of additional transmit diversity gain, which is critical to address the fundamental uplink limited network performance. Studies have shown that such improvements in uplink channels equate to an approximate 20 percent increase in range at the cell edge. Why is this so important? Operators report that most mobile calls originate from within building structures (approximately 75 percent of calls are made from inside a home or an office), which causes signal degradation and decrease in cell radius. In other words, the call is operating from the cell’s edge, which is physically located further away from the base station. Thus any adjustment made toward that end will be viewed positively by the operators and help to minimize costs of 5G networks.
Beyond improving cell edge performance, 2×2 uplink MIMO improves spectrum efficiency. Since 5G NR is mostly a TDD technology above 2 GHz, and TDD cells are likely to be configured in a highly asymmetrical configuration with priority to downlink (e.g., 80 percent downlink, 20 percent uplink), improving spectrum efficiency is key to delivering high cell capacity.
Dual Connectivity (4G/5G) in Non-Standalone Modes
In the initial phase of Release 15, mobile operators emphasized the need to establish the framework for the dual connectivity non-standalone (NSA) method of operation. Essentially, network deployment with dual connectivity NSA means that the 5G systems are overlaid onto an existing 4G core network. Dual connectivity implies that the control and synchronization between the base station and the UE are performed by the 4G network, while the 5G network is a complementary radio access network tethered to the 4G anchor. In this model, the 4G anchor establishes the critical link using the existing 4G network with the overlay of 5G data/control. As you can imagine, the addition of a new radio, in this case a 5G new radio, alongside the existing 4G LTE multi-band carrier aggregation baseline, stresses system performance, size and interference mechanisms—posing additional challenges to be resolved when designing the new 5G NR RF front-end.
A simplified view of NSA option-3a network topology (see Figure 2) shows that in early generations of 5G networks, mobility will be handled by LTE radio anchors (control and user planes). This architecture leverages the LTE legacy coverage to ensure continuity of service delivery and the progressive rollout of 5G cells. It certainly seems the most plausible method of implementing 5G while at the same time ensuring that the integrity of data connections is maintained in areas where the backhaul and network infrastructure is not yet upgraded to 5G. However, this requires UE by default to support simultaneous dual uplink transmissions of LTE (Tx1/Rx1) and NR (Tx2/Rx2) carriers in all possible combinations of standardized bands and radio access technologies (FDD, TDD, SUL, SDL). As you might expect, this raises the technical barrier of getting multiple separate radios and bands functioning in a small device. When combined with a TDD LTE anchor point, network operation may be synchronous, in which case the operating modes will be constrained to Tx1/Tx2 and Rx1/Rx2, or asynchronous, which will require Tx1/Tx2, Tx1/Rx2, Rx1/Tx2, Rx1/Rx2. When the LTE anchor is a frequency division duplex (FDD) carrier, the TDD/FDD inter-band operation will require simultaneous Tx1/Rx1/Tx2 and Tx1/Rx1/ Rx2. In all cases, since control plane information is transported by LTE radio bearers, it is critical to ensure that LTE anchor point uplink traffic is protected.
Depending on Tx1 and Tx2 carrier frequencies and their relative spacing, intermodulation distortion (IMD) products may fall onto the LTE Rx anchor point frequency band and cause LTE desensitization. Figure 3 shows an example of IMD products generated by an intra-band LTE-FDD 10 MHz (left carrier) and NR-FDD 10 MHz (right carrier) NSA configuration.
In Figure 4, noise rise falling into LTE Rx1 band leads to moderate desensitization. However, there are multiple potential combinations of NR and LTE uplink allocations which, in some cases, may result in high desensitization. Figure 4 illustrates an example of high LTE receiver (anchor point) desensitization caused by non-contiguous RB operation of an intra-band EN-DC.
RFFE solution providers are accountable for mitigating as much interference as possible to allow for optimal signal usage in the UE. The complex nature of dual transmit LTE/NR concurrency and 5G-capable UE constitutes an even greater challenge for the NR RF front-end.
The second phase of Release 15 will include standalone (SA) operation, which uses a 5G core network that will not require backward compatibility to 4G LTE. However, the assumptions used in this article are based on the initial implementation of 5G with NSA, which is the principal deployment strategy for re-farmed bands, since SA is anticipated for new spectrum above 3 GHz.
While the sheer volume of Tx/Rx combinations resulting from 4G carrier aggregation, the addition of 5G NR and the potential challenges posed by 2×2 uplink MIMO would give anyone pause, Skyworks’ systems and engineering teams have worked diligently to resolve many of the issues in the RFFE and are described in the following sections.
All Spectrum is 5G, but Not All Spectrum is Equal
In just a year, the blueprint for 5G implementation has evolved notably. Discussions in early 2017 focused on TDD spectrum and 3.5 and 4.5 GHz ranges, which amounted to three bands. At the time, this was a manageable, targeted introduction of a disruptive technology. Fast forward to 2018, and it is safe to say that any and all mobile-owned operators will be candidates for 5G NR (see Figure 5).
A quick review of 5G announcements indicate that operators are targeting not only the sub-6 GHz spectrum, but, in order to support 5G NR, will also utilize new swaths of millimeter wave spectrum and look to re-farm their LTE assets in a lower frequency band. Hence, all spectrum assets owned by operators will be brought to bear in 5G networks.
One can expect that a combination of these spectra will be used to give consumers an enhanced mobile data experience. Previous discussions have indicated that wider bandwidth and higher order MIMO are key to delivering this enhanced user experience; and a combination of 4G and 5G-based dual connectivity systems will be integrated, in ways unique to each operator, to deliver the transformative experience.
New Challenges for 5G NR
With the previous information, now is a good time to reflect on what the new 5G radio challenges present for both smartphone designers and their counterparts in the RF front-end community. The list below is not intended to be exhaustive; rather it indicates some issues as we start development of commercial 5G products.
Wider Channel Bandwidth
- The new bands in the sub-6 GHz region will feature much larger relative percent bandwidth (n77 = 24%, n78 = 14%, n79 = 12.8%) than existing bands (B41 = 7.5%, B40 = 4.2%, 5 GHz Wi-Fi = 12.7%).
- The instantaneous signal modulation bandwidth for NR is extended up to 100 MHz in bands n41, n77, n78, n79
- Contiguous intra-band EN-DC instantaneous bandwidth is 120 MHz and 196 MHz for non-contiguous
- Conventional envelope tracking (ET) technologies are hard pressed to expand beyond 60 MHz. While the industry waits for new ET technologies to meet the 100 MHz challenge, average power tracking (APT) will be required for early proof of concept work.
High Power User Equipment (HPUE) – Power Class 2 (Specific to TDD bands n41/77/78/79)
- As mentioned previously, HPUE or power class 2 (+26 dBm at a single antenna) will increase radiated power out by +3 dB relative to power class 3 operation
- PAs will need to be designed to meet higher operating power output with more stringent waveforms
- Optimized system design will be critical to achieving minimal post-PA loss to deliver HPUE benefits
5G NR Inner Allocations Away from Channel Edge Can Be Transmitted at Higher Power
- 5G NR requires less Maximum Power Reduction (MPR), or power back-off, when reduced waveform allocations are a specified offset away from the channel edge. This enables much higher power across uplink modulation orders and addresses a fundamental coverage issue in LTE networks: the uplink power limited transmission and SNR for reduced RB allocations at cell edge.
5G NR New Waveforms and 256 QAM Uplink
- The new 5G waveforms, especially cyclic prefix orthogonal frequency division multiplexing (CP-OFDM), have a higher peak-to-average ratio and will, therefore, require more power back-off than conventional LTE waveforms
- 256 QAM modulation will be employed in uplink signals to increase data rates. This will challenge the RF front-end to maintain total error vector magnitude (EVM) below 3 percent, including PA and transceiver.
- Other issues such as in-band distortion, frame rate and clipping must be managed to achieve optimum efficiency
Cost-effective Support for 4×4 Downlink MIMO, 2×2 Uplink MIMO and Coherent 2Tx Transmission Modes
- 4×4 downlink MIMO is required in 3GPP for n7, n38, n41, n77, n78, n79 —either operating as a standalone band or as part of a band combination. This feature has been prioritized due to the significant benefits of doubling the downlink data rate and spectral efficiency, as well as the up to 3 dB advantage in receive diversity gain versus 2×2 downlink modes.
- Emphasis will be placed on size and cost reduction for the additional content to achieve this new feature set
New 5G NR Spectrum
- New bands for sub-6 GHz will extend the frequency from 3 GHz up to 6 GHz in the device
- This increase in frequency will push improvements in the complete radio front-end as the industry tries to maintain current performance while operating at a higher frequency
- There will also be new antenna multiplexing and tuning challenges, as well as in-device coexistence with 5 GHz Wi-Fi
Non-standalone Dual Connectivity and Uplink Carrier Aggregation (CA) Intermodulation
- Non-standalone operation requires dual connectivity, implying uplink CA between an LTE anchor and 5G. With the number of operating radios increasing over a substantially larger bandwidth, the challenge to maintain acceptable power for intermodulation products becomes even more complex.
- Intra-band coexistence in re-farmed 4G LTE bands
- Intra-band coexistence will be imposed in many bands as operators struggle to find available 5G spectrum.
- Intermodulation distortion (IMD) and RF front-end linearity will become pain points in the new 5G NR RFFE
Part two of this article will be in the June issue of Microwave Product Digest.
 R4-1706709. 2Tx UE for NR. 3GPP TSG-RAN WG4 Meeting NR #2, June 27–29, 2017. Qingdao, China.
 38.101-1-f20 (Rel 15.2). NR; User Equipment (UE) radio transmission and reception; Part 1: Range 1 Standalone. Retrieved June 14, 2018, from www.3gpp.org/DynaReport/38101-1.htm.
 38.101-2-f20 (Rel 15.2). NR; User Equipment (UE) radio transmission and reception; Part 2: Range 2 Standalone. Retrieved June 14, 2018, from www.3gpp.org/DynaReport/38101-2.htm.
 38.101-3-f20 (Rel 15.2). NR; User Equipment (UE) radio transmission and reception; Part 3: Range 1 and Range 2 Interworking operation with other radios. Retrieved June 14, 2018, from http://www.3gpp.org/DynaReport/38101-3.htm