Scratching the Surface of 5G
By Mike Rosa, PhD
The 5G wireless rollout, expected to occur over the next few years, will have a major impact on both the number and types of ICs in end-user devices, and on the base stations and repeaters needed to transmit the higher frequency signals.
Figure 1. Expected to be released in 2020/21, 5G will dramatically improve on existing 4G technologies but not without its fair share of trade-offs and technological challenges.
The 5G standard is expected to deliver 10 Gbps of bandwidth— up to 10X the data rates achievable using the advanced forms of 4G— and sub-5ms latencies. The capabilities also will impact the amounts of data generated in a 5G ecosystem, increasing demand for servers, storage, and photonic devices.
The 5G standard calls for a two-stage shift to higher frequency signals, starting out in the sub- 6-GHz range—much of it in the 3.5- to 4.5-Ghz range—and then moving to the more challenging millimeter-wave regime in the 24- to 43.5-GHz range. The millimeter waves transmit short distances and will face absorption issues, so the 5G standard will be a combination of 4G and 5G to support urban (5G) and long distance (4G) networks (see figure 1).
Progress is being made. Speaking at a recent SEMI 5G-focused event in Texas, Ian Wong, senior group manager for advanced wireless research at National Instruments (Austin, Texas), said by the end of this year “most of the baseline plumbing will be anchored” within the main 5G standards group.
China has about 1.7 million base stations deployed nationwide, and by 2020 only about 10,000 of them will be 5G-capable base stations. Wong said telecom operators worldwide want to move beyond the smartphone and the business-to-consumer models. “The carriers have been bleeding capex, and now they need to move beyond the smartphone consumer to the Internet of things, industrial IoT, and business-to- business models,” he said.
“From the standpoint of a test and measurement company, we are in the thick of it now and need to be ready today,” Wong said.
The impacts on the semiconductor industry will be widespread. End-user devices and base stations will need to manage multiple-input and multiple-output (MIMO) and beam-steering technologies, which translate into more channels and expanded demand for bulk acoustic wave (BAW) filters, antennae, power management, and other devices.
To help cope with the power issues, systems will employ more sophisticated envelope-tracking technologies. While not completely new (they were introduced during the rollout of 4G), these chips are used in RF power amplifiers to track the signal and boost the power on an as-needed basis rather than constantly supplying high power. With the speed increase of 5G these envelope-tracking technologies shift from employing laterally diffused metal oxide semiconductor (LDMOS) and gallium arsenide (GaAs) to gallium nitride (GaN) in order to manage the higher power while still accommodating the even higher switching speeds required.
The need to guide the signal also will require high-speed devices beyond what is seen today. The absorption of high-frequency 5G signals will require the transmission beam to be electronically “steered” in order to minimize losses and optimize the transmission efficiency of the system. And due to the short range of the millimeter-wave signals, multiple repeaters will be needed, driving up the chip count for 5G deployments.
The RF front-end modules (FEMs) will need to handle a larger number of higher-frequency 5G signals. That entails more high-performance BAW filtering, and may further the transition from aluminum nitride (AlN) to scandium-doped AlN (ScAlN) films and then beyond into newer films such as lithium niobate (LiNbO3), which can potentially offer electromechanical coupling efficiencies (read transmit/ receive efficiencies) twice that of the latest ScAlN films under development today.
While the technical challenges facing the 5G wireless standard are not to be trivialized, panelists at the SEMI Austin 5G event said the basic 5G air interface technology has already been demonstrated, with the best minds in the global wireless community working together to solve the remaining challenges.
Robert Topol, Intel Corporation’s general manager for 5G business and technology, said companies in the automotive, drone, and industrial space are coming to Intel’s 5G lab in Austin almost daily to test out the 5G technology, including technically challenging millimeter-wave transmissions.
“We have [5G] running in the lab today, and it works. It is really not as complex as some people think. Of course, to run a nationwide or global 5G network, there is still a lot that we need to do. But most of the work on the way the air interface works has been done,” Topol said.
While 5G can be thought of as a fifth-generation wireless standard, Topol said within Intel the thinking is that 5G is the “first generation of a connected compute” wireless technology. With automotive, mobile, and industrial systems able to connect to the cloud at very low per-bit costs, there will be a synergy between AI and connected computing.
“The timing of AI and 5G working together is major inflection point. We will be able to connect systems with deep-learning accelerators [in the cloud]. That won’t happen in 2020, when 5G debuts, but when we find that inflection point between the two, it will be pretty impactful for the industry,” Topol said.
Beyond thinking of 5G as a faster network, Topol said “deployment of 5G comes down to the cost-per-bit. It cannot be just a 10X improvement, because data consumption will go up 10X during the time frame that 5G deploys. It [cost-per-bit transmission] has to improve exponentially,” he said.
Andrea Lati, research director at VLSIresearch (Santa Clara, California) said the electronics industry is moving from a mobility-driven model to an AI-driven industry. “AI needs lots of data, from many sources, with lots of compute power. 5G will be the platform that allows data to flow between the IoT and the cloud,” Lati said.
Asked at the Austin event about the role of 5G in the automotive space, Bill Morelli, senior research director at IHS Markit, said 5G networks would serve as another layer of safety for autonomous drivers. While the car’s sensors will be responsible for seeing if an oncoming vehicle is approaching, the 5G network will look out more broadly, to determine whether cars might be blocking the road beyond a curve well ahead. Also, vehicle-to-vehicle communications will enhance mapping, traffic, and safety information. And since drivers will no longer be tied down to steering the car, 5G wireless could be used to support in-vehicle conference calls and other forms of productivity and infotainment, he said.
Morelli said IHS sees three broad categories within the still developing 5G standard. Enhanced mobile broadband includes both fixed wireless access (FWA) and enhanced mobile broadband (eMBB). FWA will be the first to ramp, he said (see figure 2).
Figure 2. 5G deployment is expected to begin commercial trials in 2018, with a full commercial mobile broadband launch expected in 2020. (Source: IHS Markit, 2017)
A machine-to-machine application of 5G is often referred to as massive IoT. These machine-type communications will require low-power and low-data-rate capabilities, supporting a large number of connections.
5G deployment in the mission-critical arena, including some industrial, robotic, medical, and automotive applications that require ultra-reliable and extremely low latency capabilities, is likely to take the longest. The mission-critical use cases depend on significant technical innovations, Morelli said. He added that most of the world’s telecoms are in rough alignment on how to use available spectrum for 5G networks, and wireless telecoms need it to develop new markets in the industrial, automotive, smart city, and commercial spaces.
“5G requires a heavy focus on technical innovations to break into largely uncharted territories, such as millimeter-wave transmissions,” Morelli said.
For additional information, contact firstname.lastname@example.org.