The telecommunications industry is responsible for connecting billions of people and millions of businesses across the globe. The growth of the telecommunications industry has been predicated by new technologies enabling connectivity in a way that provides new and attractive features to customers and justifies the investment of upgrading and expanding the cellular network infrastructure. With the advent of data communications enabled by early 4G LTE technology, there has been an explosion of communications services that have made cellphones and cellular networks ubiquitous throughout the developed world. The next generation of telecommunications technology, 5G, promises to bring another revolution of connectivity services that extend beyond calls, texts, and simple internet and may usher in the true information age.
In order to deliver the throughput and reliability necessary to realize these new applications’ requirements, new technologies are needed. Part of the issue with achieving this next level of connectivity is the cost and complexity of transmitting and receiving quality RF signals at higher frequencies while serving a magnitude or more of additional user devices in the same area. Two key enabling technologies that can help address these challenges are gallium-nitride-on-Silicon-Carbide (GaN-on-SiC) power amplifiers and massive multi-input multi-output (mMIMO) antennas.
This article aims to provide readers with a background of the changes in requirements and design challenges associated with upgrading services and base stations from 4G to 5G-ready and 5G technology. Included in the discussion are critical details that explain how mMIMO antennas are the new normal and that new communication technologies, such as GaN-on-SiC power amplifiers, are essential in deploying 5G services that meet the 3GPP specifications and with users’ growing expectations.
Many may think that now that the 5G rollout has begun, 4G technology is on its way out. This is by no means the case, as there are still plans of providing 4G services to many areas with older 3G/4G technology and upgrades and maintenance of 4G services in preparation for future 5G base station installations. It is also likely that the network infrastructure built for 4G will continue to be used and merged into the 5G rollout, much in the same way that 2G and 3G were merged into 4G services. Hence, there is still a market for 4G technology, including LDMOS power amplifiers for 4G cellular bands.
However, the buildout of 5G services will also require new technologies and new approaches to meet the 5G expectations of hundreds of megabits per second (Mbps) of throughput in highly congested areas while enhancing reliability and reducing latency. Hence, much of the discussion and planning about large-scale 5G rollouts involves the installation of small cells that are much more densely distributed throughout urban and suburban regions. Moreover, there are 4G systems currently being upgraded from 2T2R and 8T8R MIMO to 32T32R and 64T64R mMIMO antennas in anticipation of leveraging mMIMO technology to aid in upscaling 4G services to meet with 5G expectations before the full-spectrum 5G (sub-1 GHz, sub-6 GHz, and millimeter-wave spectrum) can be deployed.
These new 5G base stations and 5G-ready 4G upgrades require many more antenna elements, with a larger number of cellular transmitters. In order to keep the size and weight of these new mMIMO antennas to a minimum, careful design and selection of RF components are required. A common design decision to reduce the size and weight of mMIMO antennas is to replace existing 4G antennas with a combined 4G/5G mMIMO antenna with embedded RF hardware. This type of densification can greatly reduce costs, especially as it pertains to tower space and wind-loading charges, but it comes at the cost of requiring higher-power-density RF transmitters that must be extremely reliable to reduce the potential of increased maintenance and service failures that are a result of component failures.
Though these concerns are significant for sub-6 GHz 5G systems, they are even greater concerns for current prototype and future millimeter-wave 5G systems. Even for sub-6 GHz 5G systems, the 3.5 GHz to 5 GHz 5G new radio (NR) cellular bands experience greater frequency-related RF losses than 4G cellular bands below 3 GHz. These greater losses are also coupled with the need for greater amplifier efficiency to account for the more complex and higher peak-to-average-power-ratio (PAPR) modulation signals used in the latest communication technologies. Hence, there is an even greater need for RF power amplifiers that can deliver high efficiency over several gigahertz of bandwidth, exhibit high reliability even while withstanding higher power densities, and be cost-effective and small enough to be assembled into compact mMIMO antennas with embedded hardware.
mMIMO 5G antenna systems have much of the similar performance considerations as 4G, with many added considerations and constraints and more stringent performance criteria. With mMIMO transmit and receive antennas placed in such close proximity, there is a heightened consideration for performance factors, such as isolation and adjacent channel power ratio (ACPR)/adjacent channel leakage ratio (ACLR). ACPR/ACLR is a measure of the leakage of power to the adjacent radio channel from a transmitter. The main contributor to ACPR/ACLR is the linearity of the transmitter’s power amplifier. A more linear power amplifier will exhibit less distortion, which results in less distortion products appearing in adjacent channels.
Power amplifier linearity and distortion — specifically, amplitude distortion and phase distortion — have other impacts on deeply modulated communication signals. Even aside from meeting the transmit mask required to meet FCC, or other telecommunications regulations around the globe, excessive distortion can also lead to a power amplifier degrading its own transmissions. Error-vector magnitude (EVM), the measure of the deviations of constellation points from the ideal, is due to power amplifier non-linearity, phase noise, and amplifier noise. Hence, it is critical to use power amplifier technology that maintains a high standard of linearity and noise, even under high load and temperature.
However, more linear power amplifiers don’t necessarily deliver better isolation metrics — transmitter to transmitter, transmitter to receiver, or receiver to receiver. High isolation is critical for mMIMO systems to prevent unwanted signals from other spatial multiplexed signals from appearing in nearby MIMO antenna elements. Even though time-domain duplex (TDD), used with 5G technology, is less susceptible to transmitter-to-receiver isolation considerations, this still doesn’t address transmitter-to-transmitter or receiver-to-receiver isolation concerns. In order to address isolation concerns, careful circuit and packaging design are necessary, which is only possible if large and high-power components, such as transmitter PAs, are compact and versatile enough to allow for creative configurations aimed at meeting stringent isolation requirements.
Other power amplifier considerations include both low current consumption and high power-added efficiency (PAE). With mMIMO antenna systems requiring arrays of transmitters and receivers, the power consumption and efficiency of each element have become critical performance criteria. This effect is amplified as future 5G rollout plans include huge numbers of dense networks being placed all throughout urban and suburban environments, from macro-cell towers to the sides/tops of buildings and telephone poles to street lamps and tunnel/subway structures. With so many more 5G base stations planned, there is a greater pressure to reduce overall power consumption, of which a transmitter’s power amplifier is one of the highest-power-consuming components.
Higher-PAE amplifiers lead to reduced overall energy consumption for the same output power but also have other beneficial effects. Higher PAE also indicates that there is less heat generated by the amplifier, as more amplifier power is used to gain up the signal energy and isn’t converted to waste heat. Less waste heat also has the benefit of requiring less heat-sinking material, which can add a significant amount of weight, size, and cost to a transmitter assembly. Moreover, lower heat generation can also lead to lower operating temperatures, which, for semiconductors, will often lead to longer lifespans and even more linear performance under high-load situations.
The aforementioned RF front-end specifications place substantial constraints on a 5G transmitter, especially 5G transmitters used with mMIMO antenna systems. This is why there is extensive research and industry efforts to develop power amplifier technologies that can meet these stringent requirements under 5G operating conditions and across new 5G spectrum. Legacy power amplifier technologies, such as laterally diffused metal oxide semiconductor (LDMOS) and gallium arsenide (GaAs) power amplifier technologies, don’t deliver on the necessary power density, energy efficiency, linearity, and cost/space requirements of 5G mMIMO systems.
In the case of GaAs amplifiers, these devices are well-suited to low-noise receiver applications but have a low bandgap voltage. This means that GaAs amplifiers have necessarily low operating voltages, which makes attaining high power densities challenging, and GaAs amplifiers are less efficient at higher power levels. The result is a much hotter and comparably higher-power-consuming device, which is less attractive for 5G mMIMO applications that demand more power density at higher efficiency levels.
Though LDMOS amplifiers have been used for high-power applications below 3 GHz for some time, LDMOS amplifiers also suffer from relatively limited thermal conductivity and comparably poorer efficiency at higher frequencies. Ultimately, this results in LDMOS amplifiers using more power and generating more heat at frequencies beyond 3 GHz while also sacrificing other considerations, such as linearity and noise (which is related to temperature for most materials).
This leaves a lot of room for a new semiconductor material to fill, that being gallium nitride. There has been much hype over GaN technology for RF applications, and in many respects, GaN devices have led to dramatic performance increases in devices ranging from long-range communications to radar. This is because GaN generally outperforms most other common semiconductor materials in power amplifier figure of merit (PAFOM) — namely, power density, reliability, thermal conductivity, linearity, and bandwidth.
There are some nuances to GaN semiconductors, as GaN is generally epitaxially developed on an insulating substrate. Hence, GaN devices could use a variety of substrates, such as sapphire, silicon (Si), Silicon Carbide (SiC), GaN, and even diamond. Due to process maturity, cost, and other design constraints, GaN for RF applications is typically widely available as either GaN on Si or GaN on SiC.
For much the same reasons that GaN is superior to Si-based LDMOS devices for high-frequency RF applications, GaN on SiC is superior to GaN on Si for 5G mMIMO applications. Much of the performance advantages of GaN on SiC over GaN on Si derive from SiC being a much more rugged material, with better thermal conductivity and a better lattice match to GaN. This means that under a high load, GaN on SiC devices can be run hotter, with less wear and tear and with higher power efficiency than GaN on Si. Moreover, this implies that for the same power output, GaN on SiC power amplifiers can be smaller and require smaller heat sinks than GaN on Si devices. Moreover, the reliability of GaN on SiC has been thoroughly examined and approved for critical U.S. Department of Defense (DoD) and aerospace applications.
The deployments of 4G and 5G systems are likely to leverage mMIMO technology to offer the best coverage and capacity to customers with ever-higher expectations from modern communication services. GaN on SiC power amplifier technology offers the best performance and cost requirements for mMIMO systems compared to GaN on Si and LDMOS technologies. Wolfspeed’s GaN on SiC technology has been approved for use in high-reliability telecommunications, military, defense, and aerospace applications and offers lower total life-cycle costs than both GaN on Si and LDMOS.
