Solving 5G mmWave Phased Array System Efficiency Challenges with the Qorvo QPF4010
by David Schnaufer, Technical marketing communications manager & David Rahn, Senior product line manager, Qorvo
5G deployments are in full swing worldwide. In the U.S., 5G is being rolled out in several U.S. cities and in NFL stadiums. South Korea already has 2 million 5G subscribers, while China’s three state telecom providers have launched 5G services ahead of schedule and UK operator Telecom EE has enabled 5G in several cities. To support these deployments, mobile network operators (MNOs) are densifying their networks with additional base stations that use phased arrays to operate at the higher frequencies above 3 GHz used for 5G, including mmWave bands. These base stations require RF front-end (RFFE) modules, such as Qorvo’s QPF4010, that can solve the power amplification and efficiency challenges associated with phased arrays operating at these frequencies.
RF Front-End Technologies for 5G Base Station Power Requirements
Several different RFFE technologies are used in 5G base stations and other infrastructure devices. Each technology has its place in the 5G ecosystem. For base stations, the choice of technology depends largely on the required output power, expressed as Effective Isotropic Radiated Power (EIRP), as well as other system requirements (see Figure 1). EIRP is the maximum amount of power that can be radiated from an antenna, which determines the range and signal coverage.
MNOs are implementing 5G network densification to increase network capacity and support rising consumer and enterprise demand. This densification requires more base stations, repeaters and terminals. To support areas with high data traffic, MNOs are focusing on tiny base stations that use phased array antenna modules (PAAM) to deliver 55-65 dBm EIRP. The key challenge of these tiny (street level) base stations is that they require very small heat-sinks and so the efficiency of the front ends is critical.
As shown in Figure 1, multiple front-end technologies can be used to meet the EIRP requirements but GaAs RFFEs such as the QPF4010 provide better performance with far fewer antenna elements. As a rule-of-thumb, lower efficiency silicon technologies can be used for indoor deployments where environment conditions are less extreme but for outdoor applications the efficiency benefits of a compound-semiconductor front-end is critical.
One of the biggest challenges for MNOs is balancing base station performance, cost, size, weight and power budgets. For PAAM base station applications, GaAs can attain EIRP targets with lower total power dissipation than other power amplifier technologies like Silicon-Germanium (SiGe) or Silicon-On-Insulator (SOI). This lower power dissipation is better for tower-mounted system design, as it results in lower system weight, complexity, and power requirements.
Comparing Technologies
For 60 dBm EIRP applications, GaAs and GaN have clear advantages over silicon-based technologies in balancing power, efficiency, and the number of active elements. Using the GaAs-based QPF4010, a phased array would require half as many active elements as with even the best silicon-based RFFE. For example, to achieve a 60 dBm EIRP, a phased array would require approximately 128 elements using the QPF4010 versus 256 elements with a silicon based RFFE, as shown in Figure 2.
GaAs and GaN are also the most efficient technologies available for mid- to high-power base station applications, which means less heat dissipation and lower cooling requirements. Efficiency, typically measured in percent, defines the power dissipated by the PA. Devices that are more efficient produce less wasted energy, which mean they generate less undesirable heat as a byproduct. This means there is less need for heatsinks or fans to pull heat from the system.
Figure 2: Comparison of RFFE technologies for 28 GHz 5G phased arrays 
Figure 3: Comparing RFFE noise figures for 28 GHz 5G phased arrays
Another system requirement that is closely related to the EIRP, power consumption and number of active elements is the noise figure. The lower the noise figure in a system, the better the sensitivity and range of the signal. A low noise figure also reduces system power consumption and complexity. As seen in Figure 3, GaAs and GaN have a much lower noise figure than SiGe, thus reducing the number of antenna elements and system complexity while increasing overall system performance.
Qorvo QPF4010 FEM
The QPF4010 is a GaAs front-end module (FEM) containing a switch, LNA and PA. It is optimized for phased array base stations and terminals using the 5G n258 frequency band (24.25 to 27.5 GHz). Fabricated on the Qorvo 90 nanometer GaAs pHEMT process, this part offers best-in-class efficiency and linearity. Because it can reduce the system complexity, thermal requirements and number of array elements, this FEM is suitable for use in many small form factor applications. For example, the QPF4010 is ideally suited for phased terminals and repeaters used in train stations, subway terminals, stadiums, cities and other locations. The QPF4010 can easily be combined with a hybrid beamformer as shown in Figure 4.
The QPF4010 is a compact 4 mm x 4 mm module that can operate with an average output power of 14 dBm with 4% error vector magnitude (EVM) at 64 quadrature amplitude modulation (QAM), with a 400 MHz modulation signal.
As the end-user data rate requirements increase, high RFFE linearity becomes paramount. The use of 64 QAM in 5G supports higher data rates, but the tight constellation density means that the system must operate with low EVM in order to distinguish each point in the constellation. The ability of the QPF4010 to transmit at 14 dBm with 4% EVM helps system designers build more efficient 5G PAAM base stations that support higher data rates.
Figure 4: QPF4010 functional and system block diagram 
Figure 5: QPF4010 performance: EVM % versus Pout and gain flatness over a frequency and temperature range
In 4G and 5G, gain flatness requirements have also become more stringent. To meet system linearity requirements, a gain flatness of 1 dB or less is required; this also helps achieve the tighter EVM specifications (Figure 5).
As stated earlier, the noise figure is an important figure of merit associated with the EIRP, power consumption and the number of active elements needed in a system. It is one of the most important parameters for the radio receiver. The noise figure of the receiver is a measurement of noise added by the receiver as signals pass through it. The noise figure varies with temperature. The receiver’s noise figure limits its sensitivity—the level of signal that that the system can receive.
As shown in Figure 6, the QPF4010 has a low measured noise figure of 4 dB at 25⁰C across the n258 band. This lower noise figure improves receiver sensitivity. Additionally, because of its high linearity the QPF4010 allows stronger signals to be received (like when a user is close to the base station) without saturating the front-end and desensitizing the receiver.
In Summary
Deployments of 5G PAAM base stations are accelerating due to consumer and business demand for more reliable, faster data networks. With its high-power output, linearity, and efficiency, the QPF4010 is a good match for 5G PAAM base stations with 32, 64 or 128 antennas. Using GaAs technology, the QPF4010 supports a high EIRP with superior EVM to reduce the base station phased array element count and complexity. The QPF4010 complements the other modules in Qorvo’s growing portfolio of mmWave products, including the QPF4005 and QPF4006 transmit/receive GaN based integrated front-end modules which are widely used in base station applications delivering 65 dBm EIRP and above.
(603)
