Proper PA development, validation, and characterization are important because a PA often accounts for a significant portion of a transmitting device’s power consumption.

Silicon has proven a reliable, cost-effective, and easy-to-manufacture material in most chipsets and components. As the world moves more and more to a digital, interconnected, and device-driven ecosystem, however, the need for more performance, throughput, and efficiency increases. While silicon still has endless use cases, it can’t meet the performance requirements needed for 5G New Radio (NR), which requires higher power, higher operating temperature, and better efficiency. Wide bandgap semiconductors will help meet this need. When it comes to high-power RF applications, Gallium Nitride (GaN) is set to change the high-power RF power amplifier (PA) game.

Depending on the application, the definition of high-power may change. For now, a high-power PA will have a P1dB compression point of at least 30 dBm, perhaps as high as 70 dBm. Due to the lower bandgap, traditional power-amplifier topologies such as HBTs and pHEMT amplifiers on GaAs substrates are not optimal. Instead, high-power PA designers typically opt for either LDMOS FETs on an SiC substrate or HEMT amplifiers built with a GaN layer on top of a SiC substrate. Figure 1 shows the differences in bandgaps among semiconductor materials.

GaN offers many advantages over traditional semiconductors. Being a wide bandgap device means that it offers better power efficiency at high frequencies, higher operating temperatures, higher power, and better power density than other processes. Because of those differences, you’ll need to alter your test strategy.

Although the exceptional power, temperature, efficiency, and frequency properties of GaN have been known for decades, certain technical challenges have limited its viability in commercial applications. For example, the ability for GaN ICs to be produced using traditional silicon semiconductor manufacturing technology has opened the door to GaN PAs on a larger scale. Furthermore, today’s increased need for higher-power and more efficient components that function across various frequency bands and compatibility with 5G NR and legacy cellular standards (Figure 2) means leads to a significant increase in interest.

Because of their wide bandgap characteristics, GaN PAs are well-suited to address many issues when implementing modern base station infrastructure for cellular communications. GaN PAs could greatly benefit the development of wireless infrastructure. Applications include the need for greater power efficiency, operation across multiple bands and frequencies that accommodate both new and legacy cellular standards, and efficient operation across wideband waveforms.

A traditional base station (Figure 3) includes three devices: a baseband unit (BBU) at the base of the tower, a remote radio unit (RRU) at the top of the tower, and an antenna. The RRU will include the hardware for separating the uplink and downlink signals, amplifying the signals, up/down converting, and signal conditioning. The high-power PA resides on the TX path within the RRU. In the base station, GaN PAs present many benefits, including the ability to accommodate multiple frequency bands to support multiple devices simultaneously.

Despite all the potential benefits, GaN PAs present many challenges in tests due to their unique characteristics. Some of these include:

• Complex test setups
• GaN linearization
• Accurate power measurements
• Time domain synchronization
• Novel processes and technologies

Complex test setups

A high-power PA is often a combination of multiple smaller PAs. Sometimes multiple stages are cascaded in series into a single high-gain PA. Another common amplifier architecture is known as a Doherty amplifier, in which two amplifiers connect in parallel, both receiving a split copy of the signal. One amplifier (known as the carrier PA) is tuned to accurately amplify the lower-power portion of the signal while the other amplifier (known as the peaking PA) is tuned for the higher-power portion. The signals are then recombined, giving improved signal fidelity across both operating regions.

Even with these multistage techniques, the amplifier’s output power is often still insufficient for commercial applications. A driver amplifier boosts the signal power ahead of the high-power PA. The driver amplifier is normally optimized for high linearity and low noise figures because its input is closer to the noise floor.

In addition to the physical test setups, the bring-up of GaN PAs can also be more intricate and involved than with other RF power amplifiers. For example, DC biasing must be applied to the DUT before generating or acquiring any RF waveforms.

Linearization

Base stations must analyze uplink signals and generate downlink signals across multiple bands simultaneously. With multiple antennas and signal chains active on a single tower, congestion can occur, both physically as towers become more crowded and spectrally as cellular traffic increases. This drives designers to optimize signal chains in several ways. Some signal chains need optimizing for multiple bands, meaning the PA must operate across these bands simultaneously. This results in strict requirements on out-of-band spectral emissions, as nearby antennae are transmitting and receiving at those nearby frequencies.

In addition, GaN PAs tend to behave with less linearity than more traditional silicon or GaAs-based PAs that operate at lower power. Because of this, digital pre-distortion (DPD) becomes an important method for maintaining a delicate balance between signal fidelity and a clean spectrum.

Power measurements

Operating at high power levels will also impact the accuracy of power measurements. Accurate power measurements require a process called system de-embedding or system calibration, in which the test system compensates for the accuracy of the signal generator and analyzer and for the losses or amplification in your signal chains.

Time-domain synchronization

A high-power PA’s power consumption pushes designers to invest in optimizing power efficiency. This is a critical metric for any infrastructure hardware provider because energy is a primary cost of operating a base station. Designers should characterize and optimize an amplifier’s power efficiency. One important strategy for conserving energy is managing when the PA is enabled. Some PAs provide an enable pin that can be toggled, while others require the power supply to start and stop at the proper time. Either way, a base station needs synchronization among the DUT, power supply, and signal generator. This synchronization is especially important for time-division duplexing (TDD) waveform tests, where certain time slots in the transmission are reserved for uplink communication while disabling the downlink chain.

Novel processes and technologies

The process for creating GaN components is still evolving and has a major influence on performance. For example, impurities in the GaN substrate can lead to charge trapping, which in turn causes gain failure in certain signal situations. Characterizing these types of phenomena is vital to understanding the performance of a full RF system based on GaN components. Using standard testing with small signals or ACLR of wideband modulated signals is not enough. You need more information on the phase impact on real-world signals and very tight synchronization between RF and DC measurements.