Audrey Martin

Two-Stage GaN HEMT Amplifier With Gate–Source Voltage Shaping for Efficiency Versus Bandwidth Enhancements

In this paper a two-stage 2-GHz GaN HEMT amplifier with 15-W output power, 28-dB power gain, and 70% power-added efficiency (PAE) is presented. The power stage is designed to operate under class F conditions. The driver stage operates under class F-1 conditions and feeds the power stage with both fundamental and second harmonic components. The inter stage matching is designed to target a quasi-half sine voltage shape at the intrinsic gate port of the power stage. The goal is to reduce aperture angle of the power stage and get PAE improvements over a wide frequency bandwidth. In addition to the amplifier design description, this paper reports original time-domain waveform measurements at internal nodes of the designed two-stage power amplifier using calibrated high-impedance probes and large signal network analyzer. Furthermore, waveform measurements recorded at different frequencies show that aperture angle remains reduced over large frequency bandwidth. In this study, a PAE greater than 60% is reached over 20% frequency bandwidth.

Modeling of trap induced dispersion of large signal dynamic Characteristics of GaN HEMTs

We propose here a non-linear GaN HEMT model for CAD including a trapping effects description consistent with both small-signal and large-signal operating modes. It takes into account the dynamics of the traps and then allows to accurately model the modulated large signal characteristics that are encountered in telecommunication and radar signals. This model is elaborated through low-frequency S-parameter measurements complementary to more classical pulsed-IV characterizations. A 8×75μm AlInN/GaN HEMT model was designed and particularly validated in large-signal pulsed RF operation. It is also shown that thermal and trapping effects have opposite effects on the output conductance, thus opening the way for separate characterizations of the two effects.

Highlighting trapping phenomena in microwave GaN HEMTs by low-frequency S -parameters

This paper presents an original characterization method of trapping phenomena in gallium nitride high electron mobility transistors (GaN HEMTs). This method is based on the frequency dispersion of the output-admittance that is characterized by low-frequency S-parameter measurements. As microwave performances of GaN HEMTs are significantly affected by trapping effects, trap characterization is essential for this power technology. The proposed measurement setup and the trap characterization method allow us to determine the activation energy Ea and the capture cross-section σn of the identified traps. Three original characterizations are presented here to investigate the particular effects of bias, ageing, and light, respectively. These measurements are illustrated through different technologies such as AlGaN/GaN and InAlN/GaN HEMTs with non-intentionally doped or carbon doped GaN buffer layers. The extracted trap signatures are intended to provide an efficient feedback to the technology developments

Characterization and circuit modeling of Graphene Nano Ribbon field effect transistors

This paper reports on the pulsed I-V and microwave characterizations of a Graphene Nano Ribbon FET (GNR-FET) for nonlinear electrical modeling. The extraction method of model parameters is based on the characterization of three specific technological structures called PAD, MUTE and FET (integrating only the coplanar access structure, the FET without graphene, and the entire GNR-FET) respectively. The differences between DC and pulsed I-V characterizations of the GNR FET and the evolution of its multi-bias S-parameters are investigated and compared to simulations. The nonlinear modeling of GNR FET is becoming of prime importance along with technological efforts to study the actual potential of this emerging technology.

Study and design of high efficiency switch mode GaN power amplifiers at L-band frequency

Activities have been carried out to determine the best electrical operating conditions of GaN HEMT that enable maximum power added efficiency at L-Band for Switch Mode Power Amplifiers (class F, inverse class F and class E). Satellite Radio navigation applications (Galileo) are targeted. Maximization of power added efficiency is of prime importance to save DC power consumption, reduce self heating effects and improve reliability of power amplifiers. At 50V drain bias, a maximum power added efficiency (PAE) of 72% and 40.3 dBm output power (Pout) are obtained using class-F operating conditions at 2dB gain compression while a 75% PAE and 41.0 dBm Pout are obtained using class E at 3dB gain compression.

Implementation of dual gate and drain dynamic voltage biasing to mitigate load modulation effects of supply modulators in envelope tracking power amplifiers

This paper presents a combination of dynamic gate and drain biasing techniques applied to a S-Band - 10 W GaN power amplifier. A GaN-based drain supply modulator (DSM) and an integrated gate biasing circuit have been built and connected to a GaN RF power amplifier (RFPA). The complete circuit architecture is then implemented in a test bench for the study of the envelope tracking power amplifier (ETPA). The work reported here focuses on the nonlinear coupling between the drain supply modulator and the RF power amplifier and proposes a solution to mitigate the load modulation effect of the supply modulator that is prejudicial for the overall efficiency and linearity performances. It consists in implementing an appropriate dynamic gate bias control of the RF power amplifier. The experimental validation of the study is demonstrated here for a QAM-16 (2 MSymb/s) modulated carrier at 2.5 GHz. At 38 dBm output power, dynamic load variations of the drain supply modulator (in the 30-500Ω range) versus instantaneous input power level variations have been drastically reduced and maintained to a 40Ω average value. The measurement of the signal constellation diagram at the ETPA output was found to be a well suited way to optimize the tuning of bias trajectories. A measured minimal error vector magnitude (EVM) of 2% and a power added efficiency (PAE) of 40% are obtained when optimal tuning is reached.

High speed and highly efficient S-band 20 W mixerless vector power modulator

This paper presents a performance evaluation of an original highly efficient and linear GaN-HEMT Vector Power Modulator (VPM) based on the design of a two-stage saturated variable gain (SVG) amplifier and a multi-level discrete supply modulator (SM). The proposed novel architecture for transforming a digital baseband data stream into an RF Vector modulated power waveform (RF Power DAC) is validated using a specific laboratory test bench. The main objective of this study is to merge signal modulation and DC to RF energy conversion functions into a single and compact GaN based mixer-less circuit. Using high-voltage 50 V GaN technology, a 20 W S-band vector power modulator having overall average PAE of around 40 % is reported. The concept demonstrator is experimentally validated up to 100 Msymbols/sec 16-QAM modulation scheme. Functional time alignment with phase and amplitude compensation procedure focusing on measured constellation at 40 Msymbols/sec enables to reach excellent EVM performances of around 3.2 %.

On-wafer time-domain measurement of pulse-to-pulse stability for microwave power GaN HEMT

For the first time, on-wafer time-domain envelope measurements of pulse-to-pulse (P2P) stability are reported in this paper. In the case of a radar burst, these on-wafer measurements are performed on a 10W power GaN HEMT in S-Band by using a digital quadrature demodulation (DQD) as complex envelope extraction technique. The impact of an irregular RF pulse train on the measured P2P stability at device level is illustrated by the influence of load impedance, input power and bias conditions.