Search

This paper presents a quantitative metric for memory effects in power amplifiers (PAs) and applies it to various active device technologies and wireless system contexts. The proposed metric is mathematically founded in the dynamic two-tone distortion response, has a clear physical meaning in the important field of PA linearization and can be easily evaluated from either harmonic balance simulations or measurement data gathered in a microwave laboratory. In addition, a memoryless PA linearizer, optimum for reducing the integrated intermodulation distortion (IMD) power in the operation bandwidth for a two-tone excitation, is derived, providing a rigorous figure-of-merit of PA linearizability under static IMD compensation. The application of this figure-of-merit is then illustrated for three different PA prototypes based on Si LDMOS, InGaP/GaAs heterojunction bipolar transistors, and GaN high electron-mobility transistors, designed for 900-MHz (global system for mobile communications), 2.1-GHz (wideband code division multiple access), and 3.5-GHz (WiMax) wireless systems, respectively

This article deals with the nonlinear memory effects observed on both AM/AM and AM/PM conversions and swept frequency-separation two-tone intermodulation-distortion (IMD) tests made on microwave power-amplifier (PA) circuits. This study is theoretically introduced via an analytic Volterra series analysis, then complemented with large-signal harmonic balance and envelope simulations of a GaN HEMT-based microwave PA built with different load terminations.

This paper presents a nonlinear equivalent circuit model of microwave power GaN high electron-mobility transistors (HEMTs), amenable for integration into commercial harmonic balance or transient simulators. All the steps taken to extract its parameter set are explained, from the extrinsic linear elements up to the intrinsic nonlinear ones. The predictive model capabilities are illustrated with measured and simulated output power and intermodulation-distortion data of a GaN HEMT. The model is then fully validated in a real application environment by comparing experimental and simulated results of output power, power-added efficiency, and nonlinear distortion obtained from a power amplifier.

This work responds to the commonly formulated question in PA design of “Why do GaN HEMTs' nonlinear behavior seems so distinct from the one of their Si LDMOS, GaAs MESFETs or HEMTs counterparts?”. Starting from some recent results on the origin of AM/AM and AM/PM distortion in power amplifiers made for these two device types, we demonstrate that there is no fundamental reason why the devices should be different, except that the formers suffer from a noticeable low-frequency dispersion. Then, we dig into this particular aspect of GaN HEMT operation, to show that this charge trapping-related phenomenon is responsible for a severe self-biasing capable of inducing soft-compression of an otherwise almost flat AM/AM gain plot - when measured with static CW tests - which, in fact, corresponds to a severe gain expansion when the AM/AM is assessed with more realistic dynamic tests performed with real communication signals. And it is this class-C PA like AM/AM that is responsible for the recognized GaN HEMT nonlinearity.

The carrier power amplifier (PA) of a Doherty architecture is designed with a load that maximizes efficiency for a particular load-pull ratio. With the increase of the input drive, the peaking PA provides the necessary current to reduce the carrier PA load impedance. From this point on, the carrier PA operates at a constant output voltage, suffering a severe voltage gain compression at its terminals. Since the input Miller reflected C gd depends on this gain, the device's total input capacitance changes substantially with drive level, inducing a significant input phase shift. Hence, in this paper we address the impact of this capacitance variation on the AM/PM distortion characteristic of a GaN HEMT based Doherty PA.

The Doherty architecture is presented as an ideally linear efficiency enhancement technique. The theory behind this statement assumes that the output of the Carrier and Peaking Power Amplifiers (PAs) are ideal current sources, which is far from the truth in practical implementations. Both the Carrier and Peaking PAs are composed of nonlinear current sources suffering from both AM/AM and AM/PM distortion. Furthermore, the AM/AM and AM/PM characteristics of the Carrier and Peaking PAs, when operating in a Doherty arrangement, are very different from the ones expected from their single-ended operation, especially in GaN HEMT based Doherty PAs. This is exactly the topic addressed by this paper.

Despite the integrability and reconfigurability offered by fully digital wireless transmitters, their efficiency is still low, and no effective ways exist to diagnose and predict its degradation at higher back-off powers. Having found that it is the linear dependence of the switching losses with the output amplitude plus the inevitable constant losses that cause efficiency degradation at higher back-offs, an efficiency model was derived for carrier amplitude burst transmitter architectures. This model, not only allows the designer to predict the efficiency for a given modulation and output stage, but also allows one to identify the origin of the losses given an observed efficiency curve.

Signal-dependent phase variation, AM/PM, along with amplitude variation, AM/AM, are known to determine nonlinear distortion characteristics of current-mode PAs. However, these distortion effects have been treated separately, putting more weight on the amplitude distortion, while the AM/PM generation mechanisms are yet to be fully understood. Hence, the aim of this work is to present a large-signal physical model that can describe both the AM/AM and AM/PM PA nonlinear distortion characteristics and their internal relationship.

The potential high efficiency of RF carrier-burst transmitters has drawn the increasing attention of various researchers. This architecture uses an on-off mode of a constant-amplitude phase modulated carrier, according to the selected envelope-to-duty-cycle conversion (PWM or sigma-delta), requiring an implementation with two transistors, often in class-D configuration. This paper proposes a significant modification to this traditional architecture, where now the phase modulated carrier is always on - and thus allowing a single transistor implementation - but where the amplitude is impressed into the carrier phase according to a bi-phase code. A proof-of-concept prototype tested with a CDMA2000 signal achieved 59% of average efficiency and more than 34dBc of ACPR for an integrated output power of 24.7 dBm.

Carrier amplitude-burst transmitters can ideally deliver 100% efficiency and perfect linearity. The amplitude to time conversion performed by a PWM or sigma-delta modulator allows the usage of switched mode power amplifiers, at the expense of introducing a large amount of out of band noise, which must be removed by a narrow bandwidth band-pass reconstruction filter. However, practical implementations suffer from several problems whose cause is the careless design of the amplifier-filter connection, which prevent the achievement of those ideal performance figures. This paper discusses how the linearity of these transmitters is affected by the interaction between the switched-mode amplifier and the band-pass reconstruction filter providing design rules for this amplifier-filter interconnection. The presented results are obtained by both simulations and practical measurements.

The importance of the bandpass reconstruction filter, BPF, in the proper operation of carrier-burst transmitters is widely accepted. However, it is often analyzed as if it could be considered isolated from the other system's components. To fill this gap, this paper presents a detailed study of the carrier-burst transmitter BPF, addressing both its most significant characteristics and its influence in the system. During this analysis, it was found that both the characteristics of the BPF and of the amplifier-BPF interconnection are extremely important to the correct behavior of the system. The proposed theoretically driven design rules are verified by several simulations and supported by corresponding laboratory measurements.

Limit-cycle or self-oscillating power amplifiers (SOPAs) are seen as a compact option to achieve efficient yet linear amplification of either complete RF signals or only their base-band envelopes. Unfortunately, one of their undesired drawbacks is that their self-oscillating frequency, and thus available bandwidth, is reduced when a significant propagation delay exists in the feedback loop, typically from the power stage. To remedy this SOPAs' pitfall, this paper presents a method to increase the self-oscillation frequency for a fixed propagation delay and loop filter. This is achieved using a relay that exhibits a novel kind of hysteresis. For that, the theory of its working principle is discussed, and a suitable implementation is proposed. To validate the concept, simulations are carried out to compare the behavior of existing types of SOPA with the proposed one.

This paper proposes an innovative power amplifier, PA, low-pass equivalent behavioral model, BM, intended to enable the simultaneous prediction of linearity and efficiency in modern wireless transmitter chains. Contrary to the main PA behavioral modeling stream that has been exclusively focused in predicting the device's input-output power characteristics in the fundamental zone, this new model also provides the system designer with predictions of the PA absorbed dc power, this way enabling power efficiency calculations. For that, this new BM gives the fundamental output current or voltage across the load, as well as the base-band current flowing from the power supply to the PA dc bias terminal, I dc . Beyond the model format, a model extraction methodology and the corresponding laboratory test bench are shown. These were then validated comparing static and dynamic I dc -versus-V in modeled and experimental data.

This article attempts to improve the average efficiency and linearize the work-horse of RF power amplification- the class-B (or class AB) PA-relying, for example, on predistortion, dynamic load modulation or ET. Nevertheless, whenever higher efficiencies are desired, the current-mode class-B PA must be replaced by switchedmode PAs (for example, class-E or class-F), completely modifying the paradigm of RF amplification. In fact, as these are saturated circuits, they cannot respond to excitations of variable amplitude; and thus the original signal envelope must be restored by operating the PA as an AM modulator via dynamic supply adaptation. A good example of this is the EER arrangement.

In this paper, two vector nonlinear characterization procedures are presented, aimed at improving available GaN HEMT models for an accurate reproduction of the device behavior operating as a current source and in switched-mode RF power amplifiers (PAs). In the case of the more traditional linear amplifying classes, a technique for simultaneously extracting the higher order derivatives of the I ds (V gs ), I gs (V gs ) and C gs (V gs ) transistor nonlinearities, along a desired load line, is described. This procedure conveniently isolates and models their contributions to the device intermodulation distortion (IMD) behavior. In the case of highly efficient switched-mode PAs, employed under drain modulation condition, a modified procedure for isodynamically measuring the higher order derivatives of the V dd -to-AM and V dd -to-PM amplifier profiles is put into consideration, as a way to refine the triode region reproduction for ON-OFF operation. The test set-ups, based on the combination of vector signal analysis and vector signal generation capabilities, are described in detail, together with some analytic expressions for the parameter extraction. Measured results for a 15 W device from Cree are finally included.

In this paper, the impact of self heating on the linearity of a drain modulated GaN HEMT power amplifier (PA) is studied. After characterizing the frequency response of junction temperature rise to power dissipation, as well as the effect of this variable on the PA modulation profiles, the dynamic envelope trajectory, under a two-tone test signal, appears to be slightly dependent on frequency spacing. Self heating related effects are shown to be negligible in presence of other nonidealities, as the case of feedthrough. In this way, punching a vector hole in the two-tone I/Q diagram and applying memoryless digital predistortion (DPD), will allow identifying the influence of self heating effects, over adjacent channel distortion. Finally, this long-term memory effect is proved to be responsible for only a minor residual distortion in a linearized EDGE polar transmitter.

This paper reports the design and implementation of a RF GaN Oscillator for test in space. This circuit/system will be applied in a space test of the European Space Agency and it will be one of the first experiments in space using European GaN devices. The main objective will be to study the material radiation hardness. In this paper we will present the proposed circuit for test and its main design procedure.

This paper describes the design of a class E power amplifier (PA) destined to be used as the final stage of a polar transmitter under IEEE 802.16e Mobile WiMAX signal excitation. Based on a 60 W GaN HEMT device, the load condition for optimum efficiency is set close to the maximum value of the power generating function, as defined in [1], for an OFDMA modulating signal. The possible impact of such average-based design on the AM profile is also put into consideration.

In this paper, we discuss the available transmitter architectures in a Software Defined Radio (SDR) context and present a suitable polar architecture implementation under three different modulation formats (TETRA, EDGE and CDMA2000FW). Then, we show how its recognized nonlinear impairments can be overcome using appropriate digital amplitude and phase pre-distortion algorithms.

This paper presents a new power amplifier (PA) linearizer model whose topology has support on the physical recursive behavior of the general PA electronic circuit. The resulting model presents a topology that, in its essence, is non recursive, which makes it suited for an easy implementation on digital hardware devices (such as FPGAs and DSPs). Validation tests were performed on a 900 MHz PA excited with WCDMA-3GPP signals and comparison with other linearizer models demonstrated its excellence. Due to its support on the PA physical behavior, the proposed predistorter model has the potential to be an adequate linearizer for a wide range of PA device technologies and excitation signals.

This paper investigates the minimum supply modulator linear processing bandwidth and slew rate (SR) requirements necessary to meet the standard specs of total signal distortion and adjacent signal power ratio in modern polar wireless transmitters. System level simulations, performed under CDMA2000FW and CDMA2000RV excitations, showed that, depending on the filter behavior in the amplitude envelope spectrum main lobe, relaxed bandwidths between 2 and 5 times the original signal are sufficient to guarantee desired specs of 45dB of ACPR and −35dB of NMSE. On what the SR is concerned, we concluded that 1/(2.5) times the maximum AM(t) signal slope is enough to achieve the same desired specs.

In this paper, the main nonidealities appearing in polar transmitters will be addressed, together with several implementation considerations. Special attention will be paid to the role of AM modulation nonlinearity and parasitic AM-to-PM conversion, once architecture mechanisms such as time-delay mismatch between branches or limited bandwidth in the amplitude path are controlled. The device limiting factors for a highly efficient switched mode operation and a linear amplitude modulation will be identified. Some circuit design and implementation guidelines for the RF modulating stage and the envelope amplifier will be discussed, to finish with system-level analysis considerations under two-tone and real communication signal excitations.

This paper presents the linearization of a polar transmitter for EDGE applications. The methodology used, complements the modulation of drain bias, vDD(t) with a proper variation of the input drive level, vg(t). It is believed to improve the linearity, while maintaining the efficiency of such transmitters.

In this paper, the particular impact of switching device ON resistance variation with drain supply voltage, Ron(VDD), on polar transmitter distortion is considered. Pulsed I/V measurement results over a GaN HEMT are used to predict the deviation in the Vdd-to-AM modulation profile. System-level calculations, in the presence of other nonidealities, allow the evaluation of the relative contribution of this frequency dispersion related effect.

The aim of this paper is to promote the discussion and raise peoplepsilas awareness towards the creation of recommendations and regulations to improve the installation of different technologies in the same site. This research was accomplished with a theoretical approach and verified with real measurements of commercial Tower Mounted Amplifiers (TMApsilas).

This paper reports on a new design method to overcome the nonlinear distortion associated to polar transmitters. The proposed methodology complements the modulation of drain bias, v DD (t), of conventional envelope elimination and restoration power amplifiers with a proper variation of the input drive level, v g (t). Allowing v g (t) to vanish when v DD (t) goes to zero, it obviates the inherent nonlinear distortion caused by the amplifier carrier feed-through. In addition, and beyond guaranteeing linear operation, such a design methodology is also beneficial for power added efficiency as it prevents the dissipation of instantaneous input power for a desired null instantaneous output power.

This paper presents a procedure to design a non-recursive model, suited for simulation and able to model short-and long-term memory effects, that is the best nonlinear approximator, up to a certain polynomial order, of a feedback behavioral model whose topology was derived from the PA physical behavior. A simple and systematic model extraction procedure, based on single-and two-tone tests, is given for a model truncated to 3rd order, and its validation is demonstrated through ADS simulation comparison.

Gallium nitride high electron mobility transistors (GaN HEMTs) have been accused of suffering from soft-compression, a recognized form of nonlinearity. Recently published works showed that this phenomenon – explained with the devices' charge-trapping effects – is only observed under CW operation, which has little in common with generally used communication signals. In real operation, those effects (modeled as a self-biasing phenomenon visible as a threshold voltage variation) are also observable, but in an apparently different way. Several nonlinear models already take into account this variation and, more recently, the authors presented an extraction methodology for such complicated systems. Based on that information, this paper presents the trapping behavior of GaN HEMTs and a physically based way of adjusting the quiescent point of a desired class B GaN HEMT-based PA, designed to operate under dynamically varying signals. Copyright © 2015 John Wiley & Sons, Ltd.

This paper explains the nonlinear dynamic memory effects observed on both AM-AM and AM-PM conversions of nonlinear microwave power amplifier, PA, circuits. The study is introduced via an analytic Volterra series analysis which is then illustrated with envelope simulations gathered from a GaN HEMT based microwave PA. This way, a theoretically supported explanation of the small- and large-signal dynamic asymptotic behavior, observed in the amplitude and phase conversions of microwave/RF PA's, could be achieved.

It is well known that AM/AM distortion introduced by power amplifiers, PAs, is responsible for the degradation of the transmitted signals. Understanding the origins and behavior of that deterioration is thus of paramount importance for PA performance improvement. This paper is devoted to analyze the impact of bias networks on the dynamic AM/AM contours of microwave PAs and, using that knowledge, to give an interpretation to the often observed hysteretic paths' shape.

This paper presents a new method for phase characterization of nonlinear mixing products based on statistical averages of signal windows captured by a synchronous two-channel sampler. Passing some of the signal processing operations, previously performed by complex laboratorial arrangements, to the digital domain, the present method simplifies the measurement bench and allows absolute phase measurements, i.e., not depending on any arbitrarily selected phase reference nonlinearity.

This paper presents a new nonlinear device model, for microwave power GaN HEMTs, amenable for integration into a commercial harmonic balance simulator. All the steps taken to extract it are explained, starting with the extrinsic elements' determination and ending with the intrinsic ones. This model was validated by comparing measured and simulated output power and intermodulation distortion data of a GaN HEMT. Very good agreement was obtained from small- to large-signal excitation regimes.