Memory Effects In Power Amplifiers Research Articles

A Novel Digital Predistortion Identification Algorithm Based on Variable Forgetting Factor Recursive Least Square Method

The transmitting signal of wireless communication system is impaired by the nonlinearity of RF power amplifier (PA) which leads to signal distortion and spectrum spillover, by which the signal transmission quality is affected. Digital predistortion (DPD) is an efficient and economical way to correct the nonlinear effects of power amplifiers. The recursive least square (RLS) recognition algorithm is commonly used to extract the correction coefficients of the DPD model, and the accuracy of the extraction directly affects the system performance. In this paper, a new variable forgetting factor identification algorithm (new variable forgetting factor recursive least square, NVFFRLS) is proposed for recursive least square (RLS) identification algorithm. The 64-QAM signal is combined with a memory polynomial (MP) predistortion model for predistortion system simulation. The experimental results show that, compared with the RLS identification algorithm and two kinds of variable forgetting factor RLS identification algorithms, the algorithm has smaller estimation error, faster convergence, and better tracking capability, stability, and adaptability; the predistortion system based on NVFFRLS identification algorithm can compensate the nonlinear memory effects of power amplifier more effectively.

A Multiple-Time-Scale Analog Circuit for the Compensation of Long-Term Memory Effects in GaN HEMT-Based Power Amplifiers

In this article, we present an analog circuit for the compensation of long-term memory effects in power amplifiers (PAs) based on GaN high-electron-mobility transistors (HEMTs). The analog compensation circuit (ACC) is supported on a charge-trapping behavioral model that features a state-dependent variable emission time constant, rather than a fixed emission time constant. This allows not only for the more accurate modeling and compensation of the nonlinear long-term memory effects experienced by GaN HEMT-based PAs, but also for the adaptation of the dynamics of the ACC to varying operating conditions, such as the ambient temperature. We demonstrate that the proposed ACC is able to successfully eliminate any multiple-time-constant transient symptom of current collapse following a high-power radio-frequency pulse or two-tone excitation and accurately preserve the class of operation of a GaN HEMT-based PA. Moreover, we also demonstrate that the proposed ACC can aid the linearization of GaN HEMT-based PAs for fifth-generation (5G) New Radio (NR) communications, allowing for the reduction in the complexity of the digital predistorter. For a 5G NR time-division duplexing signal with a bandwidth of 20 MHz, the inclusion of the ACC improved out-of-band emissions by over 15 dB and improved the error-vector magnitude by over 10 p.p. at essentially no cost to the PA’s power-added efficiency.

Instant Gated Recurrent Neural Network Behavioral Model for Digital Predistortion of RF Power Amplifiers

This article presents two novel neural network models based on recurrent neural network (RNN) for radio frequency power amplifiers (RF PAs): instant gated recurrent neural network (IGRNN) model and instant gated implict recurrent neural network (IGIRNN) model. In IGRNN model, two state control units are introduced to ensure the linear transmission of hidden state and solve the problem of vanishing gradients of RNN model. In contrast with conventional RNN model, IGRNN can better describe the long-term memory effect of power amplifier, more in line with the physical distortion characteristics of power amplifier. Furthermore the instantaneous gates are used to express the input information implicitly to reduce the redundancy of the input information, and a simpler IGIRNN model is proposed. The complexity analysis indicates that the proposed models have significantly lower complexity than other RNN-based variant structures. A wideband Doherty RF PA excited by 100MHz and 120MHz OFDM signals was employed to evaluate the performance. Extensive experimental results reveal that the proposed IGRNN and IGIRNN models can achieve better linearization performance compared with RNN model and traditional GMP model, and have comparable performance with lower computational complexity compared with the state-of-the-art RNN-based variant models, such as gated recurrent unit (GRU) model.

Augmented Time-Delay Twin Support Vector Regression-Based Behavioral Modeling for Digital Predistortion of RF Power Amplifier

In this paper, a dynamic behavioral model for digital predistortion (DPD) of RF power amplifier (PA) based on amplitude and phase augmented time-delay twin support vector regression (AP-TSVR) is proposed. Unlike other SVR-based methods, the TSVR model finds a pair of non-parallel planes by solving two related support vector machine (SVM) type problems, namely, the $\varepsilon $ -insensitive up- and down-bound functions. Furthermore, in order to accelerate the training process, an effective linear regression algorithm was used to solve the paired quadratic programming problems (QPPs) of the TSVR model involved. The simulation results show that the proposed model is able to give improved modeling and distortion mitigation capability than the traditional memory polynomial-based model, and reduce CPU training time than the ordinary SVR model, even when the effects of both nonlinear characteristics and memory effects of PA are considered. To verify the effectiveness of the proposed method, experimental verification was performed using single-device gallium nitride (GaN) PA and GaN Doherty PA, respectively. The experimental results show that the new modeling approach can provide very efficient and extremely accurate linearization performance with improving generalization ability.

Impacts of Crest Factor Reduction and Digital Predistortion on Linearity and Power Efficiency of Power Amplifiers

The linearity and power efficiency are two main factors of power amplifiers (PA), which can hardly be improved together. Digital predistortion (DPD) compensates for nonlinearity and memory effects of PAs in high efficiency zone. However, for advanced communication standards the high peak-to-average power ratio of the modulated signal constrains the PA operating point and thus limits the power efficiency. Crest factor reduction (CFR) techniques are applied to address this problem though they degrade the PA linearity relatively. In this brief, we compare different CFR approaches with respect to their performances in terms of linearity, PA power efficiency, and computational complexity. Classical hard clipping, clip-and-filter as well as joint CFR/DPD methods are considered. To complete the comparison, we propose a modeled CFR identified using indirect learning architecture combined with DPD. The simulation results validate its effectiveness on the trade-off among linearity, PA power efficiency and CFR computational complexity.

Optimal Sizing of Two-Stage Cascaded Sparse Memory Polynomial Model for High Power Amplifiers Linearization

The nonlinearities and memory effects of power amplifiers (PAs) can be compensated by multistage cascaded digital predistortion with low complexity. Compared with full multistage models, sparse multistage models may have the same linearization performances while their complexities are even lower. However, the choice of the model structure is very difficult, especially for the sparse models. For instance, if there are (K+1)(L+1) full models, then 2 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">K+L-2</sup> different corresponding sparse models exist. An algorithm with the optimal search space definition is proposed in this paper to search for the optimal cascaded sparse model structure. The search criterion represents tradeoff between the modeling accuracy and identification complexity with a weight coefficient. A method to determine the value of the weight coefficient is proposed in this paper. The sparse model solution found by the proposed algorithm is evaluated with a three-way Doherty PA using a long term evolution-advanced signal. It is also compared with the solution of full model structure search.

A Novel Algorithm for Determining the Structure of Digital Predistortion Models

The generalized memory polynomial (GMP) model is one of the most commonly used models in digital predistortion to compensate for the nonlinearities and memory effects of power amplifiers (PAs). A very difficult, yet crucial, aspect of behavioral modeling is model dimensioning, i.e., determining nonlinear orders and memory depths. In this paper, we propose an algorithm based on hill-climbing (HC) heuristic to search for a GMP model structure, which provides the best tradeoff between modeling accuracy and complexity with different search criteria. Algorithm performance and convergence are evaluated for two different search criteria. Some optimization methods of the algorithms are then proposed to reduce the execution time by controlling its search path. The algorithm has been evaluated using real measurements from two different Doherty PAs. The solutions of the algorithm are proved robust in linearization of PA on the test bench. The results show that the proposed algorithm allows us to decrease the searching time by a factor typically greater than $10^4$ compared with exhaustive search.

Advanced design and characterization methodologies for memory-aware CMOS power-amplifier implementation

Abstract. This paper reports on an effective root-cause analysis method of memory effects in power amplifiers, as well as introduces compensation techniques on a circuit design level. Despite conventional memory-effect approaches, the discussed method uses a two-tone scan over a wide operation and modulation range. This enables an in-depth study of physical causes and helps to implement compensation techniques at design stage. On the one hand, this circuit investigation is optimized using an automated SystemC model parametrized with real device and measurement values. Hence, computation time is widely reduced which shortens design cycles. On the other hand, the implementation of the derived circuit compensation means will reduce the complexity of digital pre-distortion due to a reduced memory-effect induced AM/AM and AM/PM hysteresis. The approach is demonstrated on a 65 nm CMOS power amplifier with an OIP1 of 27 dBm and a PAE of over 30 % using WCDMA and LTE signals. In fact, mismatch could be reduced by more than 8 %.

Cartesian augmented Hammerstein model for non‐linearity and I/Q impairments compensation in concurrent dual‐band transmitters

In this study, a novel model for the joint compensation of dual-band power amplifier (PA) distortion and in-phase/quadrature (I/Q) imbalance, which are the dominant impairments of wireless signal transmitters, in a complexity reduced structure is proposed. This model is mainly based on the augmented Hammerstein approach in a Cartesian form so that a static block is intended to model the PA non-linearity and the modulator I/Q imbalance while a finite impulse response filter based block is used to model the PA memory effects. The proposed approach reduces significantly the complexity of the proposed model compared with the recently published memory polynomial models. Experiments with and without I/Q modulator impairments have been carried out on a dual-band PA to verify the accuracy and the performance of the linearisation technique based on the proposed model. The experimental results have revealed a good impairments reduction with much less model coefficients compared with the recently published ones.

Generalised two‐box cascaded Hammerstein‐like digital predistorter for wide‐band RF power amplifiers

A generalised two-box cascaded Hammerstein-like (GTBC-H) digital predistorter is proposed for linearising wide-band RF power amplifiers (PAs). The GTBC-H predistorter is composed of a static nonlinearity block for countervailing the strong static nonlinearities and a dynamic nonlinearity block for compensating the memory effects of RF PAs. The proposed predistorter adds extended cross-terms of the leading terms and some lagging envelope terms in the dynamic nonlinearity block to compensate the memory effects of wide-band RF PAs more effectively. A 460 MHz and a 1.94 GHz Doherty RF PA are utilised to validate the performance of the proposed predistorter when a three-carrier wide-band CDMA and a single-carrier long-term evaluation signal are applied separately. The validation results illustrate that the proposed GTBC-H predistorter can further suppress the residual out-of-band emission over the augmented Hammerstein, the enhanced Hammerstein and the parallel-LUT-MP-EMP predistorter.

Quantifying the Memory Effects of Power Amplifiers: EVM Closed-Form Derivations of Multicarrier Signals

This letter deals with the analysis of the memory effects of the power amplifier (PA) on multicarrier signals. We analytically derive the expressions of the error vector magnitude (EVM) of amplified signals jointly considering a memory polynomial PA model and clipping effects. We provide analytical EVM expressions that depend on the PA coefficients, PA order, and memory depth as well as the clipping threshold of the input signal. These analytical expressions are new general formulas which allow to predict in-band distortions at the PA output.

An Accurate Digital Predistortion for Wideband Power Amplifiers Based on Directly Solving the Inverse Model Function

Digital predistortion (DPD) is one of the most effective techniques that can compensate for the distortions caused by the nonlinearities and memory effects of power amplifiers (PAs). In this paper,...

FPGA-based test bed for measurement of AM/AM and AM/PM distortion and modeling memory effects in RF PAs

Using a field-programmable gate array (FPGA) development board, a digital signal processor (DSP) builder, and the phase-to-amplitude conversion principle, a low-cost system for measuring the amplitude-to-amplitude (AM/AM) and amplitude-to-phase (AM/PM) distortion curves of radio frequency (RF) power amplifiers (PAs) is presented. The state of the art based on the measurements and preliminary studies of AM/AM and AM/PM distortion curves is discussed. A full digital control of the test bed simulated/emulated in Matlab/Simulink is introduced to recalculate the known AM/AM and AM/PM measurements stored as look-up table (LUT). Finally, the low-cost system comprises the memory polynomial model (MPM) that involves the nonlinearity order and memory effects of real PAs.

Research on Modeling Methods of Power Amplifier with Memory Effect in Broadband Communication

Memory effect of power amplifier (PA) due to broadband signals which adopt new modulation technique is becoming obvious, and worsening the linearity of PA. Behavioral modeling for RF (Radio Frequency) power amplifier is indispensable to design digital pre-distortion system by which we can promote the efficiency of PA. According to the simulation data in the software of ADS (Advanced Design System), behavioral models of MRF21085 amplifier can be established by MATLAB. Comparison of memory polynomial model and ADS simulation of MRF21085 amplifier demonstrate memory polynomial model can represent the electric characteristic of it exactly.

A New Metric for the Envelope of OFDM Signals Based on Memory Power Amplifier

Orthogonal frequency division multiplexing (OFDM) signals with large envelope variations usually requires power de-rating operation on the power amplifier (PA) to meet the requirement of adjacent channel leakage ratio of systems. Peak-to-average power ratio (PAPR) and cubic metric (CM) are two metrics commonly used to characterize the envelope of OFDM signals. And the amount of power de-rating is determined by their values. However, both PAPR and CM do not take the memory effects (MEs) of PA into account. Actually, for wideband system such as OFDM, the inherent MEs in PA can not be ignored any more. In this paper, we study the behavioral model of PA with memory and sufficiently analyze the impacts of memory PA on OFDM signals. Based on our analysis, a new metric for the envelope characteristics of OFDM signals is proposed. The simulation results show that the proposed metric provides much more accurate prediction on required power de-rating of PA than both PAPR and CM. So the new metric is a better choice for the envelope characteristics of OFDM signals.

An Improved Modeling of Nonlinear Circuits based on Elman Neural Network

In this paper, an improved modeling based on Elman neural network was proposed to analyze the nonlinear features of nonlinear circuits with the memory effect. The input vector of the hidden layer in neural network is normalized to enhance the neural network convergence precision. A group of Chebyshev orthogonal basis functions was employed to activate hidden layer neurons. Computer simulation results of the nonlinear power amplifier (PA) have shown that the proposed behavioral modeling not only accurately describes the nonlinear distortions of PAs, but also well depicts memory effect of PAs. And, the proposed approach could be also applied to analyze linear circuits and RF power amplifiers.

Method of Determining Delay Dependence of the Memory Effect in Power Amplifiers and Derivation of Inverse to Cancel the Nonlinear Distortions

This paper describes the details of the iteration process used to determine the transfer functions of linear time-invariant (LTI) circuits causing the memory effect of power amplifier (PA). An outline of the method is reported in our work presented at ICCS2012. The accuracy of the method is improved by using cross-correlation spectra at three signal levels, and its validity is confirmed by a computer simulation. The method can be applied to online updating of PAs operating in mobile communication systems. The updating is realized separately from the fast varying nonlinear coefficients. The possibility of updating with a short interval is indirectly shown for the nonlinear coefficients using a procedure similar to that of memoryless PAs. For PAs characterized by the method, this paper also describes the inverses that cancel the nonlinear distortion with minimum complexity. The validity of the inverse is confirmed by a computer simulation on the power spectrum of the PA for orthogonal frequency-division multiplexing (OFDM) signals with 500 subcarriers. The simulated spectra show that the fifth order or higher inverses are effective in keeping adjacent channel leakage power ratio (ACLR) lower than -60dB at the practical signal level. Improvements in the error vector magnitude (EVM) due to the inverse were also confirmed by reductions of gain and phase variations under varying envelope conditions.

Formulations for the Estimation of IMD Levels in an Envelope Feedback RFIC Amplifier: An Extension to Dynamic AM and PM Behavior

This paper extends the formulations presented in an earlier paper on intermodulation distortion (IMD), to include the important aspect of memory effects in power amplifiers (PAs). They are intended for the estimation of linearity requirements for circuit blocks typically found in the error signal paths of envelope feedback amplifiers in the context of computer-aided design and test of Radio Frequency Integrated Circuit (RFIC) PAs. A system of nonlinear equations for computing IMD levels is derived and an algorithm is given in the Appendix. They facilitate implementation within a computer-aided IMD test setup based on the methodology presented in the earlier paper.

A Novel PA Behavioral Modeling Using Memory Rational Function

Aiming at the nonlinear characteristics and memory effects of power amplifier, a memory rational function behavioral model, which based on the memory polynomial model and memoryless rational function model, is proposed. The model coefficients are identified by conjugate gradient method. The estimation of the best model nonlinearity order and the best memory depth is carried out by minimized the normalized mean squared error (MNSE). The model is validated using multi-carrier WCDMA signal and a MRF6S21140H power amplifier by measurement. The result indicates that the proposed model can achieve better model performance with less number of coefficients compared to the memory polynomial model and memoryless rational function model

Pruned Volterra Models with Memory Effects for Nonlinear Power Amplifiers

In this letter, a novel model is proposed for modeling the nonlinearity and memory effects of power amplifiers. The classical Volterra model is modified through a function of the sum of nonlinearity order with sum of memory length. The parameters of this model can be extracted in digital domain since the model is analyzed based on the envelope signals. The model we proposed enables a substantial reduction in the number of coefficients involved, and with excellent accuracy.