Derivation of a behavioral RF power amplifier model with low normalized mean-square error

Abstract

In this paper a behavioral power amplifier model with substantially lower normalized mean-square error than the well-known and widely used parallel Hammerstein or tapped delay line model is presented. This is significant since the parallel Hammerstein model currently must be considered to be the baseline for behavioral power amplifier modeling efforts. The proposed model does also have at least as low total error as any other model known to the authors for measurements on class-AB LDMOS power amplifiers. Further, the proposed model exhibits an equal or lower out-of-band error than the parallel Hammerstein model. The parameters of the model can be extracted using straightforward system identification techniques. Together this makes the proposed model a strong candidate to replace the parallel Hammerstein model as the model of choice for behavioral power amplifier modeling for many applications.

A New Approach to Pruning Volterra Models for Power Amplifiers

“…Slightly more complex than the MP, and exhibiting an exponential dependence, is the VBW model (dotdashed and downward triangles). The model of [6] presents the same dependence, but with a significant increase in the number of parameters. The RPV model (continuous line and left-pointed triangles) has moderate complexity, and thanks to its linear behavior, the number of coefficients is maintained below that of the model of [6] for Q ≥ 4.…”

“…Other discrete-time complex envelope models based on a priori knowledge of the PA circuit model have been recently presented [5], [6], [7]. In [5], the authors propose a model for wideband amplifiers deduced under the assumption of transfer functions with a flat response in all harmonic zones except at dc where the (even) nonlinear transfer functions cannot be considered flat, because of low-frequency dispersive effects and load impedance dependence on baseband frequency.…”

“…This Volterra behavioral model for wideband amplifiers (VBW), presents a distinct "off-diagonal" structure. In [6], a behavioral power amplifier model is also derived starting from previous knowledge about the amplifier, and the authors claim an NMSE substantially lower than for the well-known and widely used MP model. The two models, VBW and [6], present a similarity in the fact that filtering involves samples of the instantaneous power.…”

“…In [6], a behavioral power amplifier model is also derived starting from previous knowledge about the amplifier, and the authors claim an NMSE substantially lower than for the well-known and widely used MP model. The two models, VBW and [6], present a similarity in the fact that filtering involves samples of the instantaneous power. Although in the case of the VBW no filtering is applied to the complex-envelope samples, the fidelity of this model is remarkably better than the MP.…”

“…An extended version of the VBW model and its efficiency are analyzed in [9], where an equivalent discrete-time baseband PA model considering memory effects and bandpass nonlinearity is formulated by joining the MP and VBW approaches. The accuracy of the so-called extended VBW (EVBW) model is reflected in significantly lower NMSE than for the MP model and similar to [6], which contains a larger number of coefficients.…”