Research On Behavioral Modeling For Radio Frequency Active Components Based On Nonlinear X-parameters

With the rapid development and great progresses of the semiconductor fabrications in the past few decades, the characteristics of the Radio Frequency/RF microwave components have been completely changed. Taking a two-port active component as an example, the power amplifiers are intending to operate in the nonlinear region to achieve a higher efficiency, output power, and linearity. In the meantime, the amplifiers under large-signal stimulus will show a series typical nonlinear distortions, such as harmonic distortion, intermodulation distortion, gain compression and memory effects. Under such circumstances, the classical linear system theory（such as small signal S-parameters） is no longer suitable for the accurate modeling and prediction for the active components. As the industrial standard solution, the nonlinear X-parameters are most suitable for the modeling of multi-frequency situation. As for the extraction platform, NVNA（Nonlinear Vector Network Analyzer） is capable of measuring the absolute magnitude and phase information of the cross frequency spectrums, and it has also becoming a hot research topic in microwave measurement fields.As the most important reference measurement channel of NVNA, the HPR（Harmonic Phase Reference） affects the measurement accuracy of the absolute phase spectrum value. However, on consideration of the disadvantages of the commercial products, an improved prototype HPR is presented in this dissertation, with a wider bandwidth and a lower complexity. Moreover, a series of innovative work about the X-parameters model’s extension towards a higher dimension, accuracy and complexity operating conditions have also been achieved together. It extends the X-parameters from an idea condition to a more general application. The main research contents and innovative research results are shown below:Firstly, in view of the current application background, this doctoral dissertation provides a completed design, simulation, fabrication and test procedure for the 67 GHz HPR（or called comb generator） based on the coplanar waveguide nonlinear transmission line structure. Besides, in order to solve the quantity traceability problem, an innovative calibration procedure for phase measurement based on the electro sampling system and the equivalent sampling oscilloscope was proposed as well. The calibration results satisfied the expected objectives.Secondly, this dissertation analyzes the basic assumptions of the conventional X-parameter model, and proposes a more universal and general behavioral modeling theory of the load-dependent X-parameter. It extends the application background from the nearly matched situation to the large mismatch cases. Furthermore, referring to the application of the S-parameters in a two-port microwave network, this dissertation also proposed the analytical expressions for the input reflection coefficient by utilizing the load-dependent X-parameters. It breaks the research blank of nonlinear amplifier design procedure, and provides a theory foundation for the analytically descriptions of the nonlinear performance, such as power gain, TOI（Third Order Intercept） and ACPR（Adjacent Channel Power Ratio）.Finally, the harmonic dynamic X-parameter model, which aims at providing a powerful analysis capability for the long-term memory effects of the power amplifier, was proposed. Besides that, a novel validation experiment based on the “narrow RF binary pulse envelope” signal was presented as well. During the extractions of ZFL₁1AD+, the fundamental and harmonic large signal responses were measured, as well as the memory kernels at different frequency. Comparing to the traditional two-tome method, the fundamental kernel keeps the same, but the harmonics’ memory kernels are extracted for the first time in the world by using this dynamic X-parameter model. This research results will undoubtedly promote the further developments of the nonlinear memory effects modeling theories.