Table of Contents
Fetching ...

Comprehensive Signal Modeling for Talkative Power Conversion

Jan Mietzner, Cerikh Chakraborty, Peter A. Hoeher, Lutz Lampe

TL;DR

This paper addresses the lack of rigorous signal models for joint information and power transfer in talkative power conversion (TPC) by deriving both continuous-time and discrete-time models for the information-carrying output voltage of a buck converter with an LC filter and Ohmic load. It develops a comprehensive continuous-time model that decomposes the output into a transient and a data-driven ripple, establishes multiple discrete-time approximations with accuracy analyses, and introduces a generic end-to-end signal model applicable to arbitrary modulation signals, enabling end-to-end equalization design. The work further analyzes implications for receiver design, including transient subtraction and LC-filter-aware equalization, and generalizes the modeling framework to include parasitic effects and arbitrary impedance loads, thereby bridging power electronics and communications for future TPC systems. The results provide closed-form time-domain expressions, Fourier-domain ripple analyses, and a principled pathway to ISI mitigation and system optimization in practical TPC implementations across power grids, VLC, and related domains.

Abstract

Talkative power conversion is a switching ripple communication technique that integrates data modulation into a switched-mode power electronics converter, enabling simultaneous information transmission and power conversion. Despite numerous research papers published over the last decade on various theoretical and practical aspects of this emerging topic, thorough signal modeling suitable for analysis and computer simulations is still lacking. In this article, we derive the continuous-time output voltage of a DC/DC switched-mode power electronics converter for a broad range of pulsed-based modulation schemes. We also develop corresponding discrete-time signal models and assess their accuracies. Finally, we devise a generic end-to-end signal model for arbitrary modulation signals, discuss implications of continuous-time and discrete-time signal modeling on equalization, and consider generalizations to include parasitic effects as well as the influence of general impedance loads.

Comprehensive Signal Modeling for Talkative Power Conversion

TL;DR

This paper addresses the lack of rigorous signal models for joint information and power transfer in talkative power conversion (TPC) by deriving both continuous-time and discrete-time models for the information-carrying output voltage of a buck converter with an LC filter and Ohmic load. It develops a comprehensive continuous-time model that decomposes the output into a transient and a data-driven ripple, establishes multiple discrete-time approximations with accuracy analyses, and introduces a generic end-to-end signal model applicable to arbitrary modulation signals, enabling end-to-end equalization design. The work further analyzes implications for receiver design, including transient subtraction and LC-filter-aware equalization, and generalizes the modeling framework to include parasitic effects and arbitrary impedance loads, thereby bridging power electronics and communications for future TPC systems. The results provide closed-form time-domain expressions, Fourier-domain ripple analyses, and a principled pathway to ISI mitigation and system optimization in practical TPC implementations across power grids, VLC, and related domains.

Abstract

Talkative power conversion is a switching ripple communication technique that integrates data modulation into a switched-mode power electronics converter, enabling simultaneous information transmission and power conversion. Despite numerous research papers published over the last decade on various theoretical and practical aspects of this emerging topic, thorough signal modeling suitable for analysis and computer simulations is still lacking. In this article, we derive the continuous-time output voltage of a DC/DC switched-mode power electronics converter for a broad range of pulsed-based modulation schemes. We also develop corresponding discrete-time signal models and assess their accuracies. Finally, we devise a generic end-to-end signal model for arbitrary modulation signals, discuss implications of continuous-time and discrete-time signal modeling on equalization, and consider generalizations to include parasitic effects as well as the influence of general impedance loads.

Paper Structure

This paper contains 21 sections, 65 equations, 10 figures, 1 table.

Figures (10)

  • Figure 1: Circuit diagram of the considered synchronous DC-DC step-down converter with coupled switches, $LC$ lowpass filter, and Ohmic load (cf. Hoeher2021).
  • Figure 2: Illustration of the basic pulse shape $g_{\sf tx}(t)$ for the unmodulated case ($t_{\sf c} = T_{\sf p}/2$, $T_{\sf p}=\delta \cdot T$), given the parameter values in Table \ref{['tab:num-para']}.
  • Figure 3: Illustration of the transient component (dashed line, blue color) and the data-dependent components (solid lines, mixed colors) of the output voltage $v_2(t)$ for the parameter values in Table \ref{['tab:num-para']} ($L\!=\!10$$\mu$H, $C\!=\!1$$\mu$F, $R_{\sf L}\!=\!10$$\Omega$), an arbitrary fixed input voltage $V_1$, initial conditions $v_2(0)/V_1=\delta=0.75$, $\dot{v}_2(0)=0$ V/s, and VPWM for an example data sequence $x[k]=[1,0,1,0,1]$. The duty cycle variation depth was chosen as $\pm 0.2$ for illustrative purposes.
  • Figure 4: Illustration of the overall output voltage $v_2(t)$ according to (\ref{['eq:v2_final']}) (solid line, blue color) compared against the result of a corresponding LTspice® simulation (dotted line, red color) for the parameter values in Table \ref{['tab:num-para']} ($L\!=\!100$$\mu$H, $C\!=\!0.1$$\mu$F, $R_{\sf L}\!=\!20$$\Omega$), an arbitrary fixed input voltage $V_1$, initial conditions $v_2(0)/V_1=\delta=0.75$, $\dot{v}_2(0)=0$ V/s, and VPWM for an example data sequence $x[k]=[1,0,1,0,1,...]$ of length 64. The duty cycle variation depth was chosen as $\pm 0.025$.
  • Figure 5: Normalized ripple power spectrum $20\log_{10}(|V_{2,{\sf ripple}}(f)|)$ for the parameter values in Table \ref{['tab:num-para']} ($L\!=\!100$$\mu$H, $C\!=\!0.1$$\mu$F, $R_{\sf L}\!=\!20$$\Omega$) and VPWM for an example data sequence $x[k]=[1,0,1,0,1,...]$ of length 64 (blue color). The duty cycle was chosen as $\delta=0.75$ and the variation depth as $\pm 0.025$. The cut-off frequency $f_{\sf 3dB}$ of the second-order $LC$ lowpass filter is indicated by the vertical red line. The theoretical frequency response of the second-order $LC$ lowpass filter with Ohmic load $R_{\sf L}$ is included as a reference (magenta color).
  • ...and 5 more figures