A Hardware-Enabled Tool for Nonlinear Analysis of Digitally Controlled High-Freq. DC-DC Converters

High frequency digitally controlled DC-DC converters have become popular because of offering high performance and efficiency with reduced size and thermal overheads. The output voltage of a high frequency DC-DC converter is sampled once per switching cycle followed by using a quantizer. The effects due to finite discretization may lead to various fast-scale instabilities. This primarily leads to non-smooth bifurcation, which may not occur using analog controllers for the same operating condition. These significantly increase the root-mean-square (RMS) current, thereby degrading the efficiency and violating the ripple limits. In this paper, a hardware-enabled methodology is developed to capture noise-insensitive experimental bifurcation diagrams in a high frequency DC-DC converter. This would enable one to use system, controller, and timing parameters as different bifurcation parameters. A digitally current-mode controlled buck converter is tested with 500 kHz switching frequency. Bifurcation diagrams using the proposed tool are shown to be much more informative than using a 1 GHz oscilloscope. Further, analytical methods are provided for accurate measurement of the RMS current and power spectral density under period-1 and higher periodic orbits. The proposed tool is extended to analog controllers.