Enhanced Stability with Fast Transient Performance in Digitally Current Mode Controlled Multi-phase Buck Converters using Event-based Sampling

Digital current mode control (CMC) architectures have been gaining popularity in multiphase DC-DC converters for high-current low-voltage applications, in which a configurable digital platform enables one to achieve (i) effective phase current balancing and ripple minimization with reduced output capacitance and better thermal distribution, (ii) high efficiency by dynamic phase shading, and (iii) fast transient performance by online controller tuning. In fully digital CMC (DCMC), the phase currents are sampled once per switching cycle, and the current ripples are emulated inside the digital platform, thereby reducing ADC sampling requirement. In DCMC, the current samples can be made insensitive to switching noise, and the emulated ripple can be made insensitive to variations in phase inductances. However, one cycle delay significantly degrades cycle-by-cycle stability boundary with limited closed-loop bandwidth. This paper proposes an event-based sampling method in fixed-frequency DCMC in a multi-phase buck converter. Stability boundaries are formulated, discrete-time large-signal model is derived for stability analysis and controller design. Using mathematical analysis and simulation study, the proposed sampling is shown to significantly enhance stability boundary and transient performance over uniform sampling.