A novel current-balanced adaptive zero current detection scheme for Buck DC-DC converter

This brief proposes a novel current-balanced adaptive zero-current detection (CB-AZCD) scheme for buck DC-DC converters. The traditional adaptive zero-current detection (ZCD) scheme adjusts the turn-off time of the low-side power MOSFET ML by detecting the switching node voltage VX value or slope at the turn-off time of ML and calibrating ML conduction time through digital feedback loops. The high-frequency ringing of VX results in detection errors and the accuracy of digital calibration circuits is limited to the least significant bit (LSB). Instead of detecting VX, the proposed CB-AZCD scheme builds feedback loops according to inductor volt-second balance equation to adaptively adjust the turn-off time of ML, which avoids the detection error caused by VX ringings. Compared with the conventional calibration scheme, the proposed CB-AZCD uses analog methods to realize adaptive calibration and improve accuracy. The proposed CB-AZCD is designed and verified in 180nm CMOS process. The simulated results show that the CB-AZCD can realize near-optimum ZCD under various processes, voltage, temperature (PVT), and mismatch conditions. Compared with the situation without the proposed CB-AZCD technique, the buck converter can effectively improve the efficiency by 3.7% in discontinuous conduction mode.