Dynamic operation of sensorless dead-time optimization in digitally controlled synchronous buck converters

As demands for high bandwidth high current switching converters for microprocessor and other electronic loads increase, realizing digital control loops around switching converters can give significant benefits as it allows easy programmability and fine-tuning converter parameters in real-time to achieve the best performance and efficiency at any given operating condition. In a digitally controlled synchronous buck converter design, the problem of setting the dead-times to optimal values is addressed in this thesis. Several previously proposed approaches to dead-time control are reviewed. The sensorless dead-time optimization approach to dead-time control is found to be attractive as it renders simple gate driver design and avoids sensing of any power stage signals, thus alleviating the noise issues and hardware associated with sensing. A need for dynamic operation of sensorless dead-time control is explained. A more detailed analysis of the effects of dead-times on the converter efficiency and duty cycle is provided and a new dynamic algorithm is proposed to enable real-time sensorless dead-time optimization even in the presence of persistent disturbances in the converter.;A digitally controlled Buck Converter prototype is developed and successful convergence to optimal dead-times even in the presence of persistent load transients is demonstrated.