This thesis describes the development of a system to continuously monitor the battery in a car and warn the user of an upcoming battery failure. An automotive battery endures enormous strain when it starts the engine, and when it supplies loads without the engine running. Note that the current during a cranking event often exceeds 500 Amperes. Despite the strains, a car battery still typically lasts 4-6 years before requiring replacement. There is often no warning of when a battery should be replaced and there is never a good time for a battery failure.All currently available lead-acid battery monitoring systems use voltage and current sensing to monitor battery impedance and estimate battery health. However, such a system is costly due to the current sensor and typically requires an expert to operate the system. This thesis describes a prototype system to monitor battery state of health and provide advance warning of an upcoming battery failure using only voltage sensing. The prototype measures the voltage during a cranking event and determines if the battery is healthy or not. The voltage of an unhealthy battery will drop lower than a healthy one, and it will not recover as quickly.The major contributions of the proposed research to the field are an algorithm to predict automotive battery state-of-health that is temperature-dependent and a prototype implementation of the algorithm on an ARM processor development board.iii ACKNOWLEDGEMENTS I would like to acknowledge and thank a number of people and organizations that have helped make this thesis a reality.Dr. Dong Ha for providing me with the advice and support to see this project through.

No abstract

No abstract

No abstract

No abstract

Power consumption of smart cameras varies significantly between sleep mode and active mode, and a smart camera operates in sleep mode for 80 -90% of time for typical use. To prolong the battery life of smart cameras, it is essential to increase the power converter efficiency for light load, while being able to manage heavy load. The power stage of traditional buck converter is optimized for maximum load, at the cost of light-load efficiency. Wei proposed a multiphase buck converter incorporating the baby-buck concept and optimum number of phases (ONP) control. This thesis research investigated Wei's multiphase buck converter to improve the light-load efficiency for smart cameras as the target application.The proposed two-phase buck converter aims to provide power for microprocessors of smart cameras. The input voltage of the converter is 5 V DC, and the output voltage is 1.2 V DC with power dissipation range of 25 mA (30 mW) for light load and 833 mA (1 W) for heavy load. Three methods are considered to improve light-load efficiency: adopting baby-buck concept, adapting ONP control for low-power range, and implementing a pulse frequency modulation (PFM) control scheme with discontinuous conduction mode (DCM) to lower switching frequency. The first method is to adopt the baby-buck concept through power stage design of each phase to optimize efficiency for a specific load range. The baby-buck phase is optimized for light load and the heavyload phase is designed to handle the processors maximum power consumption. The second method performs phase selection from sensed load current information. Rather than have all phases active for heavy-load as in ONP control, optimum phase selection (OPS) control is introduced to adaptively select between phases based on load current. Due to low-power constraints, OPS is more efficient for the medium to heavy-load range. The transition between phases due to load iii change is also investigated. The third and final method implements PFM control with DCM to lower switching frequency and reduce switching and driving losses under light load. PFM is accomplished with a constant on-time (COT) valley current mode controller, which uses the inductor current information and output voltage to generate switching signals for both the top and bottom switches. The baby-buck phase enters DCM to lower switching frequency under very light load, while the heavy-load phase remains in continuous conduction mode (CCM) throughout its load range.The proposed two-phase buck converter is designed and prototyped using discrete components. Efficiency of the two-phase converter and a power loss breakdown for each block in the control scheme were measured. The efficiency ranges from 64% to 81% for light load ranging of 30 mW to 200 mW, and the efficiency ranges from 81% to 88% for heavy load ranging from 200 mW to 1 W. The majority loss is due to controllers, which are responsible for 37 % (8.6 mW) for light load of 60 mW and for 10.9 % (9 mW) for heavy load of 600 mW. The gate driver loss is consider...