Projects per year
Class-D audio amplifiers are well known for their excellent power-efficiency. They were traditionally used for higher power applications, where cooling requirements at peak power determined the system costs. More recently, Class-D amplifiers have entered the mobile audio market, where energy efficiency is the main concern. In mobile devices, lower power demands and stringent space requirements have led to the use of fully-integrated and filterless Class-D amplifiers. This work covers three research questions concerning integrated, filterless Class-D amplifiers: First, established methods for design and characterization have to be evaluated for their suitability in low power, filterless amplifiers. Second, Class-D amplifiers can be a potent source of electromagnetic interference (EMI), methods are needed to characterize EMI and consider EMI already in the design process. Third, Class-D amplifiers are traditionally build in an analog manner. But in mobile devices, the signal sources typically have a digital nature. Approaches to implement high-fidelity, digital Class-D amplifiers are investigated. In this work an efficiency measurement setup is developed, which allows more accurate results than state of art setups due to lower capacitive loading. Then, a power-efficiency model for filterless amplifiers is derived, which is based on physical implementation parameters. The model allows to optimize the power efficiency before transistor-level implementation and thereby reduces design time. A main contribution of this work is a new emission measurement setup, which allows to quantify and compare the ability of different Class-D amplifiers to create EMI. Using this setup, it is shown that filterless amplifiers are especially prone to creating EMI. Further, a simulation tool is developed to evaluate EMI already during the design phase. Using this tool, several circuit level measures to reduce EMI have been investigated. Spectrum-shaping methods have been analyzed in depth. A novel integrated implementation of Slew-Rate-Control is proposed. These measures allow to build Class-D amplifiers with reduced EMI considering the tradeoff on audio-performance and power-efficiency. Several prototypes were built to demonstrate the feasibility of the presented circuit level measures and to evaluate the simulation and modeling tools. Finally, a novel circuit concept for a digital input Class-D amplifier is developed. This new concept allows to implement a highly linear, filterless amplifier which performs digital to analog conversion and power amplification inside the same circuit. The concept is suitable for implementation in deep sub-micron technology and has significantly relaxed requirements on the analog circuitry, compared to previous works.
|Qualification||Doctor of Philosophy|
|Award date||12 Apr 2018|
|Publication status||Published - 2018|