High data-rate communication systems require the transmission of radio frequency signals which are modulated in both amplitude and phase, presenting peak power envelope swings exceeding ten times their average power level. In wireless communication systems, it is a significant challenge to transmit high peak-to-average-power-ratios signals whilst maintaining a high degree of energy efficiency. However, it is fundamental from an environmental and economical point of view. Within radio hardware, power amplifier modules are the most power hungry elements, and it is therefore of utmost importance to develop modules capable of maintaining high levels of efficiency over large output power dynamic ranges. A common approach to preserve the power amplifier efficiency is to use load modulated systems. This thesis investigates circuit level approaches to push the efficiency of load-modulated architectures to their fundamental limits, focusing on the power amplifier harmonic terminations. A theory is proposed to determine the intrinsic optimal load modulation of harmonically tuned power amplifiers using closed form equations. The theory is validated by simulations and load-pull measurements. The impact of the harmonic tuning on highly saturated Gallium Nitride (GaN) power amplifiers is investigated experimentally and found to be significant, with up to 3 dB fundamental output power and 50 percentage points efficiency variation on a prototype at 900 MHz. The outphasing load modulated technique was examined further due to its high efficiency enhancement potential that arises from its dual drive nature. It was observed that the optimal harmonic terminations vary for different output power levels. A design methodology was then developed to simultaneously maximise the back off efficiency and power utilisation for a given device. A 26 Watt 900 MHz GaN outphasing amplifier is presented, based on the proposed methodology, with a saturated power added efficiency of 76% and exceeding 58% over 8 dB of output power back-off.
|Date of Award||6 Nov 2018|
- The University of Bristol
|Supervisor||Kevin A Morris (Supervisor) & Souheil Ben Smida (Supervisor)|