Design and Implementation of a Re-Configurable Arbitrary Signal Generator and Radio Frequency Spectrum Analyser

Abstract

This research is focused on the design, simulation and implementation of a reconfigurable arbitrary signal generator and the design, simulation and implementation of a radio frequency spectrum analyser based on digital signal processing.

Until recently, Application Specific Integrated Circuits (ASICs) were used to produce high performance re-configurable function and arbitrary waveform generators with comprehensive modulation capabilities. However, that situation is now changing with the availability of advanced but low cost Field Programmable Gate Arrays (FPGAs), which could be used as an alternative to ASICs in these applications. The availability of high performance FPGA families opens up the opportunity to compete with ASIC solutions at a fraction of the development cost of an ASIC solution. A fast digital signal processing algorithm for digital waveform generation, using primarily but not limited to Direct Digital Synthesis (DDS) technologies, developed and implemented in a field-configurable logic, with control provided by an embedded microprocessor replacing a high cost ASIC design appeared to be a very attractive concept. This research demonstrates that such a concept is feasible in its entirety.

A fully functional, low-complexity, low cost, pulse, Gaussian white noise and DDS based function and arbitrary waveform generator, capable of being amplitude, frequency and phase modulated by an internally generated or external modulating signal was implemented in a low-cost FPGA. The FPGA also included the capabilities to perform pulse width modulation and pulse delay modulation on pulse waveforms. Algorithms to up-convert the sampling rate of the external modulating signal using Cascaded Integrator Comb (CIC) filters and using interpolation method were analysed. Both solutions were implemented to compare their hardware complexities. Analysis of generating noise with user-defined distribution is presented. The ability of triggering the generator by an internally generated or an external event to generate a burst of waveforms where the time between the trigger signal and waveform output is fixed was also implemented in the FPGA. Finally, design of interface to a microprocessor to provide control of the versatile waveform generator was also included in the FPGA. This thesis summarises the literature, design considerations, simulation and implementation of the generator design.

The second part of the research is focused on radio frequency spectrum analysis based on digital signal processing. Most existing spectrum analysers are analogue in nature and their complexity increases with frequency. Therefore, the possibility of using digital techniques for spectrum analysis was considered. The aim was to come up with digital system architecture for spectrum analysis and to develop and implement the new approach on a suitable digital platform.

This thesis analyses the current literature on shifting algorithms to remove spurious responses and highlights its drawbacks. This thesis also analyses existing literature on quadrature receivers and presents novel adaptation of the existing architectures for application in spectrum analysis. A wide band spectrum analyser receiver with compensation for gain and phase imbalances in the Radio Frequency (RF) input range, as well as compensation for gain and phase imbalances within the Intermediate Frequency (IF) pass band complete with Resolution Band Width (RBW) filtering, Video Band Width (VBW) filtering and amplitude detection was implemented in a low cost FPGA. The ability to extract the modulating signal from a frequency or amplitude modulated RF signal was also implemented. The same family of FPGA used in the generator design was chosen to be the digital platform for this design. This research makes arguments for the new architecture and then summarises the literature, design considerations, simulation and implementation of the new digital algorithm for the radio frequency spectrum analyser.