Back to list of Group's MSc Theses
Abstract:
The Square Kilometre Array (SKA) is an ambitious multi-billion rands project to build the worlds’s largest and most sensitive radio astronomy telescope. The telescope will have a collection area of 1 million square metres and will be able to look at a larger portion of the sky than current telescopes. The sensitivity of this telescope is of such a nature that isolation of this instrument from the negative effects of man-made noise is paramount. As such a campaign to measure the levels of electromagnetic radiation at the prospective sites had to be conducted. In response to this, the South African (SA) SKA team developed the RFI Measurement System 2 to perform these measurements.
This dessertation documents the findings of a study of the functionality of RFI Measurement System 2 that was conducted by the author. An in depth study of MEMO 37 (RFI Measurement Protocol) was employed first. After which, an exhaustive audit of all the equipment used and a thorough analysis of the method of data collection and processing was conducted. A sample of raw uncalibrated, MODE 1 field measurement data was scrutinized. The data chosen was for RF signals in the 70 - 150 MHz frequency range and following similar data processing procedures as that of the SA SKA team, the plots of Spectral Flux Density versus Frequency were generated in the laboratory using MATLAB.
It was found that for the frequency range 80 MHz to 150 MHz, these graphs closely matched that of the SA SKA team and that for this bandwidth, both the SA SKA graph and that replicated by the author were within their tolerence limits. A relatively strong correlation coefficient of 0.4430 was observed for the maximum values graph. The range 70 MHz to 80 MHz lies outside the operational ability of the antennas as prescribed by the manufacturers and for this reason some differences between the work of the SA SKA team and that of my work was observed for this range.
Back to list of Group's MSc Theses
Abstract:
The Karoo Array Telescope [KAT] is a South African project that is attempting to build a worldclass radio telescope in the Northern Cape. The first prototype phase of the project was called the eXperimental Development Model or XDM. This MSc project involves the development of a carrier card that was planned to be used for XDM. The card, called the XDM Carrier Card, or XCC, was designed to be used as part of a modular Digital Signal Processing [DSP] architecture.
The XCC’s external connections are summarized in the context diagram in Figure 1. The card was designed to support two daughter cards, high-speed data network connections, a control network connection and power and synchronization signals from a backplane. The modular architecture was chosen to allow the boards to operate in different modes, depending on the application, for example, as digital receivers or data processors. The primary function of the carrier card was to perform communication and control tasks.
Typically, radio astronomy applications require substantial DSP hardware to sample, process and transfer incoming data. The demands on DSP performance are growing rapidly, as new instruments incorporate more dishes and seek improved performance. In the past, these instruments were designed for very narrow applications and took anything between three and ten years to develop. However, the fast growth of the processing power of electronics often rendered these designs obsolete even before they were deployed.
In the last decade, Field Programmable Gate Arrays [FPGAs] have provided an agile means of performing DSP. They facilitate rapid development of digital systems and lend themselves to relatively simple hardware as they increasingly allow more tasks to be performed on a single Integrated Circuit [IC]. These features allow for frequent updates of hardware, allowing the Moore’s Law curve to be followed. For these reasons, FPGAs have been widely adopted for use in radio astronomy projects, such as the Allen Telescope Array and in KAT itself. Thus, FPGAs form a central theme in the design of the XCC.
Back to list of Group's MSc Theses
Abstract:
This thesis describes the design and implementation of several instruments for digitizing and processing analogue astronomical signals collected using radio telescopes.
Modern radio telescopes have significant digital signal processing demands that are typically best met using custom processing engines implemented in Field Programmable Gate Arrays. These demands essentially stem from the ever-larger analogue bandwidths that astronomers wish to observe, resulting in large data volumes that need to be processed in real time.
We focused on the development of spectrometers for enabling improved pulsar2 science on the Allen Telescope Array, the Hartebeesthoek Radio Observatory telescope, the Nan¸cay Radio Telescope, and the Parkes Radio Telescope. We also present work that we conducted on the development of real-time pulsar timing instrumentation.
All the work described in this thesis was carried out using generic astronomy processing tools and hardware developed by the Center for Astronomy Signal Processing and Electronics Research (CASPER) at the University of California, Berkeley. We successfully deployed to several telescopes instruments that were built solely with CASPER technology, which has helped to validate the approach to developing radio astronomy instruments that CASPER advocates.
Back to list of Group's MSc Theses
Abstract:
This thesis describes the design and implementation of a low-cost high speed data capture card for the Hubble Sphere Hydrogen Survey (HSHS).
The Hubble Space Hydrogen Survey was initiated in an effort to build a low-cost cylindrical radio telescope for an all sky redshift survey with the observational goal to produce a 3-dimensional mapping of the bulk Hubble Sphere using Hydrogen 21cm emissions.
This dissertation first investigates the system design to see how each of the user specifi- cations set by the planning team in [33] could be achieved in terms of design decisions, component selection and schematic capture. The final design, AstroGIG, satisfies the user specifications by capturing data up to a full power bandwidth of 1.7GHz with an instantaneous bandwidth of <250MHz while maximizing the dynamic range. AstroGIG buffers, processes, stores and finally transmits the data through a 4-lane PCI-Express interface to a standard PC where the majority of the processing is performed. The system implementation is then described where issues relating to the process of transforming schematics into a physical PCB, and HSHS integration are discussed. The design is verified through Hyperlynx simulations to give a high degree of certainty that physical implementation and production would be successful. Results from tests on the actual hardware characterizing the overall system performance are presented. Conclusions are drawn based on these results and suggestions for future work and design improvements are recommended.
Back to list of Group's MSc Theses
Abstract:
This dissertation provides a comprehensive discussion around bistatic radar with specific reference to PCL, highlighting existing literature and work, examining the various performance metrics. In particular the performance of commercial FM radio broadcasts as the radar waveform is examined by implementation of the ambiguity function.
The FM signals show desirable characteristics in the context of our application, the average range resolution obtained is 5.98km, with range and doppler peak sidelobe levels measured at -25.98dB and -33.14dB respectively.
Furthermore, the SDR paradigm and technology is examined, with discussion around the design considerations. The USRP, the TVRx daughterboard and GNURadio are examined further as a potential receiver and development environment, in this light.
The system meets the low cost ambitions costing just over US$1000.00 for the USRP motherboard and a single daughterboard. Furthermore it performs well, displaying desirable characteristics, The receiver's frontend provides a bandwidth of 6MHz and a tunable range between 50MHz and 800MHz, with a tuning step size as low as 31.25kHz. The noise characterisation of the receiver reveals a NF of 10dB, a sensitivity of -105dB and a dynamic range of 62dB.
Finally, the investigation into the stability of the daughterboard frontend oscillators due to ageing effects is shown to be steady with acceptable levels of variation, showing a fractional frequency variation from 2.3 to 0.8 parts per 10 million and a maximum frequency drift of 4Hz.
Back to list of Group's MSc Theses
Abstract:
The full range of radar types and innovations can be complex and difficult to prototype, especially for institutions that wish to perform a wide range of experiments for low financial cost. Radar and sonar system development can benefit from digital technology that is powerful for research purposes, easy to use, and inexpensive.
The purpose of this thesis was the development of a sonar application using the Universal Software Radio Peripheral and the GNU Radio software framework. These are Open Source tools created for the software-defined radio community. These tools provide a powerful yet flexible means to experiment with a wide range of radio frequency applications, using a minimal amount of relatively cheap hardware. In this thesis, these tools were modified from their original telecommunications purpose, to produce a sonar system that could eventually be scaled to a prototype radar system using the same device and software framework.
Back to list of Group's MSc Theses
Abstract:
The Square Kilometre Array (SKA) is an international effort to build an advanced and highly sensitive radio telescope. The South African Karoo Array Telescope (KAT) project is initiated to show the international commitee of SKA that South Africa has sufficient technological background and resources to make a contribution to them. Therefore this research investigates and evaluates the performance of the Australian built 3x4 receiver module and then to verify that this prototype is suitable to be integrated into the 24- channel RF rack for KAT.
This disseration starts off with a brief introduction of the SKA project, and explain how it relates to the KAT project. Then certain receiver design techniques and parameters will be discussed together with receiver design trade-offs will be presented. This dissertation will then focus on the actual simulations of the 3x4 receiver module using the time-domain RF simulator, SystemView. An overview of the design for the 24-channel RF rack integrated locally by Tellumat (Pty) Ltd is presented and acceptance tests will be conducted and the test results will be presented. Results obtained from both simulations and measurements are analysed and compared.
This dissertation is concluded by discussing the conclusions and recommendations are presented for how to improve the accuracy when simulating the 3x4 receiver module.
Back to list of Group's MSc Theses
This page was last updated in January 2009 (RL)