Fourth Year Projects for 2008-2009


Group Project in Radio Frequency and Analog Design: Ultra-Wideband Radio

Joint between students reporting to Prof. Plett and Prof. Rogers.

The following is the blurb from April 2008

RFICs FOR UWB APPLICATIONS

While cellular and WLAN applications are some of the best known applications for radio frequency integrated circuits (RFICs), another very active area that has generated much interest is the design ultra wide band (UWB) systems for personal area networks (PAN) and other low power applications. Unlike narrow band wireless applications, UWB systems must handle many signals simultaneously over a wide frequency range.

Students in the project will be responsible for the design of a high frequency (either in the 3-10GHz range or in the 60GHz range) circuit building block and may also be expected to participate in designing the radio architecture. The circuit blocks will be designed in either a 130nm or 65nm state of the art CMOS process, using the same tools used in industry.

This project will make use of the skills students have developed in ELEC 2507, and ELEC 3509. Enrollment in courses such as ELEC 4505, and ELEC 4707 will provide additional background for the project. This project will be run jointly with Professor Rogers.

What is really being done is the following

Students are working on a 60 GHz radio in 90 nm CMOS. Components for both the receiver and transmitter are being designed. Individual students are designing the following:

Questions and Answers:

Question Are there any available spots?

Answer I don't determine that - if you fill out your preference form, the project coordinator places students into available spots. If you are interested in my project, since Prof. Rogers and I are combining our teams, you can increase your chances by listing both of us as preferences.

Question What software would be used for the analog signal circuitry design (possibly ORCAD?)

Answer Cadence based software to do schematic entry, simulation, and layout, etc. Orcad is owned by Cadence, but it is for printed circuit design - the software we use is for integrated circuit design and is the same software you would use in industry for IC design.

Question Could we get more details on the circuit blocks? design itself?

Answer There are many references available in the literature (if you use XPLORE from ieee.org you can search for UWB papers). Some blocks would be a low noise amplifier with bandwidth from 3.1 GHz to 10.7 GHz, gain over 10 dB, noise figure from 3 dB to 5 dB. This would be followed by a mixer or multiplier depending on which type of system we implement - this will convert the high frequency input signal to a low frequency signal. This would be followed by a low pass filter, then an A/D converter. On the transmite side, digital information might go to a D/A converter, or the bits would directly be used to modulate a high frequency signal (high frequency again being 3.1-10.7 GHz). A power amplifier would apply this high frequency signal to an antenna. There would also be an frequency synthesizer, or pulse generator depending on which type of UWB system we are implementing.

Question But how will we go about designing them?

Answer You would probably start with known topologies from the references, or what we provide for you. Later, as you become more familiar with the design, you may wish to make some design changes. You start by entering a schematic into cadencs, setting the circuit parameters, according to the function of the circuit and design conditions - we will provide more information on this, but essentially you will be sizing the transistors and setting bias currents and voltages to result in the best performance. Here performance might be low noise, or broad bandwidth, or high linearity. After the schematic design is complete, you do the layout, which is placing the transistors, running metal connecting lines, connecting inputs and outputs to pins. After layout you use another tool which compares layout versus schematic (LVS) to make sure they match. Then, another took extracts parasitic capacitance, then a simulation is done which includes these parasitics. If the resulting performance is no longer good enough, more design iterations may be required. During the design process, interfacing to the neighbouring blocks must also be considered, and at some point simulations will be run verifying that circuits function together as a system. Other considerations: designs must be protected against electrostatic discharge, it is hoped the circuits will continue to work even if there are temperature, voltage and process variations.

Question Will we be learning about the industry tools in the courses you requested to take in parallel?

Answer Courses will likely not directly teach Cadence based tools as we have limited licenses, typically not enough for a large class, however, there are other more simple tools used in courses which have similarities. Part of your first few weeks in the project would be to learn about the tools - and we will provide tutorials.

Question Are we going to be soldering the circuit together?

Answer No, this is an integrated circuit project, so interconnect between components is all on chip.

Question Which concepts for the courses we've done till now will we be applying?

Answer The main one would be analog circuit design in ELEC 3509 - designing analog bloack where you determine the transistor size and proper biasing, like the cascode amplifier design, computer based simulation of blocks, e.g., the opamp simulation. Concepts from other courses exist but much less than from ELEC 3509, e.g., some communications concepts, like AM, FM signals in SYSC 3501(?), the concept of impedance matching, which you will have touched upon in ELEC 3909, some things about transistor operation in ELEC 3908.

Question How much learning/reseach time do you think we will need before starting more on the project?

Answer The first number of weeks in the project will be learning tools, learning about the system, learning about the integrated circuit process, all in parallel with learning about the particular block you will be responsible for. But that is not before you start the project - you don't necessarily need to do research before you start the project, but of course you are welcome to do so as learning and doing research is never a waste of time.

Question Is it a continuing project or are we starting from scratch?

Answer Previous project students have worked on analog and radio frequency projects, but every year, we try to make it a different system, or change the specifications considerably, so that you have the opportunity to design your own circuits. But, of course there is a lot of literature and reference material available on related cicuits.

Question Will we be able to create a physical working prototype at the end of our project?

Answer The physical working prototype at the end of the project is not so physical, but the simulation of all the parts working together, demonstrating the complete system. During the eight months, we wouldn't have the time to fabricate and measure the circuits since fabricating an integrated circuit takes about an extra four months, after all the design is completed. So, to be able to measure them before the end of the project time, we would have to have completed the design by about December. Because of all the learning, the many steps in doing design and verification, this is not feasible, so instead, by the end of the 8 months, we try to have the circuits completely ready to be fabricated. If particular students are interested in continuing on to graduate level research after the project, then we have sometimes made these proejcts the basis for further research in which case they might be fabricated and measured.