skip to main content
Department of Electrical
and Computer Engineering
Menu
Home  Biography  Courses  Publications  Research  Gallery

  

Dr. Madhu S. Gupta 
Professor

RF Communcation Systems Industry Chair 
Professor of Electrical Engineering 
Director of the Communication Systems and Signal Processing Institute

Current Research Projects

  • Phase Noise in Oscillators

The importance of studying phase noise in signals arises from the fact that, in modern digital communication systems, information is increasingly encoded in the phase of a carrier signal for a variety of reasons, such as power efficiency, or robustness in the presence of nonlinearity in communication channel. Among the various possible goals of such a study, two having high practical value are the simulation of phase noise in nonlinear RF integrated circuits (RFIC), and comparison of alternative strategies for phase noise reduction. Accurate simulation, or computation, of phase noise spectra in an RFIC permits the optimization of circuit design for noise, design to specifications, tradeoffs, and first-pass success. Evaluation of phase noise reduction methods, that are based on device-circuit interaction, identify the degree of improvement, limitations, and attendant tradeoffs for each method. This investigation spans the physics of device noise generation, noise modeling, computer-aided design algorithms, and measurement of noise in devices and circuits.

  • Very Low-Power Low-Noise Devices and RFICs

Field-effect transistors in GaAs and other III-V compound semiconductor technologies are the principal family of candidate devices for applications where low current consumption, low noise, and high linearity are simultaneously required. Traditional low-noise FETs that require bias voltages of 3V or more, and optimal bias currents in the neighborhood of a fifth of the saturated drain current, are not suitable in portable and other battery-operated applications, and must be modified for operation in the range of 1 to 1.5 V. Optimization of such devices involves design for low knee voltage and low contact resistance; renegotiation of the tradeoff between noise and third-order intermodulation level; biasing schemes that tolerate unregulated voltages from a single supply source; and circuit designs having low idle currents and rapid turn-on and turn-off of power for power conservation. These techniques, the corresponding models, and design optimization are under investigation.

  • Noise in RF SOI MOSFETs

Silicon-on-insulator(SOI) technology is a promising approach for extending the application of silicon ICs into the RF region. It is also a promising approach due to its lower supply voltage, lower power consumption, lower parasitics, better isolation, simpler processing, smaller (compact) circuit designs, and lower cost compared with other Si-based RF technologies. The characterization and modeling of noise in these devices is needed to guide device design, selection of operating conditions, and establishment of achievable performance. This task involves careful measurement of low-level signals, and examination of the assumptions implicit in theoretical noise models. The ultimate purpose is to develop physically-based noise models for the device that can be used for device design, scaling, parametric study, and prediction of noise characteristics beyond the usually limited range of variables over which noise equivalent circuit parameters have been found by measurement.

  • RF and Microwave Packaging

It is well known that the cost, performance, and size of many RF and microwave subassemblies and modules is now constrained by their packaging rather than by their constituent components and devices. Moreover, packaging is a multidimensional problem, having multiple simultaneous goals of thermal management, mechanical strength, electromagnetic compatibility, reliability, and low lifetime cost, in addition to the RF performance. The packaging consists of the interconnection and housing of a variety of building blocks, that must collectively be reliable, testable, and at the same time the packaging should be as close as possible to the ideal of lossless, linear, reflectionless, non-interacting, non-resonant, transparent embedding for those building blocks. The objective of this task is to establish figures of merit for the performance of interconnects in a package so as to assist in making design choices, evaluating alternative designs methods, and negotiating tradeoffs among the multiple, often competing, performance requirements.

  • Physical Basis of Accelerated Life-Testing

The lifetime and failure rates for devices having a long lifetime under normal operating conditions are estimated by extrapolation from lifetime test data collected under highly-stressed operating conditions that shorten the lifetime to a convenient measurable interval. While much statistical research has focused on the parameter extraction problem (robust estimation of failure model parameter from data that is limited, truncated, or fluctuating), there has been little examination of the extrapolation validity problem (applicability of the model parameters, determined under stressed conditions, to the normal operating conditions). Physical models are being constructed that describe the degradation of a device on the way to failure in generic (rather than detailed microscopic) terms, and that employ weakly restrictive postulates based in statistical mechanics. The goal of this work is to arrive at testable criteria, expressed in terms of measurable quantities, that can be employed to validate the mentioned extrapolation.

  • Linearization of High-Efficiency Power Amplifier

A power amplifier is typically the largest, most power consuming, and most expensive component in a wireless transceiver circuit. Successive generations of wireless communication protocols imposed increasingly higher linearity requirements on these amplifiers, that are in direct conflict with other design goals of low cost, high DC-to-RF conversion efficiency, and compactness. We are exploring feedforward methods for linearizing the power amplifiers, that rely on high-speed sampling and signal processing.

  • Multimedia Instructional Modules in RF Engineering

In a rapidly evolving technical field such as RF engineering, the educational needs for the preparation of new professionals in the field are not adequately met by the usual existing means : university-based undergraduate and graduate education, professional short courses, and vendor-provided application materials and training sessions. Multimedia instructional modules can become a significant source of instructional material, with numerous desirable features. They do not require collocation of multiple learners to make instruction cost-effective; they can be updated frequently as and when the need arises; they permit easy collaboration of multiple experts; they can be concatenated to construct a customized program for each learner; they offer the convenience of learning at one’s own choice of time, place, and pace; and they can address highly specialized topics with insufficient demand at one time and place. We are developing such specialized offerings, and simultaneously gaining insight into the preferences and perspective of the learners.

 

Research Laboratories

  • Communication Systems and Signal Processing Institute

This institute is engaged in educational, research, and service activities in the field of RF wireless communications, which is an area of major strength special in the Dept. of ECE, as evidenced by the number of faculty involved in it, educational offerings and student enrollments in the graduate program, externally-supported and sponsored research activity on going, and industrial collaborations and support received. Faculty, students, and industrial partners collaborate to advance the state-of-the-art in the institute’s core areas of expertise, which include digital signal processing, and digital communication theory, and radio-frequency and electronics. Specific activities include research and design projects; development of advanced products, software, algorithms, and techniques; and training programs including short courses in rapidly moving fields. The faculty’s research interests encompass OFDM modems and polyphase DPS filters; RF power and low-noise amplifiers; MIMO and other wireless antenna systems; and spread spectrum communication technology. Prof. Madhu S. Gupta serves as the Director of the Institute, in which some half a dozen faculty members participate as investigators.

  • High-Frequency Electronics Laboratory 

This is a research laboratory for graduate students and faculty with interests in electronic circuits, devices, and systems that operate in the radio frequency range, nominally between 0.5 to 50 GHz. The laboratory has instrumentation for measurements and characterization of components, modules, and devices in this frequency range, including RF wafer probe stations for accessing integrated circuit chips and semiconductor wafers; network analyzers for broadband scattering parameter evaluation; spectrum analyzers for signal identification, analysis, and measurement; antenna measurement and characterization; and instrumentation for the observation and measurement of short-pulse phenomena. The laboratory is particularly equipped in the area of instrumentation for high-sensitivity measurements of very weak signals and of noise generated within the radio-frequency devices and circuits. The laboratory has been used for graduate thesis research, and for conducting sponsored research work from NSF, DARPA, U.S. DoD, and industrial sponsors.