Electromagnetics & Microwave Research
Laboratory
INTRODUCTION
In this information age, the needs in the
microelectronics industry for manufacturability-driven design and
time-to-market demand powerful and efficient Computer-Aided Design (CAD)
techniques. Furthermore, the recent advances in radiofrequency (RF) and
Microwave Integrated Circuits (MICs) require a
permanent upgrading of existing CAD tools. In fact, with ever-higher
integration and miniaturization, the needs for concurrent and
multi-disciplinary design become increasingly important, requiring that the
available CAD tools be not only fast but also accurate so that the design can
be achieved reliably. Fast, because of the repetitive
computations involved in simulation and optimization of modern MICs. Accurate, because to be close to experimental data,
simulated circuit responses need to be obtained from component models that
fully integrate higher-order
nonlinear, thermal, and electromagnetic (EM) effects.
The central
focus of the Electromagnetics &
Microwave Research laboratory is the development of Computer-Aided Design
(CAD) models and software tools for high frequency linear/non-linear system
design, and the validation of these tools by checking their predictions against
measurements.
Led by Dr. Mustapha C.E. Yagoub, our team has a good
experience in advanced device modelling, radiofrequency and microwave
simulation algorithms, neural-based modelling, and circuit analysis, simulation
and optimization. Development of computational electromagnetic techniques for numerical modeling of
guiding as well as radiating structures is
also being carried out.
RESEARCH TOPICS
We are always interested in solving
problems that engineers in industry could face. However, as regard to our
planned research, we are currently working on the following research tasks:
EM-Based CAD Tools for RF/Microwave Circuit Modelling and Design
Any
RF/microwave circuit simulated response, like gain or dispersion, depends
closely on how the implemented miniaturized components are modeled. To be
efficient, this aspect requires the resolution of several EM issues related to
the complexity of the integrated structure due to the hybrid nature of the EM
field. Strip thickness, finite conductivity, and number of layers, are factors
that can significantly affect the propagation and attenuation characteristics
of EM field in high-density multilayered hybrid/monolithic MICs.
Such characteristics have been widely investigated using various techniques
like the perturbation technique, the mode-matching method, and the method of
lines. However, the perturbation approach is not suitable for MICs since the skin depth and the strip thickness are in
the same order, while the above fullwave methods are
CPU time consuming. Thus, easier and faster methods need to be developed to
meet the evolution of MICs.
Our team has
developed various EM numerical tools using the spectral domain method through
the derivation of dyadic admittance Green’s functions via a recursive process,
which significantly enhances the CPU time.
EM-Based CAD Tools for RF/Microwave Circuit Simulators
Another key aspect to be
considered in MIC design is the parasitic coupling. Coupling is becoming more
of a concern in MICs as packing densities of devices
are increased and frequencies are pushed higher. Although such coupling plays
an important role in circuit performance, this quantity is very complex to
evaluate and several numerical techniques in bringing this forward into circuit
design space have been investigated. Mainly based on the resolution of
Maxwell’s equations, such EM methods have demonstrated their efficiency, but
still require huge computation time and memory space. This aspect is crucial
when modern design tools lead to massive and highly repetitive computational
tasks during simulation, optimization and statistical analysis. Furthermore, recent lumped EM-based models seem to be helpful in obtaining a quick
initial design, but they are developed under perfectly shielded conditions
excluding parasitic coupling between neighbouring components, and hence, they
tend to be inefficient for circuit-level simulation.
In
this topic, our team has proposed a novel technique for computation of EM
parasitic coupling in passive structures based on simple circuit theory and
de-embedding concepts.
Compact Transistor Models For RF/Microwave Circuit Design
Since Field Effect (FETs) and Heterojunction Bipolar
Transistors (HBTs) are widely used in the
RF/microwave range, a large number of modelling approaches are being proposed.
Detailed physics-based transistor models are accurate, but slow. Table look-up
models can be fast, but suffer from the disadvantages of large memory
requirements and limitations on number of parameters. Nevertheless, they are
difficult to develop, equivalent circuit models remain
the most common modelling approach, where the element values can be determined
either by direct extraction or by optimization-based extraction. Fast and
simple to implement, direct-extraction techniques provide adequate values for
the more dominant circuit elements, but they cannot determine all the extrinsic
elements uniquely. On the other side, optimization-based extraction techniques
are more accurate but computationally intensive and relatively sensitive to the
choice of starting values. Furthermore, to make them attractive to
non-experienced users, such extraction techniques often assume a prior
universal circuit topology referred as the FET standard topology or the HBT
standard topology.
Determining the most suitable
small-signal equivalent circuit configuration (topology) and accurately extracting
its element values was the preliminarily target of our team. Based on an
exhaustive literature review, a circuit library has been created that contains
the most widely used topologies. Thus, by combining the Fuzzy c-means method
and neural network capabilities, a method was developed that efficiently
selects the most suitable small-signal circuit topology for a given set of
measured S-parameters. The goal is to extend this generic approach to generate
nonlinear/thermal FET and HBT models. In fact, many roadmaps now recognize
advanced nonlinear and thermal analyses as two of the major challenges in
electronic product innovation due to ongoing push of technology towards higher
power and complexity, and reduced size.
Advanced Tools for RF/Microwave Antenna Design
Our team has demonstrated its know-how in antenna design. We completed
successful contracts for the Defence
Research and Development Canada (DRDC) on the topic of designing planar antennas for ultra wideband (UWB) applications, especially for direction finding systems. We developed an innovative
technique for UWB (2-18GHz) sinuous
planar antenna design and balun integration, along
with original non-uniform tapered UWB couplers for the sinuous antenna feed.
Furthermore, we designed new kinds of very sensitive and highly
miniaturized dosimetric probes for accurate measurements of EM radiation effects on the human body. Such
probes will enhance the knowledge of EM radiation effects of cellular phones,
microwave ovens and other daily RF equipment on human health.
RESEARCH
COLLABORATION
Our
team is developing a
very fruitful collaboration including common research axes and training through
supervision, intensive courses and seminars. Part of this collaboration is successfully
supported by granted research programs. In fact, we maintain a vast network of collaborations:
·
·
·
·
·
RESEARCH FACILITIES
The team forms part of the RF &
Microwave Group (whose website can be found at www.genie.uottawa.ca/research/rsrch_site.php).
The following facilities are available in our laboratories:
•
Far-Field
Anechoic Chamber
•
Planar
Near-Field Anechoic Chamber
•
Network
Analyzers (Up to 60 GHz)
•
Spectrum
Analyzers
•
load-pull bench for
embedded and on-wafer transistor characterization
•
Commercial
and In-House Electromagnetic Simulation Tools (HFSS, Ansoft
Designer, IE3D, FEKO, Agilent ADS)
Our technician Mr.