Radio-Frequency
(RF) and Microwave Engineering have grown rapidly
in importance in recent years, stimulated in particular
by the exceptional world-wide growth in wireless communications.
A broad fundamental trend is evident towards higher operating
speeds, higher centre frequencies and greater bandwidths. The
main focus of the RF and Microwave Research Group is to
address a selected set of core basic research problems underpinning
future progress in high frequency/high-speed design, as
follows:
New Approaches
to High-Frequency Non-linear Active Semiconductor Device
Modelling
Non-linear
Equivalent Circuit Modelling and Characterisation of High
Frequency Transistors
High Efficiency Power Amplifier Circuit Design: Switch-Mode and Beyond
Digital Predistoriton of RF Power
Amplifiers for Broadband Wireless
Communication Systems
Digitally Compensated High Power and High Efficiency Digital Mixing RF-DAC
Broadband Millimetre-Wave Transceiver Architectures for 5G Cellular Communications
Time-Domain/Frequency-Domain
Simulation Algorithms and Model Reduction Techniques
New Approaches to High-Frequency
Non-linear Active Semiconductor Device Modelling
It is widely recognised
that semiconductor device modelling is one of the most critical
issues in high-frequency design. In principle, the most
powerful and general approach to the modelling and simulation
of high-frequency transistors is to perform a detailed numerical
solution of the basic semiconductor transport equations
for charge carriers in several spatial dimensions and time,
consistent with the Poisson equation. Nano-scale geometries
and quantum mechanical aspects of operation mean that complex
hydrodynamic-type formulations of carrier transport are
generally considered necessary, thereby increasing the computational
effort involved. While important advances have been made
in the efficiency of such codes, they remain generally far
too complex for routine circuit design tasks.
Direct numerical solution
of device transport equations for a transistor, and modelling
approaches based on equivalent circuit representations of
the device, are often seen as essentially competing approaches
within nonlinear high-frequency CAD. Each method has clear
advantages and limitations. The objective of this research
is to demonstrate the potential benefits to be obtained
by combining the best features of both approaches, using
a simplified physical model which is highly computable and
yet retains key consistencies with the internal semiconductor
dynamics in terms of both particle and displacement current.
back to project
list
Non-linear Equivalent Circuit
Modelling and Characterisation of High Frequency Transistors
This activity has continued
as one of the major research themes within the Group, and has been
extended to new device technologies and material systems,
both III-V and silicon-based. The importance of this kind
of modelling is growing in high frequency circuit design,
which is increasingly oriented towards MMIC implementation.
Stringent system specifications place severe demands on
model predictive accuracy, while at the same time models
must be "compact" and efficient. The COBRA model
for PHEMT’s and MESFET’s developed at UCD has
proven to be an excellent general-purpose scalable representation
for these kinds of devices. It has been extended and improved
to act as the basis for equivalent circuit models for other device classes:
SiGe-based MODFET and RF CMOS. We are currently developing models for wide band gap GaN HEMTs devices.
back to project
list
High Efficiency Power Amplifier Circuit Design: Switch-Mode and Beyond
The RF Power amplifier (PA) is a system that converts DC electrical power into added radio frequency (RF) output power, and it is an integral part of all base stations and mobile units in wireless communication systems. Normally, the PA consumes over 50% of the total transceiver power. Remarkably, even the most advanced existing power amplifiers in cellular base stations only have an average efficiency of 40%. When considering other factors in the system, such as, feeding/coupling loss, the average overall power efficiency of base station PAs is less than 20%, which means that over 80% of the power used to transmit the information signal is wasted in the form of heat. This power waste is not only environmentally undesirable and contributing to an increasingly large share of global CO2 emissions, but it also significantly increases the operational cost of the networks, as, for instance, a large cooling system must often be installed to reduce the temperature of the transceivers. The objective of this research is to develop novel power amplifier architectures and new forms of amplifier circuitry, in order to satisfy the stringent requirements of future wireless systems.
back to project
list
Digital Predistoriton of RF Power Amplifiers for Broadband Wireless Communication Systems
It is well known that power amplifiers are inherently nonlinear, and when operated near saturation, cause inter-modulation products that interfere with adjacent channels. To reduce distortion, the PA must be backed-off to its linear region, which leads to high energy consumption and therefore low efficiency. Digital predistortion (DPD) technique is proposed to compensate for the nonlinear distortion generated by the PA thereby allowing it to be operated at higher drive levels for higher efficiency, using advanced digital techniques. The attraction of this approach is that the nonlinear PA can be linearized by a low-cost standalone add-on digital block, freeing vendors from the burden and complexity of manufacturing high-cost complex analog/RF circuits. Through extensive research led by Prof. Anding Zhu in the past years, the UCD's RF & Microwave Research Group has built up excellent expertise in this field and the UCD group is widely recoginized as one of the very few top-level teams in digital predistortion of RF power amplifiers in the world today.
Although DPD has been researched and developed for a number of years and several commercial products/solutions are available in the market, future DPD deployment faces significant challenges because the continuously-increasing demands of higher data rate and wider bandwidth. For example, in 5G systems, over 1GHz instantaneous modulation bandwidth is required. This will create severe challenges for DPD design. In this project, we aim to develope novel DPD techniques for future ultra wideband systems.
back to project
list
Digitally Compensated High Power and High Efficiency Digital Mixing RF-DAC
With increasing demands for enhanced data services, wireless communications networks are facing significant challenges since different radio technologies, multiple transceiver architectures and increasing frequency allocations make the wireless transmission more complex to control. The digital mixing RF-DAC combines digital to analog converter and up-conversion mixer in a single building block. Thereby, the analog baseband blocks are eliminated and replaced by reconfigurable digital blocks. The RF-DAC extends the scope of digital signal processing right up to the antenna and it enables synthesis of digital baseband signals directly at the final RF frequency, which provides great flexibilities and high performance in multi-standard reconfigurable radios. The aim of this project is to develop a wideband high power, high efficiency and fully integrated digital mixing RF-DAC with high resolution and superior linearity on 28 nanometer CMOS.
back to project
list
Broadband Millimetre-Wave Transceiver Architectures for 5G Cellular Communications
With the continuously increasing demands for higher data capacity, it has been growing interests in developing cellular systems to be operated in the millimetre-wave (mmWave) bands, between 30 and 100 GHz. The development of cellular systems in mmWave however faces significant technical obstacles. The propagation can be very poor and power efficiency is very low with the existing architectures and technologies. Linearity is another serious issue. New research breakthroughs will be essential to make cellular communication at mmWave frequencies possible. In this project, we are aiming to develop novel mmWave transceiver front-end architectures and related circuits and systems using advanced CMOS process and combining with the emerging GaN device technologies. In particular, the research will focus on power efficiency and linearity enhancement in power amplifier design and digital intensive transceiver architectures.
back to project
list
Time-Domain/Frequency-Domain
Simulation Algorithms and Model Reduction Techniques
A fundamental problem
in large-signal microwave system simulation is that active
devices are best described in the time-domain, whereas the
circuit environments external to these devices are usually
distributed in nature and may show significant loss and
dispersion effects, so that they are much more satisfactorily
described in the frequency-domain. Research at UCD has pioneered
the use of discrete-time convolution techniques that allow
an efficient and highly accurate solution to be obtained
combining information from both domains. This approach has
been developed further to include multiple active devices
and complex digitally-modulated signal formats. It has allowed
a rigorous evaluation of the relative accuracies of several
architectures of widely-used AM-AM/AM-PM -type behavioural
models. In addition, it is widely recognised that a fully
integrated electrothermal simulation is of great importance
in properly accounting for distortion and dynamic behaviour
in microwave high power amplifiers.
back
to project list
top
