Transcript tom_brazil - University College Dublin

Electronic Engineering –
Research that Transforms
our Lives
Tom Brazil PhD, FIEEE, MRIA
Professor of Electronic Engineering
University College Dublin
In the past 4 decades, no
other discipline has more
profoundly changed the way
we communicate, find
information, do business,
entertain ourselves…
… why is this?
… will it continue?
… what more is to come?
EE: “Electronic Engineering”
Some distinctive characteristics..
• EE involves both (charged) matter
and (electromagnetic) radiation;
• EE is highly research-intensive with a
strong coupling to mathematics and
physics;
• For over 40 years implementation has
been driven by an exponential
underpinning technology trend
(‘Moore’s Law’)
Various Perspectives on EE
• Level of Physical Reality:
– Electronic properties of materials/
Electromagnetic (EM) radiation
(devices, transistors, electronic circuits,
Integrated Circuits (ICs) etc..)
• Abstract Level:
– EE is about generation, processing, storage and
transmission of signals (i.e. information in
electrical form)
(algorithms, coding, modulation, software …)
These Two Perspectives Evident in
UG Degree Subjects…
• Solid-State Electronics; Electromagetics;
Electronic Circuits
– Physics; Mathematical Physics, Chemistry…
• Circuit Theory; Control Theory;
Communications Theory; Computer
Engineering; Linear and Nonlinear Systems;
Software
– Mathematics; Statistics; Computer Science …
… of course yet another
perspective of EE could be based
on Applications …
How to Proceed…?
• Overview of Major Trends In EE:
– taking the perspective of (dual) physical
reality in EE: PARTICLE (Charge) &
WAVE (Electromagnetic)
• EE at UCD:
– current status, future opportunities
Wave-Particle Duality
• The everyday physical world seems to us to be
composed of matter (dust, stars…) and radiation (heat,
light…)
• The insights of Einstein (1905) and de Broglie (1924)
contributed to the view that this may be just our
perception: at a deep level they could the same
• Eventually this concept of ‘wave particle duality’
became a fundamental aspect of the development of
the modern theory of quantum mechanics
E   
p   k
particle
wave
Duality in EE
• This wave-particle duality is also
fundamental to EE:
– The engineering of charged matter
– The engineering of the electromagnetic
(EM) spectrum
• EE is unique among the Engineering
disciplines in placing such strong
emphasis on both matter and the EM
spectrum!
Theme (1): The Engineering
of Charged Particles
Electronic Devices & Circuits
Electronic Devices & Circuits
• Era of Valve (Tube) Electronics ~1900s1950s
– First electronic device: audion tube (Lee de
Forest, 1905)
• Semiconductor Era ~1950s to 2020 (?)
– Effectively began with invention of transistor
in 1947 by Bardeen, Brittain and Shockley
• Post-Semiconductor Era?
– Devices based on new materials and/or more
directly exploiting quantum mechanics ….
Moore’s Law
In 1965, Gordon Moore, the founder of
Intel, postulated what was to come to
be known as ‘Moore’s Law’:
“the power of the silicon chip
will double every 18-24 months,
with a proportionate decrease in
cost “
222 ~ 4,000,000 over 40 years
Microprocessors
• In 30 years the number of transistors on a
microprocessor chip has increased by a factor
of over 77,000, from 2,300 transistors on the
4004 in 1971 to 178 million transistors on an
Intel Pentium® 4.
• Product speed has increased in a similar
fashion, with the earliest 4004 processors
running at 100kHz vs. the current Intel
Pentium® 4 Processor at 3.8 GHz, a factor of
38,000.
ICs: Huge Increase in
Complexity at lower cost…
Smaller means faster!
Current IC Technology
• Current technology ‘node’ is 65nm in
CMOS (‘nanoelectronics’)
• Severe challenges being overcome: material
dispersion, modelling, signal integrity,
thermal management…
• Reconfigurable computing concepts are
increasingly dominant
• Emphasis on System-on-Chip (Soc) or
System-in-Package (SiP) combining various
technologies in a single platform
Detailed Cross-Section near surface of
advanced CMOS IC
~ 2km of interconnect
on one 10mmx10mm IC!
6 separate layers
of metal
interconnect, with
connecting ‘vias’
MOS
transistor
What does the future hold?
“The scaling of CMOS device and process technology, as it is
known today, likely will end by the 16nm node (7nm physical
channel length) by 2019.” Source: ITRC Roadmap 2004
• The end of CMOS will effectively be
the end of the evolution of chargebased electronics;
• Beyond that new device concepts are
likely to emerge more directly
exploiting quantum mechanics;
Molecular Quantum-Effect Device
Theme (2): The Engineering
of Electromagnetic Waves
Wireless, Microwave…
Wireless, Microwave…
• Linked to the development of
electronic devices and components,
the history of EE has involved
exploiting the EM spectrum at
steadily higher frequencies
• Each advance has been critically
dependent on the availability of
sources, amplifiers and detectors
Wireless, Microwaves…
• Maxwell 1865 Theory of EM Waves
• Hertz 1885 Experimental demonstration of
radio
• Marconi
– 1895 (Bologna) First communication by microwaves
– 1899 (Dun Laoghaire) Worlds first outside sports
broadcast
– 1933 (Vatican City) First voice link by microwaves
• WWII Huge advances in radar technology
• Post-war: network TV, satellites, space ..
• 1990’s: Wireless revolution!
Vision of Future Wireless Systems
Technical Trends
• Data rates and system center frequencies
will increase.
• The most advanced transmission techniques
in space-time-frequency domain must be
used to increase spectral efficiency.
• Radio systems will be adaptive.
• Energy-constrained wireless nodes
• Networks are partially based on Ad-Hoc
techniques.
• Flexible radio transceiver techniques are
needed.
79GHz Anti-collision Radars
sensor
height
47 cm
The Final Frontier? TeraHertz
historical trend in EE
A vast
~unexploited
part of the
EM spectrum
optoelectronics
Problem to Date:
lack of sourcesbeing addressed
NB: Terahertz is
non-ionising!
Applications:
Numerous!!
(Radio
Astronomy)
A Radio Telescope for
Ireland (ARTI)
• Between 1845 and 1918, the largest (optical)
telescope in the world was located in Birr, Co.
Offaly;
• In 2002 Lord Rosse donated a site at Birr for a
proposed new radio telescope, linking Ireland to a
European network of such telescopes
• Just €10M needed now!
Research in Electronic
Engineering
• Intense levels of research and innovation
are characteristic of EE
• The research is highly rigorous and
scientific (e.g. IEEE Transactions…): often
closely linked to and even driving
comparable research in mathematics,
physics etc
• Research may often be very fundamental,
but a breakthrough can be immediately
applied in industry or to society
Research in EE in UCD
• Very strong, long-established ethos
of top-level international research in
Electronic Engineering in UCD
• Built up by Sean Scanlan (Professor
of Electronic Engineering, UCD 19732002)
• Excellent individual achievements:
challenge now is to expand on a large
scale at institutional level
EE Research Performance
• October 2004: 15 full-time academic staff;
• 48 PhD students, 12 research Masters (3.x
PhDs per academic)
• €2.4M cash income in previous 12 months
• Xx publications in leading peer-reviewed
international journals and conference
proceedings
Research Themes: UCD EE
• Broad Theme of Physical Layer
Communications
• Specific world-class teams in signal
processing, nonlinear circuits & systems,
microwave/RF and optoelectronics
• Generic competence in electronic
simulation, modelling and design
• Strong interdisciplinary focus around
rapidly growing area of biomedical
engineering
Design in Electronic Engineering
• Traditional Approach
– Use a mixture of intuition, past
experience, simple analysis, prototyping,
trial & error…
• Computer Aided Design (CAD)
– Create a mathematical model of possible
realisation and relate parameters of
model to measurable quantities. Then
analyse, tune, optimise…
EE CAD:
Circuit
Theory
CAD of Microwave Systems
e.g. GaAs MMIC for 77GHz automotive radar (ca. 1mm2)
Standard
Lumped/
Distributed
Circuit
Elements
Non-linear
Transistor
Models
dy t 
 f y t ,t 
dt
S-parameter blocks
A special
feature of
microwave
analysis
My Research: Microwave CAD
• Non-linear equivalent circuit and physicsbased models of microwave transistors;
• Algorithms for combining frequencydomain (S-parameter) blocks within
transient time-domain analysis;
• Behavioural level models for describing
complete high frequency systems
– SFI Investigator Award (€1.3M)
– EU Network of Excellence (TARGET)
– UCD Patent, EI Support, industry etc..
Non-linear Modelling of Microwave FET
V
Devices
G
RG
VS
RS
n
n 
   nv   


t
 t c
p
 v   p  p  v 
t
p 
 qnE  nT   


 t c
W
   v W  
t
Gate
QGD
QGS
VD
RD
N-epi
Source
Drain
iDS
(Intrinsic part only)
DGD
CGD
G
D
RGDI
CGS
DGS
RGSI
S
IDS
IAC
CX
GDS
 qnv  E    v nT 
W 
  u  


 t c
  E
  q N
d
 Na  n 
Hydrodynamic equations:
(n = electron conc.; p = momentum; W=
energy density; u = heat flux)
Equivalent circuit model
Time-Domain/FrequencyDomain Nonlinear Simulation
source
FET
model
General, linear
time-invariant
network
[S(f)]
FET
model
FET
model
behavioual or system-level model
load
Example of General Network: S21(f)
0
-10
-20
-30
-40
-50
-60
I.R. Weight
4
2
0
-2
-4
-6
0
2 4 6 8 10 12 14 16
Frequency (GHz)
100
0
-100
Voltage (V)
200
-200
0 2 4 6 8 10 12 14 16
Frequency (GHz)
0
0.5
1
Tim e
1.5
2
0.12
0.08
0.04
0
-0.04
-0.08
-0.12
0
0.4 0.8 1.2
Tim e (nsec)
1.6
Bevavioural Modelling: Discrete Time
Volterra Series
Linear system
m 1
y (n)   w(i )  x(n  i )
Nonlinear system
m 1
y (n)   ai [ x(n)]i
with memory
i 0
without memory
i 0
 w(0) x(n)  w(1) x(n  1) 
 a0  a1 x(n)  a2 [ x(n)]2 
m 1
m 1 m 1
i 0
i 0 j 0
2nd-order Volterra kernel
y (n)   h1 (i ) x(n  i )   h2 (i, j ) x(n  i ) x(n  j )
m 1 m 1 m 1
3rd-order Volterra kernel
  h3 (i, j , k ) x(n  i ) x(n  j ) x(n  k ) 
i 0 j 0 k 0
Volterra Series
(Complex Envelope Discrete Time
Formulation)
Nonlinear system
with memory
Baseband
Input
b
Main
Amplifier
Complex
Multiplier
Mod.
RF
Output
LUT
Predistortion
Calculation
Volterra
Model used to
Generate required
pre-distortion
Application of Volterra
Model: reducing nonlinear distortion
produced by
wireless amplifier
Dem.
RF
Output:
Without
Predistortion
With
Predistortion
Basic Research in EE
• Research is fundamental to EE
• The best of that research poses some of
the most fundamental mathematical and
physical challenges
• EE research draws on a broad range of
skills and disciplines (including circuit
theory, circuits and systems…)
• A breakthrough is immediately applicable!
EE Research in Ireland
• Science Foundation Ireland’s prioritisation
of ICT is welcome
• The Irish electronics industry does not
have a secure future without a strong base
of top-level national research
• BUT: Ireland’s academic research
infrastructure in EE is relatively weak
• Critical investments are now needed for
our future success
Conclusions
• The story of Electronic Engineering in the last 50
years is absolutely astonishing: it has truly
transformed our lives
• Yet there is vast amount more to come – the
historical pace of change will continue for at least
two more decades
• Ireland faces fundamental structural challenges in
responding to this. We are at a critical stage and
must make the right decisions for future economic
success
• Electronic Engineering is one of UCD’s true ‘star’
centres of research: it is ready and willing to
continue to play its part in building a successful
future for the university and the country