IEEE Solid-State Circuits Magazine - Winter 2014 - 10

Figure 3: The author with future wife, Liz, and a fellow student at the author's research
bench, University of Melbourne, 1965.

Daryl's example of rigorous analysis combined with presentation of
intuitive concepts and copious use
of real -world examples.
I began research toward my M.S.
under Daryl's supervision in 1964.
The noise theory of both BJTs and
junction field-effect transistors
(JFETs) had recently been elucidated, and Daryl thought a Master's
project involving the characterization of the noise in these devices
would be a good way for me to get
up to speed in electronic circuit
research. I designed and built my
own low-noise preamps (using vacuum tubes to begin with) and spent
many long hours in the lab pulling
all-nighters with the other graduate
students. That was an exciting time
for me, learning to make real circuits work and performing difficult,
sensitive measurements on real
devices. I devised my own means of
modeling transistors, created new
methods of parameter extraction
and got good agreement with the
prevailing theories. I managed to
derive a generalization of the theory of noise in JFETs, showing how
noise in any FET can be described
by a common underlying noise
theory, independent of doping profiles. Our group discovered that the
I-V characteristics of all JFETs were
well described by a square law, but

w I n t E r 2 0 14

we were scooped in the literature by
researchers from Cal Tech.
Daryl had assembled a top-flight
group of graduate students and we
learned enormous amounts from
each other, working day and night
side-by-side in the lab over a period
of years. I learned lessons like the
need to respect voltages greater
than 50 V when one of our number
got across a 5 kV regulated supply
one night. The event was announced
with a loud crack from the far end of
the lab, and the student staggered
out with his hair standing on end.

Figure 4: The author demonstrates mid
20th Century lab measurement techniques at
his research bench, University of Melbourne,
1965. Reproduced with permission from
John Wiley and Sons, Inc. from the book
"Amplifying Devices and Low-Pass Amplifier
Design", E. M. Cherry and D. E. Hooper,
Wiley, New York, 1968. Photo by Daryl
Hooper. In those days, the ears of the human
operator were often used as the detectors of
electrical events such as impedance bridge
balance.

IEEE SOLID-STATE CIRCUITS MAGAZINE

He announced that he was quitting
for the day.
Other lessons I learned included
the existence of phenomena such
as microphonics, when I noticed
strange disturbances one night in
the output of my sensitive preamp,
correlated with someone hammering
in the next room. My scope showed
bursts of a ringing signal appearing
at my pre-amp output whenever the
hammer fell (20 meters away). I had
to make my measurements after the
hammerer went home. I also discovered that the early MOSFETs then
available were very delicate and
died easily. In retrospect, I figure
they had no ESD protection and consequently had about 20 V ESD rating.
Since they cost $40 each in 1960's
money, it was a major (distressing)
event for a penurious grad student
when one died in your hands.
After a year spent measuring the
noise properties of innumerable
devices by watching the amplified
noise time waveform on a scope at
the same time as I manually measured the corresponding noise spectrum with a wave analyzer, I got to
the point where I could specify the
frequency spectrum of any noise
signal by simply glancing at its
time-domain waveform. This was
a very useful skill that seemed to
greatly impress casual observers.
After my M.S. work on device noise
characterization I searched around
for an interesting Ph.D. research
topic, and from scanning the recent
technical literature I decided that the
electrical noise behavior of radio frequency (RF) mixers would be a challenging topic. I got the OK from Daryl
and began attacking the problem. I
immediately ran into the fact that an
overdriven transistor is, by design,
a very nonlinear circuit and there
were few theoretical tools available
to tackle such problems. The first
digital computer had arrived on campus in 1964 (IBM 7044 - very puny by
today's standards) so I decided to try
to use this resource to solve the very
nonlinear differential equation (DEs)
I was encountering. This necessitated

Table of Contents for the Digital Edition of IEEE Solid-State Circuits Magazine - Winter 2014