IEEE Solid-State Circuits Magazine - Winter 2016 - 42

This article is about the Miles I know best-
Miles the consultant, the engineer, the friend
for over 21 years.
VCOs, PLLs, and clock and data recoveries (CDRs) in what became the
world's first single chip 10-Gb/s SONNET and 10-G-Ethernet transceiver
products. He worked so hard and under so much pressure to complete the
design by the beginning of September
2000 that he confessed of dreaming
about circuits, schematics, and simulations in his sleep.
Even today, 15 years later, when
I run into Quake's early customers,
they mention Quake's low-jitter CDRs
as the gold standard for the industry.
Two years after the consultancy
work at Quake, Miles had a duodenal

cancer, which was removed by surgery. A suspicion of the neurologist
was that the GBS was responsible
for his immune system reacting to
the early stages of duodenal cancer.
Miles managed yet another spectacular recovery.

Power Series Hand Analysis
of Phase Noise Sidebands
Already 70 years old by the fall of
2002, everybody would have thought
that Miles was done with research.
Yet his determination to understand
the interaction between large-signal waveforms and intrinsic noise

Boundary of 1-Port as Defined

Load

VBASEBIAS

Lpt

(Plus Any
Following
Circuitry)

HBT as
Used
in Oscillator
C1

Rpt

(Inside 1-Port)

IBIAS

Figure 2: A simplified large-signal model used in the power series analysis [8]. (Printed
with permission of Cambridge University Press from High Frequency Integrated Circuits by
Sorin Voinigescu.)

W I N T E R 2 0 16

IEEE SOLID-STATE CIRCUITS MAGAZINE

sources in oscillators was as strong
as ever. Always mindful of practical
aspects, he was keen to exploit it in
the algorithmic design of low-phase
noise VCOs operating at the designerspecified oscillation frequency and
with the designer-specified output
power. It was art for art's sake, with
no interest in publications or consultancy. Miles was just seeking knowledge and having fun in the process.
I set him up with SiGe-HBT HICUM
models, which he could use with a
harmonic-balance simulator to verify
his analytical, power series model of
the Colpitts oscillator. We would have
conference calls almost every week,
and I sometimes I would visit him
with my family at his summer home
in Gananoque, at the Eastern tip of
lake Ontario, spending the weekend
with him and Margaret, talking VCO
phase noise, playing family games,
swimming, and sailing with Captain
Copeland proudly at the helm.
Figure 2 illustrates Miles's unique
talent for creative simplification of the
strongly nonlinear Colpitts VCO model.
This circuit representation allowed him
to develop an analytical model based
on a seventh-order power series expansion of the oscillator tank voltage and
current, mixed with intrinsic circuit
noise sources. Unlike many others who
attempted to model phase noise analytically before him, Miles understood
the importance and the complexity of
the nonlinear high speed HBT model,
whose simplification made any precise analysis an elusive goal. Instead,
he described the nonlinear part of the
oscillator as a black-box and came up
with yet another disarmingly simple
equivalent circuit addition, shown
in Figure 3. This scheme accurately
captures the non-linearity of the tank

Table of Contents for the Digital Edition of IEEE Solid-State Circuits Magazine - Winter 2016