HP Innovation Journal Issue 11: Winter 2018 | Page 55
A RETURN TO
FUNDAMENTALS
Lessons from our founders
Chandrakant D. Patel, Chief Engineer
and Senior Fellow, HP
It was the early 1990s. Even though Bill Hewlett and Dave
Packard were retired, they routinely came to their offices in
Building 3, Palo Alto, CA—then home of HP Labs and today
the home of HP’s headquarters and HP Labs.
One day, I met Bill in the HP Labs labstock. After a brief
conversation on the parts he was looking for—mostly fas-
teners and solder wire—I asked him if he would join me in
our interconnect and thermal technologies laboratory which
we fondly called the Tinker Tank. Bill, the tinkerer, agreed
and spent three hours in the Tinker Tank. The conversation
with Bill centered around what experimental projects I was
conducting in the lab, why was I doing the work, why it
was relevant to HP, what was the relevance externally with
reference to industry and academia, what was the state of
art, what were the limitations in state of art, my hypothesis
for advancing the state of art, whether my design of experi-
ments would prove or disprove my hypothesis, the results,
and the documentation of the results.
In describing the project, I started with the explanation of
the problem statement associated with removal of heat flux
from high-performance chips mounted on a substrate—a
multi-chip module. I articulated the various paths to trans-
fer the heat flux, and the state-of-the-art approach then
employed by IBM, Hitachi, CRAY, and other supercomputer
manufacturers. We also discussed the limitations of their
approaches and why those approaches would not work for
our 400 square millimeters, 100 W chip. Then, I went on to
explain our approach, the heat transfer, and fluid flow calcu-
lations in support of our approach and the experiments. In
articulating the experiments, I pulled out my lab notebook
and ran through the numbers.
That HP value of appreciation for fundamentals-based engi-
neering, thoroughness of experimental work and documen-
tation has led to breakthroughs that show up in our history
of innovations, from the audio oscillator and the atomic
clock to instruments, computing, printing, and 3D printing.
It has also influenced those of us who have spent decades
in the company. In my engineering work and in guiding my
team, my students, my mentees, and our children, I have
insisted that an engineering problem be written down.
I have required a clear problem definition, assumptions,
and first-order analysis on an engineering notepad or a labo-
ratory notebook prior to the start of any work. For example,
when devising the architecture for a software-defined data
center management system, multidisciplinary systemic 3D
Print solution, or in computer-aided design and engineering,
write it down first. Problem definition must have a sketch.
A southeast isometric view is a plus.
Understanding, assimilating and getting a first-order “feel”
leads to confidence. And that confidence is necessary for
the 21st-century cyberphysical age, an age that is driving the
integration of cyber and physical systems. It necessitates the
engineering thoroughness of the physical or the machine
age, breadth in digital sciences of the cyber age, and multi-
disciplinary collaboration across all disciplines to instantiate
systemic solutions.
In many ways, in the 21st-century cyberphysical world, old
school is new school.
“In many ways, in the 21st-century
cyberphysical world, old school is
new school.”
GREETINGS FROM HP CHIEF ENGINEER
53