WILTRONs Model 360 was regarded by customers as the most sophisticated
instrument that had ever been available to microwave engineers, but
before they invested $100,000 to $200,000 for an instrument of this
kind, the key factor in the customers mind was going to be its accuracy.
My overriding concern was that WILTRON had to have state of the art
accuracy and resolution. If there had been any shortcomings in either
of these areas, we would have been chopped to pieces by Hewlett-Packards
large sales force. Their salespeople in this specialized microwave
area outnumbered ours by 20 to 1.
My lead designer had the idea to pursue a more economical instrument,
one with reduced accuracy and resolution. I had to convince him that
due to the nature of our competition we had to first prove ourselves
in this new product area with the most accurate instrument available
and only after we had proved ourselves in this respect could we come
out with a less expensive, reduced performance instrument.
I sold this concept to my designers; it made sense to everyone. But
then came the subject of resolution. Just a cursive error analysis
showed that due to mismatch reflections one could never be sure of
more than 1.0° accuracy, therefore a resolution of .1° seemed
to my design engineers to be more than enough. From my experience
in the field I knew that engineers in many of their applications wanted
to watch little changes in phase. To them, resolution was tantamount
to accuracy. I had personally visited many of our customers
labs and production facilities and in the process acquired a year
or two of experience on all their variety of measurements. Fortunately,
catching the product in this phase of design it was virtually as simple
as going to an extra digit on the DAC (Digital to Analog Converter)
in order to realize .01° resolution. In the final analysis, it
was this extra resolution that made us competitive.
It really made me feel good to create an instrument totally useful
to the worlds scientific community. Imagine my situation, I
essentially bet my laboratory for three years on an advanced product;
I committed 20 of my 30 design engineers for a period of three years.
Consider that a man year of design time costs $150,000. That is a
ten--million-dollar design effort, or more than five times larger
than any previous design effort we had made. I made this enormous
commitment with full confidence it would be a winner.
It turns out we did just enough: every single feature built into the
instrument was necessary, and hardly any of our time was wasted with
false starts or trying difficult ideas. We planned what we needed
to do and did it. Our competition, good old Hewlett-Packard, was fast
on our heels.
About midway in our program we hired a design engineer who had recently
left Hewlett-Packard. He was a brilliant fellow who in about three
months learned essentially everything we were doing. After working
for us four months he left and rejoined Hewlett-Packard in their Santa
Rosa division where their instrument competitive to ours was being
developed. On leaving our company the engineer in question said he
wouldnt be working in exactly the same area and, of course wouldnt
breathe a word about what we were doing. In truth, he was an out-and-out
industrial spy. He had been planted in our lab but there was nothing
we could do about it.
We learned afterwards that Hewlett-Packard reviewed their whole project
the month this fellow joined them and adopted several of our concepts.
Happily, there was one fundamental improvement we had made that in
order to incorporate they would have had to start all over again and
would have lost two years. They elected not to lose those two years
and were at a competitive disadvantage from then on. Their system
phase locked to the sweeper whereas ours locked to a crystal reference.
If during a series of measurements the ambient temperature changed
just a few degrees there was a notable shift in their data, at least
ten times as much as ours.