|
|
Oral History – Jack Haenichen
(Continued)
|
|
So, I
was designing these high frequency NPN transistors. We had an applications
department, which was like a customer, who would build circuits. I would
design devices for them. Leo Lehner of Applications and I collaborated on an
effort we called “The Bellweather Customer”. What we did was pick a few
companies which we felt were at the forefront of their business – these
were people that were making apparatus that used semiconductors, and we
felt that if they needed something, sooner or later, everyone else would
too. It was a very good strategy and worked greatly to our benefit.
The
three companies that we picked were Tektronix, in Beaverton Oregon, Hewlett
Packard in the Bay area, and E.H. Research, which was a very small company
in the Bay area that made state of the art pulse generators - it was a one
man show. The president and chief engineer was a guy named Jack Hubbs. Leo
and I would make the rounds every month or so to these places; we’d
introduce ourselves at first, and then we got on a first name basis with
the chief engineers with all these places. And I designed devices for
these guys. At this time, everything was NPN bipolar. It worked out very
well, because everyone else wanted the same devices. A lot of our
commercial success came out of that strategy.
|
|
|
Oral History – Jack Haenichen
(Continued)
|
|
9) How did the
discovery of the annular process come about?
It turns
out that a lot of guys were designing equipment that needed both NPN and
PNP devices, in complementary circuitry. So, no problem, instead of
starting with an N-type substrate, you started with a P-type substrate.
Well, I very quickly came upon a major problem, which evidently everyone
else had too. Nobody was able to make PNP planar transistors that had
breakdown voltages above about 50 volts. In order to get higher voltages,
you had to use a more lightly doped substrate, and when you went to this
lightly doped substrate (for PNP transistors) a phenomenon known as
inversion took place. Under the oxide layer, which is the masking and
passivating element of these devices, an inversion would occur and it would
switch over to N-type. So, here’s this little thin layer of N-type on top
of the P-type which ran all the way out to the edge of the chip, which was
jagged from scribing, and it basically made a dead-short. So, anybody who
was making PNP planar transistors could only make low voltage. Well, of
course, our customers wanted high voltage breakdown.
Go
To Haenichen Oral History, Page 9
|
|