> Inventing the transistor was one thing. Being able manufacture it as
> a reliable and inexpensive product was a difficult long task.
> According to the IBM history, at first transistors were made by hand
> -- someone jiggled the cat whiskers and watched a scope until the
> proper effect was created. Obviously very expensive and error prone
> way to go. Even after automation yields of working transistors were
> low. IBM research not only was developing new computers to use
> transistors, but also new technology to manufacture transistors and
> circuit cards. IBM failed to patent or license the manufacturing
> technology not realizing how valuable it was and let its
> subcontractors take it and re-use it. (Kind of like PC-DOS).
> The end result was that until the late 1950s, transistors cost more
> than tubes.
The earliest transistors were the "point contact" type, and made on a
base of germanium rather than the silicon which everybody associates
with transistors these days. Semiconductor diodes were made the same
way, not really dissimilar to the "cat's whisker and crystal" diodes
which had been in use as radio detectors for many years. It was a
little later that "junction" diodes and transistors appeared to
eliminate the cat's whisker approach, although the point-contact
transistor still had certain advantages over junction types for some
applications.
> Another issue was the learning curve. Electronic engineers by that
> point had long experience with vacuum tubes--they knew what they could
> and could not do and their operating characteristics. After the war,
> both the television and computing engineers extensively studied and
> developed circuits using tubes and were hesitent to go off on
> something new and different.
The vacuum tube and the bipolar transistor do indeed have very
different operating characteristics which require something a
different approach to design and servicing. The tube, for example, is
an inherently high-impedance device, whereas the transistor operated
with much lower impedances. Even the polarities of DC voltages and
currents could be off-putting. Many of earlier transistors were the
PNP type, which means that the main supply rail is NEGATIVE.
Engineers used to years of dealing with tube circuits with POSITIVE B+
supplies suddenly had to start thinking about all the supply, biasing,
etc. polarities in a circuit in reverse.
Somewhat ironically, the field-effect transistor which was developed
some years later actually has characteristics which more closely match
those of tubes in many ways.
> Not all circuits were convertable to transistors, especially back
> then. I understand to this day electronic guitar amplifier still use
> tubes.
> It is not surprising that TV equipment still contained many tubes. It
> would do so for a number of years.
In the early days, transistors were suitable for low power
applications at relatively low frequencies. They could handle neither
high frequencies nor high powers effectively.
It's easy for people nowadays to think of the transistor as heralding
the space age, for example, and without the low-power consumption and
miniaturization of transistors, some of the early satellite projects
such as Telstar would have been impossible. Yet at the same time,
those projects could not possibly have worked without the continued
use of thermionic devices, such as the TWT (Traveling Wave Tube) used
to provide the high power signals for transmission at the extremely
high frequencies involved. Transistors at that time were simply not
capable of providing the power needed at those frequencies.
In domestic equipment, TV sets employed tubes well into the 1970s, and
even the early 1980s in some cases. Hybrid sets gradually became quite
common from the 1960s onward, with transistors being used in the small
signal stages (I.F. amplifiers, AGC circuits, sync separator, etc.)
and tubes for the parts of the circuit which demanded higher power
(line output, frame output, final video amplifier, etc.) or, in the
early days, which had to handle high frequencies (e.g. the UHF tuner).
And of course, the thermionic tube never really went away completely
even after all those high-power stages became transistorized, for the
cathode-ray tube itself is just a special type of vacuum tube.
-Paul
Paul Coxwell (paulcoxwell@tiscali.co.uk)