awylie@pyr1.cs.ucl.ac.uk (01/07/88)
Hello,
I would like to tell you about an interesting but now little-known
semiconductor device, the point-contact transistor (PCT). Most of you should
know that the transistor was invented at Bell Labs. In fact, the PCT was
stumbled upon by Bardeen and Brattain in late 1947 and announced by Bell
Labs in mid 1948. The junction transistor (JT) for which Shockley was
searching was not fabricated until 1951. Early JTs had a very poor
frequency response, but this was overcome when techniques were developed to
create a narrow base region, and the PCT rapidly fell out of favour and was
discontinued.
For those of you who are not familiar with it, the PCT was normally
made using a die of N-type germanium to which was attached an ohmic contact
to make the transistor base. Two sharply-pointed phosphor-bronze wires were
pressed into the opposite side of the die a short distance apart, these
constituted the collector and emitter. A high current was passed through
these wires for a short period in order to fuse the points onto the
germanium. This process was called "forming" and at the time it involved
more magic than physics. Later theory suggested that forming created P-type
regions at the point contacts by diffusion of phosphorous out of the
wires.
Many introductory electronics books mention the PCT in one sentence
and give the impression that it was a kind of inferior JT which was purely
a historical aberration. This is incorrect, the PCT was a very different
device indeed from what we now call a "transistor", and its operation was
never fully understood, although elaborate theories were thought up to
explain it.
If you obtain an older electronics book which devotes a few paragraphs
to the PCT, it will surely mention the one fact which clearly distinguishes
it from the JT - the PCT has an alpha which is greater than one, and can
exceed 10 under certain conditions! It requires only a moment's thought to
conclude that the base current in a PCT flows in the opposite direction to
that in a JT. I have never seen this obvious fact mentioned in print. In a
PNP JT an electron current flows into the base and out of the emitter, it
is typically 1% of the collector current. In an N-type PCT under typical
conditions 50% of the electron current flowing into the collector flows OUT
of the base contact.
This tells us that the physics of the PCT are very different from
those of the JT. The device may not seem very attractive as an amplifier
since all three electrodes carry currents of similar magnitude, but large
power gains are possible because of the low input impedance in common-base.
In addition, apart from its development earlier than the JT, it possesses
two further advantages which were responsible for its survival for a number
of years. Firstly, the points can be placed very close together, about 1
thousandth of an inch apart, which results in a narrow base region and
an extended high frequency cutoff, 100 times higher than that of early JTs.
Secondly, again revealing its different physics, the collector and emitter
characteristics can exhibit negative resistance. This is ideal for making
oscillators and switches, for example a flipflop can be made using one
transistor. PCTs were employed in early telephone equipment and in
computers, as well as in linear applications such as radios and hearing
aids, to which they were less well suited.
The first PCTs were made by Bell Labs, but they licensed their
technology to all comers at a symposium in 1951. Subsequently at least 14
American companies manufactured PCTs, and 3 British ones. Over the next few
years several million PCTs were produced, most of which were sold to the
US military. Early types were unreliable and their characteristics varied
wildly, but later types were much more stable. These were extremely robust,
and could withstand thousands of g's of acceleration. The notion that a PCT
could be destroyed by dropping it on the floor is totally wrong!
Now comes the hard sell - obviously I am very interested in PCTs, and
I would like to obtain specimens of any which have survived. So many were
produced that thousands must still be lying around in junk boxes and the
like. I have a few British types but American ones seem very hard to find.
I am also keen to obtain early junction types, and early data sheets
or data books. Other semiconductor oddities also interest me - did you know
that several types of junction tetrode transistors existed in 1956? And
that even pentode transistors were made? If you have anything of possible
interest to me, please send a mail. I would be happy to pay postage
and a nominal price for any material which does interest me, but I am not a
millionaire and you will not get rich by selling me a few old transistors!
Even if you have devices which you do not want to part with, or you have an
interesting story about transistors, or you would just like to know more,
do contact me. I have a list of most of the transistor types which I want to
obtain and I will mail it on request. Below is a condensed version which
covers only point-contact transistors: it has taken 10 years of research to
make this list - information about PCTs is hard to come by!
I hope to hear from you soon,
Andrew.
Please give a copy of this article to anyone you know who might be interested!
Point-Contact Transistors - any specimens or data sheets wanted
***************************************************************
CBS-Hytron PT-2A, PT-2S
Clevite 2N50 to 2N53
GE USA G11, G11A, 2N30
Hydro-Aire A-0 to A-3, S-0 to S-2
LCT 3698, 3768
National Union T18A, T18B
Radio Receptor R1698, R1729, R1734
Raytheon CK716
RCA 2N32, 2N33
Sprague Electric 5A, 2N159
Sylvania 2N32
Texas Instruments 102, 103
Transistor Products 2A to 2H, 2L, 2N32, 2N33, 2N50 to 2N53
Western Electric 'A' series, M1689, M1729, M1734, M1768 etc.,
" " 2N21A, 2N21 to 2N26, 2N67, 2N110
Westinghouse WX-3347
GEC UK EW51, EW61, GET1, GET2
Mullard UK OC50, OC51
STC UK LS737, 3X/100N, 3X/101N, TP1, TP2
------------------------------------------------------------
Andrew Wylie, University of London Computer Centre,
20 Guilford Street, London WC1N 1DZ, England.
JANET: andrew@uk.ac.ulcc.ncdlab
UUCP: ..!mcvax!ukc!cs.ucl.ac.uk!awylie
ARPA: awylie@cs.ucl.ac.uk
BITNET: andrew%uk.ac.ulcc.ncdlab@ac.uk
If replying via satnet please give a UUCP path from mcvax to
your site.