radiomuseum.org

 
Please click your language flag. Bitte Sprachflagge klicken.

Transistors - early research - first applications

Moderators:
Ernst Erb Otmar Jung 
 
Please click the blue info button to read more about this page.
Papers » Technical History: Beginning up to 1909 » Transistors - early research - first applications
           
Vincent de Franco
Vincent de Franco
Editor
D  Articles: 183
Schem.: 0
Pict.: 119
18.Feb.08 21:12

Count of Thanks: 42
Reply  |  You aren't logged in. (Guest)   1 [Note: This text has been translated from the original publication of Ernest Erb in German. The original document can be found HERE.]

Transistors - the way to it - early research - other possibilities - first applications

Transistors and its derivate integrated circuits are still after 60 years at the front of the technology, therefore I had written in 2004 a contribution to the semiconductors, from which I am referring below.

At least during the 19th Century, someone gave some thoughts to the structure of crystals. For the German linguistic area, Ferdinand Braun with its systematic studies is well-known. He published them in 1874. Someone also found out that connections Steel-Carborundum which would let the current flow under certain circumstances only into one direction. Someone found thermal techniques to achieve a practical use of semiconductors as detectors. Starting from 1904 someone could obtain the same effect with an Electron Tube of Fleming.

Still before the triode of Lee de Forest someone found a replacement for the coherers or the “Magical” possibilities with different kinds of crystals and metal connections, but wanted to be able to build something for telephony and for wireless transmissions amplifiers. All these mentioned possibilities remained up to the penetration of the triode in use, thereafter also the crystal detectors for particularly cheap receivers.


Amplification for telephony and radio

For telephone amplification, the C. G. Brown (England) found a “mechanical” solution for the amplification of sound signals via a kind the relay amplifier, more a repeater. The most well-known example is the “Microphone Amplifier type C”. In particular Siemens in Germany got itself a license and improved the telephone amplifier extraordinarily. This mechanical amplifiers packed in an aluminum cylinder was used for a long time for certain applications. Someone built also proper AF-stages for the operation of crystal detectors to build a radio with loudspeaker. An example is the in 1923 released active loudspeakers Brown Crystavox (amplifying speaker), that has been praised at the “Viennese autumn fair” as “tube-less active loudspeaker” - with the remark: “replaces a two-stage tube amplifier”. As power source a battery of 2-6 V was sufficient. There are still different mechanical amplifier principles such as the Frenophone described in the book “Radios von Gestern” on page 60.

The mechanical amplifier of Siemens in the aluminum cylinder caused probably Robert Denk to his hoax (swindle or joke) in the newspapers in 1948 . Only very few humans knew the Siemens relay still, who will it surprises, if they described what they had seen as a “miracle”. Miracles can only be tied up to the naivety of people who describe it as a “epoch-making German invention”. Hokus Pokus (magician) is since 1796 common as hoax (newspaper duck) a term, but only with the Internet and/or email generally in German-language countries.

The tubes dominated as amplifiers since the invention of the triode by Lee de Forest end of 1906, patent 2007 until the 50's-60's. A triode is an amplifier, which can be used for AF and HF-purposes. De Forest was at the beginning only interested in the audion as HF detector amplifier, so that Robert von Lieben secured the patent for AF-amplification at the end of of December 1910, after his unsuccessful attempts with the cathode ray relay starting from 1906. However its mercury steam tube was not successful, this being an explanation why the whole technology was inspired from the USA. By the way, many collectors mix up these two completely different developments of Lieben: the man-high cathode ray relay from 1906 to 1910 and the large mercury steam tube starting from the end of 1910 to 1914. He wanted to optimize the triode of de Forest for AF-amplification, went however the wrong way (seen from today's point of view). Nobody took over something from the tubes of von Lieben, except Telefunken who had an extremely strong market position because of the patent claim of the AF-amplification.

About the comparison Europe-USA at the beginning of 1914

Germany was still experimenting in 1914 around the mercury steam tube from von Lieben, while France owing a sample of the triode of de Forest developed a socket-based high-vacuum valve. Von Lieben had announced the triode as telephony amplifier in December 1910 with express mention of Lee de Forest. The practical efforts went however into the wrong direction and were actually unsuccessful. At the end of of 1914 Germany had likewise realized the high-vacuum technology with blade contact tubes.

AT&T in the USA developed in 1913 a telephone network going from coast to coast (3400 US miles) with the triode of Lee de Forest . However the inauguration took place with Theodore Vail, the president of AT&T only in July 1914.

Only later, Europe could recover the tube development - and only in certain areas - with the invention of the Pentode (Philips) and the oscillator mixer tube as combined Triode-Hexode and late Triode-Heptode (Telefunken).


The way to Solid State technology

In the 20's, once the structure of the atoms was better understood, the scientific study of the semiconductors began, what finally led to the transistors... mainly by Karl Lark-Horowitz at the Purdue University in the USA. Lark-Horowitz placed the name of his wife in front after his marriage. He came as Horowitz from Vienna (1892 Vienna - 1958 Purdue) and was transmission officer during the First World War. Supported by the Rockefeller Foundation, Horowitz could occupy the position of professor in Purdue in 1928. He could focus fully on the study of the germanium and other crystalline elements.

A.H. Wilson published in1931 the theoretical principles and in 1938, Walter H. Schottky publishes his Diffusions-Theorie - after which the “fast” Schottky diodes were built. Also the work of http://www.radiomuseum.org/forum/perpetuum_mobile_und_sagenhafte_erfindungen.html, who came from Austria-Hungary to study in Germany and then migrated to the USA as a Jewish emigrant, brought in 1926 what could possibly become the field-effect transistor. Oskar Heil announced a patent, in which he suggested capacitively working gate electrodes in semiconductors, towards the FET in March 1934. Before the war, he works in England on microwave tubes. You can find more information on Hilsch, Pohl (1938 “three-electrode crystal”), A.L. Gorelik, A.H. Wilson, N.F. Mott, W. Schottky, B. Davydow, S. Ohl, J.H. Scaff, C.M. Zener (Zener diode) etc. in “Radios von gestern” on page 234ff.

With the practical realization of the radar, England had succeeded in reverting the supremacy of the air to their own advantages. An invasion would have been otherwise the result. With the Second World War, it was crucial to see the opponent before he could see you, both with ships as well as with airplanes. Particularly for airplanes it was necessary to achieve as high frequencies as possible, in order to make the wide antennas smaller.

England shared therefore their know-how over radar immediately with the USA. Solid-State diodes have the task within the radar to change the signal coming back into a direct current for the display on the screen and in addition, are used a as diode mixers. They were often overloaded by the high-frequency and high (counter) current and burned; this is the explanation why the research was coordinated between Purdue, Bell Lab, MIT and the university of Chicago. Results were achieved in better and suitable crystals “breeding” methods.

By doping with traces of tin with germanium, the group around around Lark-Horowitz could reach around 1942-43 an increasing tenfold of the tension. More by coincidence, the student Seymour Benzer discovered this in Purdue, because he accomplished more than one year of attempts with germanium. From this point on, the research was focused on germanium.

Concrete steps to the transistor

The way to the transistor can be read like a detective story, but at the same time a reader can recognize that many individual theoretical and practical steps were necessary, in order to achieve the goal "transistor".

First the expensive not very stable point-contact transistor


In 1945, directly after the war, Mervin Kelly who was the boss of AT&T recognized the necessity to replace amplifier tubes, because the developments of those could not be further improved. He assumed the fact that semiconductors could bring a solution and put together a department for semiconductor research (Solid State Physics Group).

William Shockley and the chemists Stanley Morgan became responsible. Additional researchers from the Bell laboratory joined such as Walter Brattain, Robert Gibney and Gerald Pearson, and later John Bardeen. Shockley focused with his group, above all, on to the semiconductor research, continuing to drive the work from Frederick Seitz and Eugene Wigner of the 30's.

On 17 November 1947 Walter Brattain had put an experimentation arrangement for a semiconductor amplifier in the water, instead of working via vacuum, in order to prevent a certain saturation; the attempts again and again were failing. It generated thus the best amplification. The idea was to achieve not only a steered diode, but an amplification element.

Robert Gibney tried then different tensions. On 21 November John Bardeen recommended a metal contact, surrounded by distilled water, into the silicone. Brattain finally obtained a small, but measurable gain. They tried it with different other materials instead of with water. Thus a substantially larger success could be obtained, but completely differently than expected. The arrangement was not suitable even for a speech frequency. It was decided that the problem came from the liquid (the electrolyte), and a test was finally made on December 12 with germanium oxide (green). Only after Brattain pressed the gold contacts several times on the material and created recesses, a gain compatible with audio frequencies would be observed. However no oxides were present any more on the contacts.

With an arrangement of a piece of germanium and two punctiform gold contacts, placed with a plastic in a fraction of a millimeter apart, Bardeen and Brattain had realized an electronic Solid-State amplifier. Only some days later, after most diverse attempts, they gave on December 23 a demonstration to Shockley and further important people. They kept their secret more than one week, even towards Shockley.

The point-contact transistor has however important disadvantages and was used only during a short period of time particularly for military purposes. It was much too expensive in the production and also much too unstable in the application, but reached initially higher frequencies than with the junction transistor. From “Radios von gestern” page 235: “The first public demonstration took place on 30 june 1948 and R. Brown presents thereby the first transistor radio among other things.” … “J.R. Pierce gives later the name “Transistor” to the element out of Transfer reSISTOR.”

At different places, such as at Philips and in England, people were involved also intensively with the semiconductor technology. Also Herbert F. Mataré and Heinrich Welker developed with Westinghouse of Paris a similar mechanism as a transistor and patented this “Transitron” some months later than the Americans, on August 13 1948. (Link may take some minutes). They worked before at Siemens on radar technology and did research on “point-contact diode” ED705 as mixer-detector for cm-wave (also ED704, ED707). The Mataré-Welker mechanism led to telephone amplifiers between France and Algeria in 1950. Mataré creates in 1952 (with US capital of new England Industries) the Intermetal. The “Current Biography Internationally Yearbook 2003" records Mataré as Swiss.

Only with an important further step the junction transistor develops

It was clear for everyone after the demonstration that Shockley should not have a portion of this invention, because he was not directly involved in it. Bell refused even that the name of Shockley should appear on the patent. Shockley spent some days - disappointed and sad - alone, because finally it was actually his idea. Thus he started to research on fundamental improvements: he wanted to build a semiconductor with three layers - with the middle layer acting like a grid of a vacuum tube to control the flowing current. During the time his group improved the primarily discovered product, he worked alone on a completely different solution. Actually it was more and more clear for everyone that the point-contact transistor produced too much noise and was not stable.

On January 23, William Shockley decided to dope both outside layers with too much and the internal layer with too few electrons (N-type). He wanted to feed the middle layer with a control voltage, in order to control the way of the electrons would flow between the outside layers. Doing so, the electrons run inside the material, with the previous structure along the oxide outside. On February 18, Shockley could see the fact that the arrangement would work: Joseph Becker and John Shive with another experiment could let the current flow in the semiconductor and Richard Haynes could finally validate this with exact measured results. Heynes discovered also that the middle part must be particularly thin and pure. Now Shockley made its findings public, which annoyed again Brattain and Bardeen.

Gordon Teal found the best way, which required to breed the crystals in one piece instead of putting 3 crystals together to realize the 3 layers. In March 1949, Shockley also accepted this way since with this method the system functioned more than hundred times longer.

On June 23 1949 the invention of Shockley could be presented to the military, but since they did not wanted to move forward and hold this the invention as military, it as been possible to make it public after one week. It is only on April 12 1950 that it has been possible to breed crystals in sandwich structure, already as N-type or P-type. It was still difficult to amplify language and it was recognized that the middle part would have to be thinner. On January 1951, Morgan Sparks found also the way to it. The internal layer was then thinner than a sheet of paper. On July 1951, Bell could finally present a practicable transistor to the press.

Naturally it was clear for Bell Lab that it was not sufficient to cannibalize this invention now. Therefore Bell Lab organized in September 1951 a symposium, in order to show what can be done with the transistor. The license of the rights for the production of transistors were provided to 26 companies. Among them large ones like IBM, General Electric and Siemens; in addition also small companies like at that time nearly unknown Texas Instruments (TI). For these licensees a symposium was organized in April 1952. Bell Lab united all experiences from this symposium in the book “Transistor Technology”. At least during 1952 (monthly 6,000, GE = 1,000) and 1953, Raytheon had the largest numbers of transistors produced.

As it can be read in CQ DL 12/97, Philips in Eindhoven was able to produce transistors one week after the Bell discovery, like also the British GEC after a further week. Therefore it is to be considered that there as also in France - due to the experiences during the war over crystal diode mixers - the research continued. I know the details however only for France, because there were attempts and realizations to improve the crystal diode during decades. After the realization of the transistor there was rapid progress up to the MOSFET of 1960 and further special types.

Kinds of transistors

1947 saw the Point-Contact Transistor from Bardeen and Brattain. It was always afflicted with problems, could however master higher frequencies. But soon it was replaced by the Junction Transistor of Shockley (February 1948). Both were germanium transistors until it was possible to do the same arrangement with silicon (silicium). Initially the layers were brought on top of each other approximately or around, which was called Planar Transistors or when diffused Diffusion Planar Transistors. Afterwards the Mesa Technology came (Tafelberg/Tisch).

Above all, there were important variants of semiconductor constructions - up to high-complex IC's. It would blow up this framework, but as examples here the most important advancements: Uni-Junction Transistor, Overlay-Transistor, Photo Transistor, Thyristor, Triac, Darlington-Transistor etc. Special developments such as the the inter-digital transistor arrangement, e.g. 256 single transistors in one fingers-big structure resulted in Overlay-Types, e.g. for 4W with 400 MHz etc.

For different applications there is the very important Field-Effect-Transistor (FET), which behaves “tube like”; it is available as V-FET (vertical FET), as JFET (Barrier Layer and/or Junction FET) and finally as MOS-FET. It can be found as N channel and P channel types, as well as with both Enrichment and Depletion types. The first type closes without signal, the latter leads without signal.

It is obvious that an uncontrolled growth of designations would have developed to probably more than hundred thousand different kinds and values of semiconductor components.

After the first years it was agreed in the USA and Japan over unified designation pattern: Joint Electron Device Engineering Council (JEDEC), Japanese Industrial Standard (JIS). In addition there is the pattern “pro electron” and numerous proprietary designations, where companies did not want to reveal the kind of the transistor, so that it was necessary to order them exclusively. It is possible to find the designations in comparison books - or also here.

Integrated circuits

The evolution did not stop with the transistor. Texas Instruments did patent the first Solid-State IC in February 1959 (Jack Kilby), in October the first Germanium IC was finished, in December 1963 came the first hearing aid with IC on the market and about at the same time the radio IC2000 from Philips. Gilbert Hyatt of Micro computer Co patented in 1971 the microprocessor and Ted Hope from Intel presented in February the 4-bit-Prozessor 4004 with nevertheless 2300 components. In 1969 the Arpanet, the predecessor of Internet, was started. Thus we see inconceivable applications coming within 20 years. Also Windows and the mouse were there, not far away: a development of Xerox including Ethernet Boards for the network in 1974. In 1975 it was possible to buy the Altair, the first “personnel computers” - as kit and Steve Jobs and Steve Wozniak brought the first Apple a year after. IBM came only in 1981 (in the USA, Europe 1982) with the “Personnel Computer” on the market.


Early practical applications with transistors

The first goal was clear: develop much sturdier and better telephone amplifiers. This was realized in France with the “Transitron” for the first time practically. Bell Telephone System, came in October 1952 with point-contact transistors to a transistorized telephone system (Englewood, New Jersey).

It was clear also that it would be possible to build better computers with the transistor. Tubes have a shorter life span and a much larger current consumption, thus high heat emission. IBM brought the first transistorized computers with the model 604 in 1953 and with the model 608 in 1957. At this time the tube costs comparison to the transistor was 8 $ to 75 cents.

In accordance with Hannon Yourke, the developer of the ECL technology (Emitter Coupled Logic) at IBM, IBM used the new technology in 1959 in its 7090 and an evolution in 1961. At least the legendary IBM 1401 of 1959 had a kind of IC with some transistors per IC when the IBM 608 featured 6 transistors mounted per board. The transistor was not yet trusted completely.

Hearing aids

It was particularly interesting to use transistors to design hearing aids which became immediately substantially smaller and with lower operating cost. Zenith came (based on the information found in its business report of October 1953) with the transistorized, tubeless hearing aid Royal T to the market. Before it was probably Maico with “Transist Ear” (March 1953) with 3x CK718 transistors from Raytheon. The first hybrid came at the end of of 1952 already from Sonotone for 229.50 $ on the market. The transistor was built by “Germanium Products Corporation”.

Generally the licensees however still looked for applications for mass-produced goods. Particularly for the small Texas Instruments, for which the royalties of 25 000 dollar - to the value at that time - were an expenditure which should bring return on investment. In principle the transistor could replace the tubes, but not for higher frequencies. Besides this, the transistor was still very expensive - and it was required to use often more transistors than tubes for the same task. There are many details about the first transistors applications and development here in English.

Transistor receivers

The radio was for some decades the most effective public-product with transistors - and remains also today, if we include the mobile phones using this technology. In mobile phones (and PC) you can find millions of transistors.

Prototypes make the beginning

It is “Natural” that an experimental radio with transistors was demonstrated in 1948. Other developers started from the end of 1949, because an article of the January 1950 edition of “Radio and Television News (magazines)” showed a Crystal Receiver with Transistor Amplifier” (R. Turner); it had three CK703 (audion, 2 AF-stages). Also in May and June, there were other publications and the CK703 was available at a market price of 18 $. Nevertheless Wikipedia writes: “Mataré created 1951/1952 the company Intermetall in Düsseldorf, that offered the first diodes and transistors worldwide - only de first statement is correct...

The developers of Bell lab gave in 1950 to the well-known caricaturist Chester Gould a sample of an “Arm-Band Transistor Radio” built in the spare time, just to show around what could be possible. And other companie, Texas Instruments showed in 1952 the first samples of transistors radio, and the electronics booklets in 1953 were filled with building proposals with the cheaper CK722 - e.g. February 1953 in “Radio Electronics Magazine”. In Germany, Intermetal shows in 1953 at the Düsseldorfer Radio Fair a sample radio without housings - which is to be found somewhere today on the Internet under “The first transistor radio of the world"...

First transistors radio in Serial Production

There was a clear outsider, Regency (USA). As next we can mention Raytheon 8TP and/or 8TP1 etc. and a small company, which invested fully on the new possibility: Sony (Japan). For portable radios, expensive transistors could still bring great advantages, because transistors are much smaller and above all, tubes a need large amount of energy to heat the cathode.

The licensees looked for mass-applications, therefore projects and samples for radios were already developed early - some came to production. TI had found a small company, which was ready to realize a transistor radio with TI-transistors. However Regency insisted after the building of some samples on its own structural drawings. The model came on October 18 1954 as Regency TR-1 (TR1) on the market. It uses four germanium transistors and measures 12.5 cm (5 x 3 x 1.25 inches). A 22,5 Volt Battery allows more than 20 hours of music, while more expensive batteries provided only some hours of music from a tube radio. TI had its public hanger and concentrated on transistorized pocket calculators.

Regency built transistors radio until 1960 (TR-99). From the TR-1 (TR1), it produced nevertheless more than 100,000 units. The reception and audio characteristics were much worse than the February 1955 Raytheon 8TP-1 for 80 $ with eight transistors. Soon Zenith with Royal 500, Emerson with the 842, RCA with 7-BT-9J and 7-BT-10K as well as General Electric with model 675 followed. These small devices allowed the young Rock & Roll generation to become independent from the radio of their parents.

A very small tape recorder manufacturer in Japan, Tokyo Tsushin Kogyo, founded by Ibuka and Akio Morita worked not many later than Regency on a transistor radio. Since they thought particularly at the American market, they looked for an easy name for this market. At that time, young successful Americans were called "Sonny Boy” and the “artist” name SONY was found in 1958. The company founded in 1946 was still “Tokyo Telecommunications Engineering Corporation”, which we put together in our database under Sony. Sony TR-55 (TR55) came for $ 49,95 into 7 different colors in April 1955 on the market, as well as additional series with special colors. The TR-55 came actually only on the Japanese market and its successor, the TR-72 (TR72) particularly into Canada (Gendis). The third model, Sony TR-6 (TR6) used the Mitsumi variable capacitor for the first time.

Only from March 1957 on- there is a really small device sold in higher quantity, the TR-63 (TR63) with again a smaller successor, the TR-610 (TR610). The TR-610 sold itself almost one half million times. You know the further history of SONY and those of the transistor devices; it took around 10 more years to the home radios to be equipped more and more with transistors and finally also for the TV sets, in which now gradually also the picture “tube” disappears. Note that there were also different early transistor radio manufacturers.


PS: If you find translation errors, please do not hesitate to contact me preferably per email, and I will make the necessary corrections.

This article was edited 18.Feb.08 22:08 by Vincent de Franco .

Ernst Erb
Ernst Erb
Officer
CH  Articles: 5486
Schem.: 13673
Pict.: 30634
20.Jan.09 09:40

Count of Thanks: 63
Reply  |  You aren't logged in. (Guest)   2

With following posts we can add here the first transistor radios - if possible including date - for the different manufacturers / brands in the USA - and firsts from other countries.

We remember the very first from October 1954 was Regency TR-1, followed by Raytheon 8TP-1 in February 1955. Then came the Sony TR-55 (Japan) in April 1955, followed by Zenith Royal 500, Emerson 842, RCA 7-BT-9J and 7-BT-10K, General Electric 675 and 676 (colors). This seems to be the lot for 1954 and 1955. No, according to the Philco Annual Report 1955, Philco claims: "The Philco "Surface-Barrier" transistor made possible the development and production of the first completely transistorized auto radio." It is Mopar 914-HR with chassis C-5690HR.

1956 will not be so easy and I can only just add what I know to this list - without Hybrids:
USA:
Philco T-7-126 (T7)
Motorola 56T1


Gret Britain (UK):
Pam Model 710 (made by Pye, on the market in June 1956)

 

This article was edited 07.Mar.13 18:57 by Ernst Erb .

Vincent de Franco
Vincent de Franco
Editor
D  Articles: 183
Schem.: 0
Pict.: 119
29.Jan.09 23:02

Count of Thanks: 73
Reply  |  You aren't logged in. (Guest)   3

The first commercial transistor radio from France came in 1956 and was called the "Solistor".

More information on the birth of the French transistor industry can be found HERE.

The posts are in French language, but if you would like some of it to be translated in English, just drop me an email.

 

Joe Sousa
Joe Sousa
Editor
USA  Articles: 639
Schem.: 185
Pict.: 390
07.Feb.09 07:14

Count of Thanks: 37
Reply  |  You aren't logged in. (Guest)   4

I submit the following first in Transistors:

An article published in the October of 1948 issue of Radio and TV News magazine in the USA,  is probably the first transistor construction article. In this case, you don't just build a circuit with ready-made transistors; instead, the article shows how to make Point Contact Transistors from 1N34 germanium Point Contact diodes. There are also measured results for gain, and orientation is given on proper bias levels.

This link contains the title of the article and has complete scans of the article

A_crystal_that_amplifies_Radio_News_October_1948

The author was a Tung-Sol engineer named C. E. Atkins. He did the work shown in the article as a home hobby project and not for his employer.

Best Regards, and thanks for the good overview of transistor history.

-Joe Sousa

Ernst Erb
Ernst Erb
Officer
CH  Articles: 5486
Schem.: 13673
Pict.: 30634
11.Apr.15 09:07

Count of Thanks: 21
Reply  |  You aren't logged in. (Guest)   5

Mark Burgess and Wolfgang Gebert have published an outstanding article about Herbert Mataré and Intermetall and about what was achieved - in detail. See: 
History of Intermetall Semiconductors
 

Michael Watterson
 
Editor
IRL  Articles: 825
Schem.: 625
Pict.: 2384
31.Aug.16 18:39

Count of Thanks: 7
Reply  |  You aren't logged in. (Guest)   6

Joe's earlier link is broken, this one works (August 2016)

This link contains the title of the article and has complete scans of the article

A_crystal_that_amplifies_Radio_News_October_1948

The author was a Tung-Sol engineer named C. E. Atkins. He did the work shown in the article as a home hobby project and not for his employer.

 

  
rmXorg