History of AWV Semiconductors

Copyright: Mark P D Burgess 2009

Genesis

Australian Wireless Ltd (AWL) was set up in 1909 by Australian investors and Telefunken. In 1912 an agreement between two wireless giants, Telefunken and Marconi settled litigation over patent rights between them in all territories except the USA. The resolution led to a merger of the Australian interests of Telefunken and Marconi and Amalgamated Wireless Australasia Ltd (AWA) was formed in 1913. The new company obtained rights to present and future patents of Marconi and Telefunken. [Given 2008]

AWA produced its first tube, the Expanse B, in 1920 in a small plant at Clarence Street, Sydney. The

Acknowledgements

The author gratefully acknowledges the assistance of Mr John McIlwaine who worked for AWA from 1954 and now maintains and extensive archive of AWA publications and company information. He provided all of the AWV references cited and further information and advice.

John encouraged Ted Watt, one of the AWV engineers that were responsible for developing its semiconductor production lines, to record his recollections of that time. This work has been an important source of information for this article and is greatly appreciated.

For further information please refer to the following links:

Manufacturing the AWV Transistor, a detailed description of 15 key manufacturing steps in producing AWV transistors in 1958.

The Transistor Story a history of the production of transistors at AWV by Ted Watts

Schematic of the Radiola Transistor 7 897P courtesy John McIlwaine

More information sought:

I would be grateful to have any further information, that would expand on this short history including relevant early advertising, pictures of early transistors by AWV and any magazine articles about the progress of AWV in the field. Please email me at markpdburgess@gmail.com

“Expanse” brand reflected AWA’s Marconi heritage: this was a cable code used by Marconi. The plant employed eight people: two in research and development and six in production. Between 1920 and 1923 the plant produced up to 4000 tubes pa.

In 1923 a new plant was set up at Knox St Sydney and in 1924 the company negotiated a license to RCA tube technology beginning a long association with the USA company that survived into the semiconductor era.

By 1927 six battery tube types were being produced while 32 tubes were being imported for resale.

In 1931 a decision was made to expand tube production and an agreement was made between AWA, RCA, International General Electric (IGE) and Westinghouse to form the Amalgamated Wireless Valve Company.

The first of the new tubes, the type 80 rectifier, carried the RCA “Radiotron” brand name. Over the next seven years the Radiotron range was expanded to 30 receiving types and by 1939 AWV was supplying most of Australia’s receiving tubes. Production reached one million pa.

In 1938 transmitting tubes were added to the range and during World War II production was extended to magnetrons, klystrons and TR switches for the Navy. To meet the huge war time demand for tubes a new plant of 65,000 sq ft was built at Ashfield in 1942.

In 1956 the company needed to expand again to meet the demand for the new peace time product: television. A site at Rydalmere was selected and a 95,000 sq ft plant built to produce picture tubes, in the first instance. By 1957 all of AWV operations were consolidated at this site. [Stokes 1982; Mingay 1963]

First AWA Transistor Radio

AWA’s first transistor radio was the Transistor 7 (897P) designed by Graham Moore. It was released in November 1957 selling for 44 Guineas. It was virtually identical in appearance to the AWA 581PZ tube portable and followed conventional tube receiver construction methods. [Picture courtesy Mike Melzer]

The transistor line up utilized RCA and Telefunken transistors:

2N219 Oscillator mixer

2N218 IF stages (two)

OC602 Transformer coupled AF (two)

OC604 Class B Output

The detector was a GEC GEX34 diode. The most unusual aspect of the design was use of a transformer coupled AF preamplifier stage where RC coupling would be conventional. While printed circuits were in common use for compact chassis, tag mounted components were employed typical of larger sets of this vintage.

Esquire's Monday Column editorial for the Farmer's department store published in the Sydney Morning Herald for 9th December 1957 suggested that this set was the perfect accompaniment for summer cruising on the harbour: "And apart from the steady beat of the engine, the only sound will come from his brand new Transistor Portable."

"With the weather so chancy, it is handy to have those regular forecasts coming in over a portable radio that doesn't weigh as much as a sizeable kingfish. And what's more, no fragile little valves inside to pop off in protest against the tossing and pitching and vibration. No valves? That's right - a portable radio with no valves at all!

Transistors are the Sputnik Age secret that make this valveless miracle possible."

"Handsome as well as efficient in its gunmetal leather case. The first release of the first transistor portable radio made in Australia, forty-four guineas on the Fabulous Fifth Floor. The ideal gift from the family for the outdoor man."

Media advertising from February 26^th edition of The Age is shown above. [Scan courtesy of Mike Mezler] The claim “lightweight portable” is clearly relative to a tube set but at 7 lbs this was a transistor heavy weight due to the use of full sized components and chassis based construction. Below: Illustration from Esquire's Sydney Morning Herald editorial, 9th December 1957 showing the Sydney outdoor man on a palm fringed tropical island...with the perfect accompaniment. [Scan courtesy Ken Rowlinson]

It appears generally (but not universally) accepted amongst Australian collectors that this was the first set designed and built in Australia. A contender, the Pye P123BQ may have come out months earlier but was a UK design.

Later models, the 897PZ, 897PY and 897PX, used the same RF and IF lineup but switched to RCA types for the audio stages. 2N408 were used in the pre-amplifier and driver stages and 2N270 in the Class B output.

The use of Telefunken transistors in the first model is an interesting echo of the role of Telefunken in the founding of AWA. The reasons may have been more prosaic: transistors were in short supply and manufacturers had to obtain them where they could, an issue that was also problematic for USA producers.

Thus 1957 was the year of the first portables in Australia. In this regard Australia was in the middle of the pack. Worldwide, the first transistor radio was the Regency TR1 released late in 1954 in the USA. The year 1955 saw the introduction of only six new models in the USA but by 1957 there were at least 30 models on the market. {More}

Introduction of Semiconductors

While it is not clear when AWV first began to re-sell transistors in Australia, it did launch a substantial campaign to introduce the RCA range of transistors in September-November 1957 by dedicating the three editions of its Radiotronics magazine over that period entirely to transistors. [AWV 1957] By this time the company had committed to producing transistors at Rydalmere which it did the following year.

The October Radiotronics served as an introduction to transistor theory setting the context for AWV’s new initiative:

“A relatively short time has elapsed since the first announcement in 1948 of the work on semiconductors by J Bardeen & W H Brattain and by W Shockley in which they describe a new version of the crystal valve – the semiconductor amplifier triode or “transistor.”

Before the development of the transistor, the crystal valve family comprised only one member- the semiconductor diode. As happened in the case of the thermionic valve, the addition of a control electrode converted the diode to a triode, made signal amplification possible and added the second member to the family of crystal valves.”

The last of the three editions discussed the design of transistor audio amplifiers. The introduction to the publication reflected the ambivalence of tube engineers and tube companies to the transistor and its underdevelopment in the 1950s. “Since the first transistors appeared on the commercial market, a tremendous volume has been written covering the manufacture and use of this interesting little device. It is the intention of this article to deal briefly with some of the things that an experimenter will come across while becoming acquainted with his new toy.” Tube companies were not the survivors in the development of semiconductors. While they dominated early on due to the power of their research and development and the risks taken by pioneering entrepreneurs in their research and production laboratories, they were displaced by the 1960s by innovative new start-ups; famously Silicon Valley companies.

The preface to the October edition also emphasized that the transistor was a “comparatively recent addition to the family of crystal valves.” While this is true much had already been achieved and the article failed to acknowledge the rapid acceleration of transistor technologies and applications after the first five years’ of development by many producers (1948-1953). Examples include AWA’s own first transistor radio:

1953 Widespread adoption in the USA of junction transistors for hearing aids

1954 First transistor pocket portable, the Regency TR1 by Texas Instruments. Raytheon produces its millionth transistor and Texas Instruments announce their silicon transistor.

1955 Quality transistor portables launched by five new producers: Emerson, Raytheon, Zenith, RCA and General Electric.

1957 At least 30 portable transistor radios in production in the USA alone. First AWA transistor Radio the Transistor Seven.

The September edition included extensive data on existing and planned releases by RCA.

The AWV publication was ahead of its time with respect to some transistors still subject to RCA final specifications noting that “for some types technical data is not yet available.” For example, the 2N409-2N412 intended for RF, Oscillator, Mixer and IF duties in a transistor broadcast band receiver were not finalized and registered by RCA until November of that year. [JETEC 1957]. Similarly the well known range of RCA drift transistors were described despite their first date of registration being even later, in October 1958. [JETEC 1958]

Semiconductor Production by AWV

AWV obtained assistance from CSIRO scientists and RCA under its license to AWV to set up semiconductor production.

Pioneering Role of CSIRO

The Radiophysics Division of CSIRO began research in semiconductors in 1953 initially with the assistance of Bell Laboratories. The work was led by Dr Louis Davies and began with developing technologies for the purification of germanium and silicon. Davies spent six weeks at Bell Laboratories and recalls: “I came back from the US armed with two essential precursors to making transistors: the technologies for purifying germanium and for growing single crystals of germanium. Brian Cooper and I set up a section on transistor physics and devices, in which I was responsible for purifying the materials, making the transistors and trying to develop physical interests in that material, and Brian was responsible for the testing of the transistors and for designing and building devices which would use them. Through that, CSIRO made a very positive contribution to industry in Australia.” [Craig 1999]

The symposium attracted the attention of industry including AWV: “we later had visits from individuals from each of four companies in Australia – AWA, STC, Philips Australia and Ducon – who worked with us and absorbed some of the day-to-day problems of working in the transistor field.”

Technology Transfer

AWV were licensed by RCA to produce transistors to their specifications. Although the transistor was invented by Bell Laboratories, RCA through its cross licensing arrangement with Bell, could make the original Bell rights and the RCA alloy junction transistor technology available to other producers. [Braun 1982] RCA had such relationships with many transistor producers both in the USA and Europe.

When AWV decided to manufacture transistors it needed to address two issues: a state of the art manufacturing facility and the knowhow from RCA to operate it.

“It was obvious that transistor manufacture would soon be introduced. Accordingly, plans were made to extend the Rydalmere works. Engineers were also sent overseas to study transistor manufacturing. [Mingay 1963]

The engineers were Ted Watt and Henry Banks who left Australia in December 1957 for the new RCA plant at Somerville where RCA had consolidated its semiconductors production since 1956 [Choi 2007] Watt had worked for AWA from 1941 and then AWV as a development engineer. His role was to cover manufacturing equipment, processes and know-how while Banks covered engineering and test equipment.

Back in Australia and while the new Rydalmere plant was being built production experience was gained from assembling 2N109s from imported components (germanium pellets, indium dots, the base tab, stem and can). [Watt 2009]

The Rydalmere Plant

Plant equipment was ordered from RCA suppliers in the USA or designed and built by AWV. By the end of 1958 two conveyor furnaces were installed. Ultimately four were in use. [Watt 2009]

The first extension to Rydalmere was completed in 1958 adding 25,000 sq ft. The transistor department was installed on very modern lines being well lighted and air conditioned. It was an area requiring clinical cleanliness and with all operatives dressed in white nylon uniforms.” [Mingay 1963]

The new plant was opened in 1958. To celebrate the occasion AWV produced a brochure entitled Manufacturing the AWV Transistor designed to emphasis the quality of the Australian made semiconductors. “The stringent controls employed in the AWV factory fall into two main types. Firstly, there is careful examination and testing of all raw materials and components to make sure that no material or parts which are not of the required quality reach the assembly area. Secondly, tight controls and manufacturing limits are imposed on the manufacturing operation, coupled with close inspection and testing at every stage.” The plant was air-conditioned and maintained “almost clinical cleanliness” in order to assure quality. [Transistors courtesy of John McIlwaine]

AWV Junction Transistor Production

The AWV brochure described 15 key steps in the production of alloy junction transistors.

1) Grown germanium crystals were imported and checked for purity by resistance testing

2) The crystal was cut into 10-12 mil slices with a diamond saw

3) The slices were lapped down to 7-9 mil thickness and fine surface finish

4) The lapped slices were etched to 3.5-6.0 mil thickness and high surface luster

5) The slices were cut with a diamond tipped scribing tool into small square or rectangular pellets of the desired size for each transistor

6) A roller micrometer was used to sort the pellets by thickness (typically 2 mils)

7) The key components for each transistor, the base tab, pellet and the collector and emitter dots were assembled in a graphite jig and passed through a conveyor furnace to fuse the dots to create the PNP or NPN structure.

8) The base tab assembly was spot welded to the stem or transistor base and two nickel wires 5 mil diameter spot welded to the stem collector or emitter connections.

9) The collector and emitter lead wires were connected to the collector and emitter fused dots.

10) The assembly was acid etched to clean it.

11) The active regions of the transistor were encapsulated in silicon and oven cured

12) The can for each transistor was filled with a viscous potting compound

13) The stem was hermetically sealed into the can in a controlled atmosphere within a pressurised cabinet.

14) Final testing was carried out and production transistors were sorted into family groups each conforming to the specifications for each commercial type.

15) Once passed the type number, production code and brand were printed on the can.

[AWV 1958]

Assembly was undertaken by female staff and was very labour intensive.

The complete sales range was not made in Australia. Ted Watts recalls the principal types produced:

2N109 family in large volumes

2N218 and 2N219 in large volumes

2N270 Easier to make due to its larger size

2N301 Power transistor. Easy to make in high yield

2N247 Drift transistor. Difficult to make

Semiconductor Research

Having pioneered semiconductor research at CSIRO, Louis Davies returned to Bell Laboratories under a Harkness Fellowship. He joined a group at Murray Hill working on hot electrons in silicon and theoretical aspects of zone refining of germanium. After returning to Australia and publishing the Bell Labs work there was no role remaining at CSIRO in semiconductors research. AWA, however, were keen to have him. Sir Lionel Hooke, the Chairman of AWA, had been making overtures for two years and in 1960 Davies became Chief Physicist, AWA Research Laboratory in the same building as the Amalgamated Wireless Valve Co at Rydalmere.

Sir Lionel told Davies, “Please do some research in semiconductors.” Despite Davies’ experience in purification and crystal production AWA never did this. “From our work at CSIRO I knew how to do it, so we were able to make an informed decision not to do it. If you do everything yourself, very often you end up becoming commercially unattractive and losing money.” Soon Davies was appointed to Chief Scientist of AWA at the AWA Research Laboratory. [Craig 1999]

Integrated Circuits

Graham Rigby was Senior Design Engineer, AWA Microelectronics 1968-78 and its Group Manager 1978-82 when he left for an academic position at the University of New South Wales. He recalls the role of Lou Davies hired by AWV to do research into semiconductors: At the time Lou joined AWA, the first integrated circuit had just been demonstrated and was to appear in a more practical form in 1961. Within four years of that event, the company set up an experimental integrated circuit manufacturing facility next-door to the Physical Laboratory, drawing again on their knowledge. In 1967, the first working integrated circuit was produced and the facility commenced commercial operations. Fifteen years later, as a result of management changes in the early 1980s, Lou was appointed General Manager of AWA Microelectronics. [Rigby 2003]

In the November/December 1967 edition of Around AWA the company announced it had developed and manufactured its first integrated circuit at its Advanced Devices Development Laboratory at Rydalmere. This was “a broadband, two stage voltage amplifier containing the functions of four transistors and nine resistors, all of which have been diffused into the tiny silicon square” illustrated in the publication between the point of a pair of tweezers.

In 1996 Semiconductor Inc. purchased AWA Microelectronics Pty Ltd. The assets included the AWA wafer foundry business and product design center.

References

AWV 1957 Transistor Special Issue No 1 Radiotronics 22 No 9 Amalgamated Wireless valve Company Pty Ltd

AWV 1957 Transistor Special Issue No 2 Radiotronics 22 No 10 Amalgamated Wireless valve Company Pty Ltd

AWV 1957 Transistor Special Issue No 3 Radiotronics 22 No 11 Amalgamated Wireless valve Company Pty Ltd

AWV 1958 Manufacturing the AWV Transistor Amalgamated Wireless Valve Company Pty Ltd

Braun E, Macdonald S 1982 Revolution in miniature: the history and impact of semiconductor electronics re-explored in an updated and revised second edition Cambridge University Press, 1982

Choi H 2007 The Boundaries of Industrial Research Making Transistors at RCA 1948-1960 Technology and Culture 48 758-82

Craig D 1999 Dr Louis Davies Physicist, an oral history conducted in 1999 by Dr David Craig, Australian Academy of Science

Given J 2008 The Radiola and the Radiotron: localising global products in early radio broadcasting. Paper presented at the annual meeting of the International Communication Association, TBA, Montreal, Quebec, Canada, May 21, 2008

Given J 2007 Not Being Ernest: Uncovering Competitors in the Foundation of Australian Wireless. Historical Records of Australian Science 18, 159–176.

JETEC 1957 Joint Electron Tube Engineering Council Release 2037 Filed November 4th 1957

JETEC 1958 Joint Electron Tube Engineering Council Release 2295 Filed October 20th 1958

Mingay 1963 Forty three Years of Valve manufacturing Electrical Weekly July 12 1963 35

Rigby G A 2003 Historical Records of Australian Science, 14, no.4

Stokes J W 1982 70 Years of Radio Tubes and Valves Vestal Press Ltd New York 1982

Watt T 2009 The Transistor Story Oral history of Ted Watt AWA Archives courtesy J McIlwaine

Annex: AWA and its Stakeholders

The Australian Wireless Company (AWL) was formed in 1909 by a group of investors led by Mr Hugh Dennison of Sydney and Telefunken obtaining rights to Telefunken’s technology in order to bid for wireless contracts. The company successfully bid for the first two coastal wireless stations to be built in Sydney and Freemantle by discounting its pricing. With early success in hand almost half the company was then floated in a first public offering. [Given 2007]

Marconi, which had invested considerable time and resources promoting its own wireless system in Australia, reacted strongly alleging breach of its patents and suing both the Australian Government and AWL. Marconi won in the High Court of Australia and at the Lords of Appeal in London. This gave Marconi the right to inspect the AWL installations. [Baker 1970]

The Marconi strategy was part of an aggressive approach to Telefunken taken by its new Managing Director, Godfrey Issacs who had been appointed in January 1910. One of his first successful suits was against the British Radio Telegraph and Telephone Company for infringement of the Marconi “Four Sevens” patent that covered an improved method of tuning developed by Marconi. Realising its own position was not entirely robust, Marconi then purchased the pioneering tuning patent in the field conceived by Oliver Lodge. With this in hand, Marconi held an unassailable position and when Marconi won a suit it brought against Siemens (for its use of Telefunken equipment), Telefunken sought a negotiated settlement. In 1912 Telefunken and Marconi made an agreement to pool their patents, both past and future for all territories outside the USA. [Baker 1970 Given 2007]

In 1912 Dennison went to London to negotiate a merger of the interests of Marconi and Telefunken in Australia. This led to the formation of Amalgamated Wireless Australasia Ltd (AWA) in 1913 which would own the Australian assets of Marconi and AWL. Marconi owned 50% of the new company, Australian investors owned 44% and Telefunken owned 6%. [Baker 1970]

AWA had been manufacturing tubes under license to RCA since 1927 but decided to form a dedicated company that would produce tubes on a much larger scale. It considered and rejected an alliance with either Marconi-Osram or Philips and opted to continue its relationship with RCA.

AWV was formed in 1932 in a consortium 55% AWA and 45% RCA, International General Electric and Westinghouse. [Given 2008 Mingay 1963] The company was formed the same year a Federal US antitrust action was successful against RCA and General Electric.

The International Stakeholders

Beneath the surface of the AWV deal there was a history of complex international relationships that had evolved since the beginning of the 20th Century, connected parties on opposite sides of two world wars, faced legal challenges and inquiries and ultimately petered out as new entrants enforced the retrenchment of these companies’ traditional interests.

The global ambitions of General Electric were taken on by the International General Electric Company formed in 1919. By 1938 it had 14 associate companies in addition to AWV. These included other tube and future semiconductor companies such as Companie Francaise Thomson-Houston and Philips as well as electrical companies. Its interests included substantial holdings in Germany such as the AEG company and many others. From 1922 IGE made agreements with AEG to pool patents and not to compete in agreed territories.

Telefunken was founded in 1903 by AEG and Siemens thereby connecting it to International General Electric. Telefunken had a virtual monopoly over radios and had a no-compete agreement for Germany with RCA as part of a patent pooling arrangement. [Mirow 1982] RCA licensed Telefunken its semiconductor technology and exchanged information on improvements with that company [Flaningam 1945 Herold 1953]

The Radio Corporation of America (RCA) was formed in 1919 by General Electric, AT & T and United Fruit with the encouragement of the US Federal Government which wanted to maintain US leadership in long distance communications. During World War I General Electric’s Alexanderson wireless transmitters and American Marconi’s shore stations were regarded as strategic assets having proved their value when Germany cut the trans-Atlantic telegraph cables and wireless was the only alternative. General Electric and Westinghouse made RCA their exclusive marketing channel for receivers and tubes in return for cross licenses to their patents.

In forming RCA accommodations were reached between British Marconi and General Electric. American Marconi was taken over by General Electric as part of a US Navy mediated integration of US wireless assets post World War One. These were transferred to RCA. Marconi and General Electric pooled patents and granted exclusive rights to Marconi for the Commonwealth, China and parts of Europe and RCA was granted rights to the Western Hemisphere except Canada. [Given 2008] RCA took over the business of American Marconi.

These agreements were subject to on-going regulatory scrutiny most notably in the United States where anti-trust actions where brought. The alliance that had formed RCA in 1919 was disestablished in 1932 following successful Federal antitrust actions against RCA and General Electric. In 1945 an action was commenced against General Electric in regard to its foreign subsidiaries and the nature of their anti-competitive agreements.

Consequences to AWA

Through the founding agreement of AWA the company obtained rights to Marconi and Telefunken patents. The agreement between Marconi and General Electric gave AWA the rights to General Electric patents although it needed to pay for these in the sum of £3,000. [Given 2008]

Thus AWA held all the major cards in wireless and was eager to exploit them in its own products and through selectively licensing other Australian companies. In 1927 the Australian Royal Commission on Wireless strongly criticised AWA’s business practices in this regard and recommended an open license in return for a broadcasting fee paid by each listener. This created a highly profitable income stream for AWA until the arrangement ceased in 1933.

In bringing together the original AWA stakeholders, Telefunken and Marconi and the new three parties, AWV is a unique case study of a remarkably complex intersection of interests. Both AWA and AWV obtained huge benefits from arrangements orchestrated elsewhere that had global reach. They could access the technologies of giants in their field. But access to international markets was not a benefit available to the Australian companies: the terms of their licenses confined them to them to their own backyard.

In time the pioneering patents that made global cartels possible expired and licensing became a conventional tool of technology transfer. The transistor was invented by Bell Laboratories which meant that producers either needed a license from Bell or, for example, from RCA through their cross-licensing arrangement with Bell. [Braun and Macdonald 1982] RCA maintained a network of cross licenses which made its transistor developments available to its licensees in return for their research. The cost of the licenses could be off-set by the contribution by the licensee.

This created an incentive for AWA to invest in its own research and patenting programme in order to have intellectual property it could trade and off-set license fees. “The Physical Laboratory thus made a direct contribution to the company's commercial negotiations.” [Rigby 2003]

References

Braun E, Macdonald S 1982 Revolution in miniature: the history and impact of semiconductor electronics re-explored in an updated and revised second edition Cambridge University Press, 1982

Flaningam M 1945 International Cooperation and Control in the Electrical Industry: The General Electric Company and Germany Am Jnl Economics and Sociology 5 No 1 7-15

Herold 1953 Herold’s German Visits, Tubes and Semiconductors, 1953 RCA Technical Report PEM 373

Rigby G A 2003 Historical Records of Australian Science, 14, no.4

Baker W 1970 A History of the Marconi Company Methuen & Co London

Given J 2008 The Radiola and the Radiotron: localising global products in early radio broadcasting. Paper presented at the annual meeting of the International Communication Association, TBA, Montreal, Quebec, Canada, May 21, 2008

Given J 2007 Not Being Ernest: Uncovering Competitors in the Foundation of Australian Wireless. Historical Records of Australian Science 18, 159–176.

Mingay 1963 Forty three Years of Valve manufacturing Electrical Weekly July 12 1963 35

Mirow K R and Mauerer H 1982 Webs of Power International Cartels and the World Economy Boston: Houghton Mifflin Company