Semiconductor History

Burr-Brown PCM1702-L I've just acquired from eBay a CD player to complement another very old CD player, both with Burr-Brown Digital to Analog converters. The photo shows the inside of the Denon CD-655 from the late 90s, a super low distortion and high signal to noise ratio CD player which uses 2 x PCM1702 DACs from Burr-Brown. The BiCMOS process 20-Bit PCM1702 was a really interesting DAC which was an improved version of the PCM63P. Early R-2R ladder DACs suffered from some noise/glitches at low output levels (and degraded linearity) due to the Most Significant Bit current source turning on at the lowest volume output point. Multi-bit DACs had significantly higher performance than contemporary 1-bit and "MASH" DACs, but had this one flaw. The PCM1702 was effectively two 19-bit DACs in a complementary design which negated the zero cross point issue. They had all the benefits of R-2R ladder DACs and now removed the other issue. Around this time Burr-Brown were al...

Silicon npn transistors from Toshiba and Hitachi with an NEC pnp in the middle. Following the development of the point contact transistor at Bell Laboratories (and subsequent grown junction transistor development), Tokyo Tsushin Kogyu became the earliest Bell Labs transistor licensee in Japan. They moved quickly, producing their first transistor radio in August 1955. This was one year after Texas Instruments introduced the TR-1, the world's first production portable transistor radio. Tokyo Tsushin Kogyu, Hitachi, Tokyo Shibauro Electric, Mitsubishi Electric and Kote Kogyo became the first five Japanese transistor licensees of Bell Labs and Western Electric by July 1956. Tokyo Tsushin Kogyu became Sony and Tokyo Shibauro Electric became Toshiba. Semiconductor developments progressed quickly during the second half of the 1950s and by 1959 50% of transistor radios bought in the US were Japanese, or used Japanese transistors. Unlike in the US where the Defense market was s...

Sinclair, TI and Philco Transistors Clive Sinclair formed Sinclair Radionics in 1961, having a good understanding of electronics and transistor manufacturing. Always the entrepreneur, he bought a quantity of transistor test failures from Semiconductors Limited (Semics) of Swindon. As seems to be the norm in British semiconductor history, it was convoluted. Semics was actually Plessey who had licensed Micro Alloy Diffused Transistor technology from Philco in the US. Sinclair used the transistors in a very small pre-amplifier and sold both the amplifier and retested/re-badged transistors separately. Later on he must have done this again as ST140 and ST141 transistors were on the market, albeit probably in relatively small quantities again. Above left is an ST140 Sinclair transistor. There is also an original Philco MADT transistor on the right. In the middle is a UK manufactured Texas Instruments transistor. Texas Instruments opened a semiconductor plant in Bedford, UK in 1960 to ...

Inside the Telstar. Calculator for Size Comparison The Coleco Telstar was launched in 1976. It played six games, but only three were offered in Coleco's first product. The General Instrument AY-3-8500 was the first single IC to offer multiple games, made available to every manufacturer, launching home video games as a mass market. It marked the end of discrete systems such as the Magnavox x00 products. The Coleco Telstar was a highly integrated system with only a few components for the GI chip. The RF cable/TV Out wrapped round the ferrite core in the photo above was a late addition/modification when Coleco failed its FCC approval. Ralph Baer, who had seen an early demo of the GI chip, and introduced the new General Instrument  IC to Coleco, suggested the addition to suppress EMI. Ed Saks, who was the head of GI had brought the two design guys from GI's Scottish facility to the US to develop the AY-3-8500 fully, and the further chips that came in the following years. ...

TI 904 Si transistor Texas Instruments introduced the 900 series silicon transistors in 1954, the first silicon transistors in history. The upstart TI of Dallas managed to beat the established East Coast companies to the first silicon products. It was through the hire of important Bell Labs people in the development of silicon that enabled the feat. The higher temperature operation of silicon was preferable over the dominant germanium, so getting Si products into the market would give that company an edge. TI started transistor development in Lemmon Avenue in Dallas and in 1958 moved to a 300 acre site at North Central Expressway. The Semiconductor building was the first building on the campus. The building is still there (although modified) and is now a Raytheon building following the acquisition of TI's defense business in 1997. I visited the facility during the 2000s and you could still see the roots of TI in the fabric of the building. hfe=28, Vf=725mV, low gain but it ...

Western Electric Point Contact Transistor Developed by John Bardeen and Walter Brattain at Bell Laboratories, the first transistor was a seminal moment in history. In the main because it was part of the American Telephone and Telegraph Company, Bell Labs was very open in its sharing of the invention, including its methods of manufacture. Being a monopoly, AT&T was very aware of anti-trust pressures. Bell Labs shared its knowledge with 40 licensees in 1951. From demonstrations in 1948, a number of companies had already seen the potential impact of the transistor and had started their own developments. Those with most to lose, the vacuum tube manufacturers, were quickest off the mark including Raytheon, the biggest supplier of mini vacuum tubes. Point Contact transistors were difficult to manufacture and their reliability was poor. Superior transistor manufacturing methods developed including grown junction and alloy junction. Western Electric at its Allentown PA facility contin...

Western Electric 2N27 Early Transistor To the left is a Western Electric grown junction transistor from the mid 1950s, when full rate production started on the first grown junction transistors. The technology was developed in Bell Labs after the point contact transistor, and was significantly more reliable and manufacturable. William Shockley, later to form Shockley Semiconductors on the West Coast, developed the theory of the grown junction transistor. However they didn't last too long, being pretty much superseded by Alloy Junction transistors. The 2N27, 2N28 and 2N29 were the first grown junction transistors. GE 2N167 Transistor General Electric were also an early developer or grown junction transistor technology. The 2N167 germanium grown junction transistor on the left is from the late 1950s. There is a small slab (or bar) of germanium which has been sliced from a larger grown crystal with impurities added to make the npn type transistor structure. The emitter and ...

Fairchild Micrologic Devices Fairchild introduced the Micrologic standard logic family in 1961. These were the world's first ICs, based on Jean Hoerni's revolutionary planar process which laid the blueprint for all future integrated circuits. There were eventually 25 standard logic RTL ICs. Products in the family included the 900, a buffer, 901, a 3 gate counter adapter, the 902, a flip-flop, 923, a JK flip flop and the 959, a 4bit latch. Above is a  3-input NOR device from the early 60s (uL903) in a metal can, and a later packaged JK flip flop (uL923) in a lower cost epoxy topped package. During the 1960s, Fairchild became the first semiconductor company to establish a lower cost assembly facility in the Far East (Hong Kong). Decapped uL903 Micrologic Device The 3-input NOR from Fairchild became the basic building block of the Apollo guidance computer, designed by MIT and built by Raytheon. It used 5,000 3-input NOR ICs and was the first major application of ICs. ...

IBM Standard Modular System (SMS) Card The Standard Modular System for mainframes was introduced by IBM in the late 50s for the 7030 Stretch program. The card on the left is an IBM SMS card post 1969 even though the card format had changed with the System/360. This card is from a 2420 tape reader. Tape machines and other peripherals kept the SMS format and the component technology longer. The NPN transistors are date code 1970 but are IBM 030 and 044 transistor models defined in the late 1950s. The IBM chrome coloured module on the right of the card is SLT (Solid Logic Technology) format. Later IBM Format Card Monolithic System Technology (MST) was introduced by IBM in 1968 and first appeared in the System/370. It replaced the earlier discrete transistor and diode logic SLT, SLD and ASLT hybrid logic with monolithic circuits, still on ceramic substrates in the same IBM package format.

Early IBM Alloy Junction Transistor This is an IBM germanium transistor from the 1950s, early package style. IBM were very active in the early days of transistor development and manufacturing, creating the world's largest manufacturing capability by the early 1960s. IBM also contracted companies like TI and Raytheon to produce Alloy Junction transistors for its computers. This pnp transistor works, hfe=74, Vbe=0.33V.