Forbes on the Tranistor’s birthday! –
Sixty years ago, on Dec. 16, 1947, three physicists at Bell Laboratories in Murray Hill, N.J., built the world’s first transistor. William Shockley, John Bardeen and William Brattain had been looking for a semiconductor amplifier to take the place of the vacuum tubes that made radios and other electronics so impossibly bulky, hot and power hungry. They were so instantly certain they’d found their answer that they didn’t speak a word of it to anyone for six months, until they could experiment further and apply for patents.
Then on June 30, 1948, they held a press conference in New York City. They showed the world not only a big model of a transistor but also a TV and a radio with transistors in place of the tubes. Nobody was talking about anything like computers yet, but it was a first look at the future we all live in. The world’s response? The New York Times ran an item at the bottom of its “News of Radio” column on page 46.
18 thoughts on “Happy birthday transistor!”
Comments are closed.
ADVERTISEMENT
Join Make: Community Today
That is one way to tell the story of the transistor, and its factual, but its rather a short narrative. During the war, semiconductor diodes came into use. These were made from naturally occurring intrinsic Si slabs that had a dopant impurity. These were identified by their low resistance. A Schottky diode was formed by metallic vapor deposition upon this substrate. After the war, physicists were commissioned by the US federal government to explain what was happening on a quantum level in solid-state, and these guys had the idea for both the bipolar and FET transistor. Funding for the implementation of these devices was scarce, so the development team chose to go with the bipolar, leaving the development of the FET behind by about a half-decade. There were subsequent developments which made production of these devices more practical. There were also advances in circuit topologies which overcame the relatively lesser linearity of these devices as regards thermionics and saturable reactors.
By Jove, that thing is sexy! Has anyone out there built their own?
Hey now, don’t slander vacuum tubes. You won’t ever see high power transistor amps because electrons move through a vacuum easier than they do silicon (the cooling requirements would be impossible to keep up with). Over a few kilowatts you have little choice but to use tubes.
But for switching, obviously tubes are not the best choice for low power things like ICs.
-Bob
Tubes are robust when driving inductive loads relative to semis, as they have large, metallic control surfaces. They are good at high potentials, although semis are improving. Tubes come in only one flavor, rather than two, PNP-NPN. To achieve complementarity, one needs to use a transformer to invert the control signal. Transistors tend to saturate more precipitantly, leaving room for tubes in Hi-Fi amps. Tubes are still the king of the highest-power amps, such as clear-channel RF amp finals. This is because they can be scaled arbitrarily. A good technique in high-power semi-based amps is to use parallel FETs. These transistors have a positive tempco in this regime, and thusly can be easily paralleled.
There are secondary apps that are native to both devices: Thermionic devices can be used to measure vacuum, as they cannot, typically, operate without it. Diodes and transistors can be used to measure temperature, as their drops are proportional to absolute temperature.
As far as making one’s own, the exemplar is a BJT, an FET would be far easier to construct.
Harold S Black’s invention of negative feedback allows transistors to function far more dependably in linear applications. Without this, they probably would not have supplanted thermionics in this domain.
This page has a hilarious account of some college students building their own transistors.
Actually, a german citizen named Lilienfeld built AND patented the first FET transistors – in 1928.
Shockley built his devices based on Lilienfeld’s patents – and this was all documented in court proceedings. Shockley’s germanium transistors weren’t very practical, and not suitable for practical use.
However, as history shows again and again, marketing is more important than development. GE had marketing dollars, and a strong PR department.
Shockley (and the folks who did the actual work) were bright guys. A few years later, they built the bipolar junction transistor, which did become useful and was the start of the transistor revolution as we know it.
But the FET, which forms the basis of most modern IC circuitry, was invented outside of America. In fact, another german patented and built a second FET device 10 years after the first guy.
PS – Don Hersey aka SJ – I can tell that you’re well read; thermionics isn’t exactly a common term these days. But as a former golden age of wireless hobbyist, I have to mention that some of the statements you’re making about tubes and amplifiers reflect old conceits of the Hi-Fi crowd… I shan’t dredge old soil again.
Also, negative feedback existed before Black wrote his paper: It was known and used in steam and electrical power generation for some time before he applied it to a different type of valve.
Best Regards,
Peter
Yes, in fact, steam engines can fail catastrophically if their governors fail. Black brought a more formal understanding to negative feedback and generalized its application to filters. I have little patience with the ‘golden ears’ crowd as well, but the high-gain characteristic of Qs versus tubes does require transistor amps to employ the use of more degeneration than tube ones typically do. This can lead to more IM distortion for an equal design effort. Soft-saturation characteristics of tubes can hide liabilities of Rth limitations of the power supply. Both types of amplifiers can be made well. Plans to miniaturize tubes and group more of them on a substrate were in the works at the time transistors came in. Vacuum fluorescent displays represent one case of the integration of thermionics. WGN Schaumburg uses tube finals for their AM 1000 amp. These babies are twice the size of a school bus and are replaced and refurbished every nine months. Of course, these are not the largest tubes in use in Illinois. That would be the particle accelerator named after Enrico Fermi, which is a giant ‘tube’ in itself. Another region of superiority for firebottles is the generation of radiation — Semiconductors currently can generate soft UV — Tubes can be made to generate X-rays!
Surely, ‘negative feedback’ has a long history. Scolding a child could be a considered a case of negative feedback — along with its similar attendant susceptibility to oscillation!
My only point is that thermionics are still interesting. Another domain of modern application is the magnetron-klystron I employ in the kitchen to warm my coffee, energizing my further (usu. transistor-based) design efforts! –regards, Don
Yeah, the photo shown is probably a point-contact transistor, which would be hard-to-find, nowadays. A big practical innovation in the production of semis was ‘zone melting’ a technique which acts a little like a distilling column for Si. When the early semi industry brought in chemists as consultants, their suggestion to use distilled water as opposed to tap water really helped. Fairchild Camera and Instrument’s development of the op-amp (for a third-party camera light-meter) was another major step in the dissemination of semis.
The story that the Germans had a transistor during the inter-war period is remarkable. One thing they didn’t have during the war was the pentode tube. Because of this, their RADARs weren’t quite as good as the Anglo-Americans.
//no spring chicken
//more recent vintage than ‘wireless’