Making a 2N78 Transistor

Making a 2N78 transistor

In September 1954 electronics journals carried a press release from General Electric describing their rate grown process and advising that volume production would commence in 1955. At the time, General Electric had been producing small volumes on a pilot line at Syracuse. Volume production would be achieved by mechanisation of the process of attaching leads to the transistor bars. The pictorial article carried in Radio Electronics and reproduced in Toute la Radio was largely based on the pilot production line.

This article supplements my General Electric History home page.

The genesis of the process: the crucible contains pure germanium and traces of impurities from N and P type impurities: gallium and antimony. The withdrawal rate and melt temperature are subjected to cyclic variations during the pulling of the crystal.

The single crystal, which is created from a small seed crystal brought in contact with the liquid germanium and raised slowly while revolving over two hours, now weighs over 100 grams.Inside, hundreds of PN junctions are available to provide about 2000transistors.

Each rib results from a single cycle of the process and represents a PN junction formed in this process.

The crystal is first split along its axis, and each half cut into as many half-moons (a) there are NPN or PNP sandwiches. The fragments are in turn cut into thick strips of 0.5 mm (b and c). The bands are finally split, every 0.5 mm (d).

The cutting technique uses thin diamond discs mounted side by side on a milling machine. Fragments of germanium are glued on glass plates. The bars obtained are transistors.

A single crystal yields 2000 bars 0.5 X 0.5 X 2.5 mm in which np and pnjunctions have been formed duringcrystallization.

The next step is the delicate operation of making the soldered connections.

At present, soldering takes place piece by piece, under the microscope. But automatic machines are being developed to enable annual production of several million.

The bars now have their connections: ribbons at the ends to the emitter and collector and a barely visible wire in the center connecting the base.

Finally, the transistor is mounted on its header. A metal can is sealed on the header and a vacuum is drawn. Then the transistor passes final inspection.

Transistor characteristics can be changed by adjusting the cycles of crystallization.

Sources:

Toute la Radio September 1954 Fabrication en Masse Des Transistors USA reproducing an article in Radio Electronics. Scan courtesy of Christian Adam.

Electronics September 1954 Transistor Production Increases in Volume