From Sticks To Chips

History 101 — Transistors

Once upon a time, some really smart people invented aids they could use to do arithmetic, like piles of sticks, and of course the abacus.

The abacus has been around since long before Arabic numerals joined the party. Variants are still in wide use today in many parts of the world. Here’s an image of an abacus beside an artifact that seems nearly as ancient, an Intel 80486 (the first in the x86 line to contain more than a million transistors). Both of these things were acquired as tourist souvenirs.

The abacus was the most powerful piece of calculation hardware for almost 2 millennia, then some folks began coming up with ideas for mechanical crank machines for arithmetic. Charles Babbage was the one I first heard about. He designed some cool machines, but it was over a century after his death that the first working complete Difference Engine based on his design was built. I was fortunate to have seen this mechanical marvel operating when it was on exhibit at the Computer History Museum here in Silicon Valley. It was about the size of a small pickup truck, and likely weighed more.

The next step forward in mechanical calculation was an electromechanical solution, using electromagnetic relays. There is an example of such a relay from a very early computer, in the photograph at the top of this page. In these relay machines, when the electromagnets are powered, they pull on a ferrous metal lever. That lever operates one or more contact switches which can be used to power other relays. Calculations performed using relays are an order of magnitude faster than hand-crank computation, but they require a lot of power and they generate a lot of heat. They also still have mechanical moving parts that wear out.

The first truly electronic computers used vacuum tubes, or valves as they were called at the time. An example of a quick-replace vacuum tube module from an early IBM computer is shown in the photograph at the top of this page. Vacuum tubes looked much like old fashioned hot-filament light bulbs. Vacuum tubes could switch electrical current a couple of orders of magnitude faster than electromagnetic relays, but like those old lightbulbs, they had a tendency to burn out and they were very fragile. The first entirely electronic computer based on vacuum tubes was quite unreliable, with a tube somewhere in the machine burning out on average every 2 hours of operation. Expert operators learned how to determine which tube had failed, and how to replace it quickly without breaking it or other tubes.

Of course all of that historical prelude is just there to put into context the importance of the transistor. The transistor is arguably the most important and most significant invention in the history of the human race (up until now, at least). When transistors were invented they very quickly replaced vacuum tubes. They were faster than tubes; they were almost indestructible; they generated almost no heat; they were much more reliable — no burn outs; they were much smaller; they cost much less to manufacture, and they consumed only the smallest fraction of the power of the old tubes. How often does an invention come along that is so dramatically superior to the preceding technology in so many different ways?

Early transistors were usually housed in small metal cans, like the one shown in the photograph at the top of this page. Eventually they were manufactured in small plastic enclosures, like the modern transistor shown on the right of that photo.

Over time, another advantage of transistors over tubes became apparent. That advantage quickly led to personal computers, and the explosion of portable consumer electronics devices that are now ubiquitous in the first world. That advantage is that transistors could be constructed on a microscopic scale in large quantities and pre-wired together with their neighbors on a single computer chip. The early 16 pin Intel chip on the left of the photograph at the top of this page is one of the support chips for the world’s first microprocessor, the Intel 4004.

The pictures below show an early board with dozens of transistors (left), and another board containing early integrated circuits or chips (right).

 

Many modern chips of course now have billions of transistors packed into single packages.

Typically transistors have 3 connectors, and you can think of them being arranged like a capital letter T:

The transistor diagrammed above is an NPN transistor. For this type of device you can wire up the two connectors at each side of the top part of the T (collector on left, emitter on right) into a circuit as a switch (grounding the emitter side). Then you can provide a very low intensity signal to the connector on the vertical part of the T (called the base) which will cause a much greater current flow across the top part of the T, essentially turning on the switch.

If you want to try building some basic transistor circuits, check out the local Fun With Transistors article.

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.