Inside a transistor

You can wire a transistor correctly for the rest of your life without reading this page. It is here because the number on the last page — a hundred times — looks arbitrary until you have seen where it comes from, and because the thing it comes from is a junction you have already met.

Three slabs, and the middle one is a sliver

Press the button and watch it switch on.

Two pieces of n-type silicon with a very thin piece of p-type between them. That is an NPN, and the name is just the order of the slabs.

Two junctions, one of them backwards

There are two p-n junctions in there, and each one is a diode — the exact junction from the last chapter. Wire the transistor into a circuit with nothing on the base and they sit back to back, so whichever way you push, one of them is blocking. That is the off state, and it is why a transistor does nothing until you ask.

Drive the base 0.7 V above the emitter and the near junction forward biases. Its depletion region collapses, exactly as a diode's does, and electrons pour out of the emitter into the base.

Why the middle has to be thin

Those electrons are now in p-type silicon, which is full of holes, and an electron that meets a hole stops being a free carrier. If the base were an ordinary slab that is where the story would end and you would have built a slightly odd diode.

The base is not an ordinary slab. It is thinner than the distance an electron covers before it meets anything, and it is doped so lightly that there is almost nothing in it to meet. So the electrons cross it, miss, and fall into the collector's field, which sweeps them away.

Which is why you never trust the gain

Drag it up from zero. The collector current tracks the base current at a hundred to one, right up to the point where the load has taken everything the supply can give. After that, more base current changes nothing.

Two things fall out of that plot, and they are the two things this chapter is for:

• The middle is where the heat is. Half open, the transistor has several volts across it and a large current through it, and the product is watts. Watch the third readout as you drag: it peaks in the middle and collapses at both ends.
• The right-hand side is free. Past the knee, the transistor keeps only about 0.2 V across it. Same current, a twentieth of the heat. That state is called saturation, and a transistor used as a switch lives there.

So the design rule is not "work out the gain and supply exactly that much base current". It is: work out what the load needs, then deliberately over-drive the base until the knee is far behind you. A gain of 100 in the datasheet may be 60 in the part you bought and 300 in the next one out of the bag.