How a BJT Works

The two-diode model predicts no collector current, and it is wrong. Two discrete diodes wired back to back genuinely do not amplify. What matters is that the junctions share the same thin base, so carriers injected by one arrive at the other.

The back-to-back picture is useful for identifying pins with a multimeter, and for nothing else.

1. 1Forward-bias base-emitter at about 0.7 V; the depletion region collapses.
2. 2The heavily doped emitter injects a flood of electrons into the base (in an NPN).
3. 3The base is thinner than a diffusion length, so about 99% cross it without meeting a hole.
4. 4The reverse-biased base-collector junction's field sweeps them into the collector.
5. 5The 1% that recombines must be replenished through the base terminal — that is I_B.

So β = I_C/I_B is just the ratio of those fractions, roughly 99:1, which is why it lands near 100. Kirchhoff still applies at the terminals: I_E = I_B + I_C.

• Heavily doped emitter — injection is nearly all one-way, emitter into base, which keeps the base current small.
• Thin, lightly doped base — few holes to recombine with, and a short distance to cross, so transport efficiency is high.
• Large collector — it must dissipate the power and withstand the reverse voltage.

The geometry is asymmetric, which is why swapping the emitter and collector gives a working but very poor transistor. Reverse-active β is typically under 5.

Datasheets quote a 4:1 range for a single part number, and β shifts with temperature and with collector current on top of that. Any circuit whose operating point depends on β will behave differently from one device to the next.

That single fact is why the biasing topic exists, and why every practical bias scheme is built to make β disappear from the answer.

Analogue design lives in the active region. Digital design uses only cut-off and saturation, and drives the base hard so that V_CE collapses and dissipation stays low.