an analog multiplexer

An analog multiplexer is a bank of analog switches sharing one common terminal, plus the logic to close exactly one at a time. It is a rotary selector knob made of silicon: many input channels, one output, and a digital address that picks which input gets connected through. A one-of-eight mux has eight inputs and three address lines (because 2^3 = 8), and whichever binary number you put on the address selects that channel.

Think of a railway with eight sidings merging into one main track through a points lever. The address bits set the points so that exactly one siding connects to the main line and the other seven are isolated. Inside, each channel is a transmission-gate analog switch; a small decoder reads the address and closes only the selected switch. Because the switches pass signals both ways, the same chip works as a multiplexer (many-to-one) or, run backwards, as a demultiplexer (one-to-many) that distributes one signal to a selected output. The selected path carries the same on-resistance, signal-range, and charge-injection caveats as a single analog switch.

Multiplexers are how one expensive, precise amplifier or one ADC can service many channels: scan the address through 0 to 7 and read each sensor in turn. They build keypad scanners, data-acquisition front ends, and audio source selectors. The honest catch in measurement work is crosstalk and settling: when you switch channels, the new signal must charge the shared wiring and the input capacitance through the on-resistance before a reading is valid, so you must wait a settling time after each switch, and a strong signal on one channel can leak slightly into its neighbours. Faster scanning trades against accuracy.