A multiplexer is a device that takes multiple input signals and funnels them into a single output. Often shortened to MUX, it works like a switchboard: control signals tell it which input to connect to the output at any given moment. Multiplexers show up everywhere, from the processor inside your computer to the fiber optic cables that carry internet traffic across oceans.
How a Multiplexer Works
At its core, a multiplexer is a selector. Imagine a hallway with several doors on one side and a single door on the other. A multiplexer decides which of those many doors opens into the hallway at any given time, letting only one signal through to the output. The “decision” is made by a separate set of inputs called select lines (or control lines), which act like an address telling the multiplexer where to look.
The number of select lines determines how many inputs the multiplexer can handle, following a simple rule: with n select lines, the multiplexer can choose from 2 raised to the n power inputs. So one select line handles 2 inputs (a 2-to-1 MUX), two select lines handle 4 inputs, three handle 8, and four handle 16. No matter how many inputs exist, only one connects to the output at a time. The select lines simply change which one that is.
Even the simplest multiplexer, a 2-to-1, can be built from basic logic gates. Larger multiplexers use the same principle scaled up, and they can also be constructed by combining smaller multiplexers together in layers.
Digital and Analog Multiplexers
Digital multiplexers switch between signals that represent binary values (ones and zeros). These are the type found inside computer chips, where they route data between different parts of a processor. They’re built from logic gates and operate purely in the digital domain.
Analog multiplexers handle continuous signals, like audio or sensor voltages, and they work differently under the hood. Instead of logic gates, they use transistor-based switches. In the “on” state, these switches have a resistance of less than 1 ohm, behaving almost like a closed wire. In the “off” state, resistance jumps to hundreds of megohms, effectively blocking the signal entirely. Modern analog multiplexers typically use a design called CMOS switching, which pairs two types of transistors together. This pairing keeps the switch resistance more consistent across different voltage levels, reducing distortion in the signal passing through.
Analog switches do introduce small errors. The on-resistance can cause slight signal loss, and variations in that resistance as the signal voltage changes can distort the output. For most applications these effects are negligible, but they matter in precision measurement systems where even tiny inaccuracies count.
Multiplexers Inside Computers
Your computer’s processor is packed with multiplexers, even if you’d never know it from the outside. Every time the CPU needs to choose between two possible data sources, a multiplexer handles the selection. When the processor picks which register to read from, a MUX routes the correct data to the output. When the arithmetic unit needs to decide between two operands, another MUX makes the choice. Program counters, memory address selection, and instruction decoding all rely on multiplexers to direct data to the right place at the right time.
These aren’t trivial components. In custom chip designs (ASICs), a six-input multiplexer occupies roughly the same silicon area as an adder, one of the fundamental math circuits in a processor. In programmable chips (FPGAs), multiplexers can actually take up more space than adders because of how they’re implemented. Engineers spend significant effort minimizing the number and size of multiplexers during chip design, since the way data is shared between hardware resources directly affects how many MUXes are needed.
Multiplexing in Telecommunications
The same core idea, combining multiple signals onto a shared channel, scales up massively in telecommunications. Rather than selecting one signal at a time inside a chip, telecom multiplexing lets thousands of conversations, video streams, or data connections share a single cable or radio frequency band. Three main techniques make this work.
Frequency Division Multiplexing
Frequency division multiplexing (FDM) splits the available bandwidth into smaller frequency ranges and assigns each signal its own slice. Traditional TV and radio broadcasting use this approach: every station gets its own frequency, and they all transmit simultaneously without interfering because they occupy different parts of the spectrum. Small gaps called guard bands sit between adjacent channels to prevent overlap.
Time Division Multiplexing
Time division multiplexing (TDM) takes turns instead of splitting frequencies. Each signal gets a dedicated time slot, and during that slot it has access to the full bandwidth of the channel. The slots cycle rapidly, so from a user’s perspective the connection feels continuous. Phone networks have used TDM for decades to carry multiple voice calls over a single line.
Wavelength Division Multiplexing
Wavelength division multiplexing (WDM) applies the frequency-splitting concept to light traveling through fiber optic cables. Different data streams are encoded onto different wavelengths (colors) of laser light, and all of them travel through the same glass fiber simultaneously. A denser version called DWDM (dense wavelength division multiplexing) packs 40 or more channels into a single frequency band. Modern DWDM systems carry individual channels at speeds of 100, 200, 400, and even 800 gigabits per second, all on one fiber. Networks originally designed for 10 Gb/s channels are now carrying 400 Gb/s and 800 Gb/s signals on the same physical infrastructure, thanks to flexible channel spacing that can be adjusted in fine increments to accommodate different data rates.
The Demultiplexer: Reversing the Process
A multiplexer is only half the equation. At the receiving end, a demultiplexer (DEMUX) does the opposite: it takes a single input and routes it to one of several outputs. If a MUX is a funnel that collects many signals into one, a DEMUX is a splitter that separates them back out. In a satellite communication system, for example, signals from multiple users are multiplexed together for transmission, then demultiplexed at the receiving ground station so each user’s data can be processed individually.
In digital circuits, demultiplexers use the same select-line logic as multiplexers, just in reverse. The select signals determine which output line receives the incoming data. Together, a MUX-DEMUX pair forms the backbone of any system where multiple signals need to share a single channel, whether that channel is a wire inside a chip or a fiber optic cable spanning a continent.