Multiplexing: FDM and TDM Basics: Complete Beginner's Guide

Multiplexing is the art of sharing one expensive pipe among many conversations. Laying a separate cable for every phone call would bankrupt any network, so engineers combine many signals into one shared channel at the transmitter and separate them again at the receiver — like many letters sorted into one mailbag, then re-sorted at the destination.

The two classic methods divide the pipe differently: FDM (frequency-division multiplexing) gives each conversation its own slice of frequency, all talking at once; TDM (time-division multiplexing) gives each conversation its own repeating slice of time, taking turns so fast nobody notices. These ideas underpin everything from FM radio to the internet backbone.

Key Concepts

• FDM: each signal is shifted to its own frequency band (channel) with guard bands between them; all channels transmit simultaneously.
• TDM: time is chopped into frames of slots; each user owns one slot per frame and transmits only in it.
• Guard band / guard time: small unused gaps that stop neighbouring channels or slots from bleeding into each other.
• Statistical TDM: slots are handed out on demand instead of fixed rotation, so idle users waste no capacity — the packet-switching ancestor.
• WDM: wavelength-division multiplexing — FDM's optical-fibre child, packing dozens of laser colours into one glass strand.
• Synchronisation: TDM needs all parties to agree where frames start; FDM needs stable carrier frequencies instead.

How It Works

1. FDM transmit: each baseband signal modulates its own subcarrier (e.g. voice 1 on 64 kHz, voice 2 on 68 kHz), and the subcarriers are summed onto the line.
2. FDM receive: a bank of bandpass filters splits the composite back into channels, and each is demodulated to baseband.
3. TDM transmit: each input is sampled in turn (a commutator switch in the classic picture), and the samples are interleaved into one fast stream.
4. TDM receive: a synchronised de-commutator routes each sample to its own output, and low-pass filters rebuild the original signals.

Worked Example

Bundle four voice calls, each needing 0–4 kHz. FDM version: give each a 4 kHz slot plus 1 kHz guard bands — channel centres at 4, 9, 14, 19 kHz — total bandwidth ≈ 20 kHz, all four talking continuously. TDM version: sample each voice at 8 kHz (Nyquist), so one frame holds 4 samples; the line must carry 4 × 8000 = 32,000 samples/s, i.e. a 32 kHz sample clock with each user getting one slot every 125 µs. Both carry exactly the same four conversations: FDM spends frequency, TDM spends time. Scale this to 24 voices and you get the famous T1 line (1.544 Mbps); scale FDM to fibre and you get WDM systems carrying terabits.

Common Misconceptions

• "Multiplexing creates extra capacity from nothing." Wrong: it only shares existing capacity more cleverly — the total information rate can never exceed the channel's Shannon limit.
• "FDM and TDM are obsolete." Wrong: your FM radio dial is pure FDM, fibre backbones are WDM, and every mobile standard still uses TDMA-style time slots underneath.
• "TDM users talk one at a time, so calls sound choppy." Wrong: slots repeat thousands of times per second, far faster than the ear can resolve — the conversation sounds perfectly continuous.

Real-World Applications

• FM radio: dozens of stations share the 88–108 MHz band by FDM, 200 kHz apart.
• Cable TV stacks hundreds of channels onto one coax by FDM.
• Old telephone trunks: analog FDM hierarchies and later digital T1/E1 TDM lines carried the world's calls.
• 4G/5G and Wi-Fi blend both ideas: OFDMA slices by frequency, TDMA-style scheduling slices by time.

FDM shares the pipe by frequency — everyone talks at once in their own lane. TDM shares it by time — everyone gets the whole pipe in rapid turns. Every modern network, from your FM dial to a transatlantic fibre, is some descendant of these two ways of sharing.