Modulation and Demodulation

Modulation and demodulation are the paired processes that let information travel. Modulation impresses a message onto a carrier by varying one or more of the carrier's properties, its amplitude, its frequency, its phase, or the parameters of a pulse train, in step with the information to be sent. Demodulation performs the inverse at the receiver, stripping the carrier away to recover the original message. Almost every electronic link that crosses a distance, from an AM broadcast to a fiber-optic backbone to the radio inside a phone, rests on this exchange, because raw information rarely suits the medium that must carry it, and a carrier is the agent that makes it fit.

The reasons to modulate are practical and cumulative. A baseband audio signal a few kilohertz wide cannot radiate efficiently, because an effective antenna must be a sizable fraction of a wavelength and the wavelength at audio frequencies is measured in kilometers; shifting the signal up to a carrier of megahertz or gigahertz brings the wavelength, and the antenna, down to a usable size. Modulation also lets many signals share one medium at once, each parked on its own carrier, so that frequency-division multiplexing can pack hundreds of telephone calls, television channels, or radio stations into a single cable or band without collision. And the choice of scheme buys robustness: by trading bandwidth for signal-to-noise performance, an angle-modulated or spread signal can ride through interference that would obliterate a simpler one. These advantages are why modulation has been central to communication since the first spark transmitters, and why the analog techniques developed here remain the foundation beneath today's digital schemes.

The four subcategories below move from the simplest carrier property to the most abstract, and then to the operations that knit a system together. The first develops amplitude modulation circuits, where information rides on the size of the carrier and detection can be as simple as a diode. The second covers frequency and phase modulation, the angle-modulation family that trades bandwidth for noise immunity. The third treats the mixer and frequency conversion, the multiplying element that shifts signals between frequencies and underlies every superheterodyne receiver. The fourth examines pulse modulation techniques, where the carrier is a train of pulses rather than a sine wave, forming the bridge from analog modulation to sampled and digital systems. The discussion that follows draws out the principles they share. Note that the treatment here is deliberately circuit-level, concerned with the devices and topologies that perform modulation and detection; the systems-level companion, Modulation and Signal Processing in communication electronics, surveys the same operations as they appear in complete links and extends them to the digital keying and multicarrier schemes built on these foundations.

Articles in This Category

Amplitude Modulation Circuits - Vary the amplitude of a carrier in proportion to a message, and you have the oldest and most intuitive form of modulation. This subcategory develops amplitude modulation (AM) from the multiplying action that produces it, through the spectrum it creates: a carrier flanked by an upper and a lower sideband, each displaced from the carrier by the message frequencies and each carrying a complete copy of the information. Coverage spans the circuits that generate it, from simple collector and base modulators to the balanced and ring modulators that cancel the carrier, and the variants that follow from suppressing the redundant parts of the spectrum, double-sideband suppressed-carrier (DSB-SC) and the bandwidth-and-power-saving single-sideband (SSB). Detection is treated in parallel: the envelope detector, a diode, capacitor, and resistor that recovers ordinary AM almost for free, and the synchronous (coherent) detector that multiplies by a regenerated carrier to demodulate suppressed-carrier signals. The recurring lesson is efficiency, that conventional AM spends two-thirds of its power on a carrier conveying no information, with at most about one-third reaching the sidebands at full modulation, which is precisely why the suppressed-carrier variants exist.
Frequency and Phase Modulation - Hold the amplitude constant and encode the message in the timing of the carrier instead, and you gain a decisive advantage in noise immunity. This subcategory covers the angle-modulation family, frequency modulation (FM) and phase modulation (PM), which are two views of the same operation, since the instantaneous frequency is the rate of change of phase and either can be produced from the other by shaping the modulating signal. Coverage includes the generation methods, the voltage-controlled oscillator and the varactor-tuned tank for direct FM and the Armstrong indirect method that integrates the message and applies it as phase, alongside the demodulators, the slope detector, the Foster-Seeley discriminator and ratio detector, the quadrature detector, and the phase-locked loop (PLL) that has displaced most of them. A central theme is the bandwidth-versus-noise trade: angle modulation spreads the signal over a band far wider than the message, estimated by Carson's rule as roughly twice the sum of the peak deviation and the highest message frequency, and spends that bandwidth to buy a quieter recovered signal and the threshold and capture effects that let a strong FM signal suppress a weaker interferer.
Mixer and Frequency Conversion - Multiply two signals together and the output contains their sum and difference frequencies, the single operation on which frequency translation depends. This subcategory develops the mixer, the nonlinear or switching element that shifts a signal from one frequency to another without disturbing the information it carries, the heart of the superheterodyne receiver that converts every incoming station to a common intermediate frequency (IF) where fixed, high-performance filtering and gain can be applied. Coverage spans the topologies, the simple unbalanced mixer, the single- and double-balanced mixers that suppress the local-oscillator and input feedthrough, and the Gilbert cell that realizes a precise four-quadrant multiplier in integrated form, together with the figures of merit that distinguish them: conversion gain or loss, noise figure, port-to-port isolation, and the linearity captured by the third-order intercept point. Two consequences recur throughout, the image frequency, a second input band that converts to the same IF and must be rejected by filtering or by an image-reject architecture, and the intermodulation products that a mixer's nonlinearity inevitably creates when strong signals are present.
Pulse Modulation Techniques - Replace the continuous sine-wave carrier with a train of pulses, and modulation becomes a sampling process, the conceptual bridge from analog signals to the digital world. This subcategory covers the pulse-modulation family, in which a periodic pulse train carries the message in one of its parameters: pulse-amplitude modulation (PAM) varies pulse height, pulse-width modulation (PWM) varies duration, and pulse-position modulation (PPM) varies timing, each sampling the message at a rate that the Nyquist criterion requires to exceed twice the signal's bandwidth. Coverage extends to the techniques that quantize as well as sample, delta modulation and its adaptive form, which transmit only the change between successive samples, and the closely related sigma-delta (delta-sigma) approach, whose oversampling and noise-shaping loop trades speed for resolution and underpins the high-resolution data converters used throughout modern audio and instrumentation. The unifying idea is that pulse modulation makes time discrete, turning a continuous waveform into a stream of samples that can be regenerated, time-division multiplexed, and ultimately coded into bits.

Themes Across Modulation and Demodulation

The four subcategories range from the diode detector to the noise-shaping loop, yet a handful of ideas run through all of them.

Every scheme is a chosen trade among bandwidth, power, and noise. There is no free modulation; each technique spends one resource to conserve another. Single-sideband halves bandwidth and saves the carrier's power at the cost of a more complex receiver; frequency modulation deliberately consumes extra bandwidth to buy noise immunity; pulse and sigma-delta schemes oversample to trade speed for resolution. Recognizing which resource a scheme protects, and which it spends, is the key to understanding why it exists and where it belongs, and these trade-offs are bounded ultimately by the Shannon-Hartley relationship between bandwidth, signal-to-noise ratio, and the information a channel can carry.

Modulation is multiplication, and multiplication makes sidebands. Impressing a message on a carrier is mathematically a multiplication, and multiplying two frequencies produces their sum and difference. This single fact explains the spectrum of AM, the operation of the mixer, the carrier and sidebands that appear and must be managed, and the recurring need to suppress, filter, or reject the products that are not wanted. The same multiplying element that creates a modulated signal at the transmitter recovers it at the receiver, which is why a balanced modulator and a synchronous detector are essentially the same circuit used in opposite directions.

Coherent recovery outperforms non-coherent, when a reference can be had. Demodulation divides into two camps. Non-coherent detectors, the envelope detector and the slope detector, are simple and need no knowledge of the carrier, but they pay in performance and cannot handle suppressed-carrier signals. Coherent detectors multiply the incoming signal by a locally regenerated carrier of the correct frequency and phase, recovering signals a non-coherent detector cannot touch and rejecting noise more effectively, at the cost of the circuitry, often a phase-locked loop, needed to acquire and track that reference. Much of the art of demodulation is deciding when the gain in performance justifies the cost of carrier recovery.

Frequency translation tames the receiver. Rather than build a high-performance receiver at every frequency a station might occupy, the superheterodyne principle converts whatever arrives to one fixed intermediate frequency, where a single set of optimized filters and amplifiers does the demanding work. This shift of the problem from many frequencies to one is among the most consequential ideas in the field, but it carries an unavoidable cost, the image frequency, a second input band that translates to the same IF and must be suppressed, a reminder that the multiplying operation at the core of conversion always produces more than the one product the designer wants.

Sampling is the bridge to the digital domain. Pulse modulation reveals that a continuous signal can be represented by samples taken often enough, the Nyquist criterion, without loss of information. This insight connects the analog modulation of this category to the digital communication that now dominates: a sampled, quantized, and coded signal is simply a pulse-modulated one carried to its logical end, and the sigma-delta converter that began as a pulse-modulation idea is today the standard route from analog reality into digital numbers.

Conclusion

Modulation and demodulation are how information is made to fit the medium that must carry it, and then recovered intact at the far end. Amplitude modulation circuits impress the message on the carrier's size and recover it with a detector as simple as a diode; frequency and phase modulation encode it in the carrier's timing and spend bandwidth to buy noise immunity; the mixer translates signals between frequencies and makes the superheterodyne receiver possible; and pulse modulation techniques replace the sine-wave carrier with samples, building the bridge to the digital domain. Across all four, every scheme is a deliberate trade among bandwidth, power, and noise, modulation is multiplication and multiplication makes sidebands, coherent recovery outperforms non-coherent when a reference can be had, and sampling is the path from the analog world into digital numbers. The subcategories above develop each in detail, and the related topics place these techniques within the wider practice of radio-frequency and signal-processing design.