Frequency Modulation and Demodulation: Working Principle, Waveforms, Circuits and Applications

Author : scientech scientech | Published On : 18 Nov 2026

 

Frequency modulation and demodulation describe how information is carried by the frequency of a high-frequency wave and then recovered. In frequency modulation (FM), the instantaneous frequency of a carrier varies in proportion to the message signal, while the carrier amplitude stays fixed. In frequency demodulation, a receiver circuit turns those frequency variations back into a voltage that follows the original message.

The technique was developed into a practical broadcasting system by the American engineer Edwin Armstrong, and wideband FM is now used worldwide to transmit high-fidelity sound. Its main strength is resistance to amplitude noise. Its main cost is bandwidth, since an FM signal occupies far more spectrum than an AM signal carrying the same audio.

This guide builds the subject from the ground up. It starts with why modulation is needed, derives the FM equation, explains frequency deviation and modulation index, and then moves to the circuits that generate and detect FM. It also walks through the frequency demodulation waveform at each stage, the standard frequency modulation and demodulation experiment performed on a trainer kit, four worked numericals, and the way the topic appears in Indian competitive exams.

How Frequency Modulation Works

The easiest way to understand FM is to start without equations. With no message, the carrier oscillates at its centre frequency, fc. When the message goes positive, the carrier frequency rises above fc. When the message goes negative, the carrier frequency falls below fc. A larger message amplitude pushes the frequency further from fc. A faster message makes the frequency swing back and forth more often. The amplitude of the carrier never changes.

The FM Equation

Let m(t) be the message, Ac the carrier amplitude, fc the carrier frequency, and kf the frequency sensitivity of the modulator in Hz per volt. The instantaneous frequency is

fi(t) = fc + kf · m(t)

Frequency is the rate of change of phase, so the phase must be found by integrating the frequency. The FM wave is therefore

s(t) = Ac cos[ 2π fc t + 2π kf ∫ m(τ) dτ ]

For the common single-tone case, m(t) = Am cos(2π fm t), where Am is the message amplitude and fm is the message frequency. The integral gives a sine term, and the equation becomes

s(t) = Ac cos[ 2π fc t + β sin(2π fm t) ]

Here β is the modulation index, defined in the next section. This is the standard form used in almost every numerical, so it is worth memorising.

Frequency Deviation and Modulation Index

The frequency deviation, Δf, is the maximum amount by which the carrier frequency departs from fc. For a single tone,

Δf = kf · Am

The modulation index of FM is the ratio of the deviation to the message frequency,

β = Δf / fm

Two points here cause most of the mistakes in exams. First, the deviation depends only on the message amplitude. Doubling the message frequency does not change Δf. Second, unlike AM, the FM modulation index is not limited to 1. A broadcast FM signal with 75 kHz deviation and a 15 kHz audio tone has β = 5, and values of 20 or more are routine in problems.

The Frequency Modulation Waveform on an Oscilloscope

On a dual-trace oscilloscope, trigger on the message and display the FM signal on the second channel. Three features stand out. The envelope is flat, because the amplitude is constant. The cycles bunch together at the positive peaks of the message, where the frequency is highest, and spread out at the negative peaks, where it is lowest. When the message crosses zero, the cycles return to the spacing of the unmodulated carrier.

At low message amplitudes the change in spacing is hard to see. Raising the amplitude on a kit makes the crowding obvious, and that is the quickest visual proof that deviation follows amplitude. The frequency modulation waveform also looks nothing like an AM waveform, which is why a simple diode envelope detector cannot recover FM.

FM Bandwidth: Narrowband, Wideband and Carson's Rule

An FM signal with a pure sinusoidal message does not contain just one sideband pair. It contains the carrier and an infinite series of sidebands spaced at multiples of fm, with amplitudes given by Bessel functions of the first kind. How many of them matter depends on β.

When β is very small, only the first pair of sidebands is significant, and the spectrum resembles AM with a bandwidth of about 2fm. This is narrowband FM (NBFM), used in two-way radio. When β is large, many sidebands carry meaningful power and the bandwidth grows. This is wideband FM (WBFM), used in broadcasting.

Carson's rule, published by J. R. Carson in 1922, gives a quick estimate that holds for both cases:

BT = 2(Δf + fm)

It captures roughly 98 percent of the signal power. For broadcast FM, with Δf = 75 kHz and a highest audio frequency of 15 kHz, the rule gives 2(75 + 15) = 180 kHz, and the deviation ratio is 75/15 = 5. In many countries each station is allotted a 200 kHz channel, which leaves a small guard margin. The ITU recommendation on FM sound broadcasting lists ±75 kHz as the maximum deviation (with ±50 kHz used in a few regions), and the Indian FM band runs from 88 to 108 MHz according to TRAI.

If Δf is much larger than fm, the bandwidth is roughly 2Δf, which is set by the message amplitude and not by how fast the message varies.

The Frequency Modulator: How an FM Signal Is Generated

A frequency modulator is any circuit whose output frequency can be steered by a voltage. There are two families. Direct methods vary the frequency of an oscillator. Indirect methods, such as the Armstrong system, build a narrowband phase-modulated signal first and then multiply the frequency to widen the deviation. Direct methods are simpler and are what a typical frequency modulation kit demonstrates.

Varactor Modulator

An LC oscillator runs at f = 1 / (2π√(LC)). A varactor is a diode whose junction capacitance changes with reverse bias. Connect it across the tank and let the message vary the bias, and the total capacitance moves, which moves the oscillator frequency. Because f depends on C to the power of minus one half, a small relative change in capacitance gives a change in frequency of about half that fraction, with the opposite sign.

Reactance Modulator

A reactance modulator uses a transistor with a phase-shifting network so that, from the tank's point of view, the device looks like a variable capacitor or inductor. The message changes the transistor's transconductance, which changes that effective reactance, and the oscillator frequency follows.

VCO-Based Modulator

A voltage-controlled oscillator (VCO) has a frequency that depends directly on its control voltage. ICs such as the XR2206 function generator can act as a VCO, so feeding the message to the control input produces FM with no tuned circuit to align. The free-running frequency sets fc, and the voltage sensitivity of the chip sets kf. This block is also the easiest place to measure deviation and modulation index in the lab, because it can be found by applying DC levels and noting the frequency change.

Frequency Demodulation: How the Message Is Recovered

All frequency demodulation methods rely on one idea. Build a circuit whose output voltage is proportional to the input frequency. If the input frequency moves up and down with the message, the output voltage moves up and down with it. The graph of output voltage against input frequency is called the discriminator characteristic or S-curve. It is zero at fc, rises on one side and falls on the other, and it should be as linear as possible across the deviation range.

Real FM signals also pick up amplitude noise, so a good receiver places an amplitude limiter ahead of the detector. The limiter clips the signal to constant amplitude and a bandpass filter removes the harmonics, which strips the noise off the envelope while leaving the frequency information untouched.

Slope Detector (Detuned Resonant Detector)

The simplest FM demodulator is a tuned circuit set slightly off the carrier frequency, followed by a diode envelope detector. As the FM signal swings up and down in frequency, it slides along the slope of the tuned circuit's response. The amplitude of the output therefore rises and falls with the frequency. The FM has been converted into AM, and the envelope detector finishes the job.

It works, but the slope is not linear and the circuit is sensitive to amplitude changes, so it is used mainly for teaching. A balanced version uses two tuned circuits, one above and one below fc, with their outputs subtracted. That gives a more linear S-curve.

Foster-Seeley Discriminator and Ratio Detector

The Foster-Seeley discriminator uses a tuned transformer in which the phase relationship between the primary and secondary voltages changes with frequency. At fc the two diode outputs balance and the result is zero. Above or below fc the balance tips and the output goes positive or negative. It needs a limiter, because it responds to amplitude changes as well as frequency changes.

The ratio detector is a modified version in which one diode is reversed and a large capacitor holds the total output voltage steady. The output depends on the ratio of the two diode voltages and not on their sum, so amplitude variations are rejected without a separate limiter. The price is a smaller output and a wound component that costs more, which is why it has largely disappeared from modern designs.

Quadrature Detector

In a quadrature detector, the incoming FM signal goes two ways. One path goes straight to a multiplier. The other passes through a phase-shift network, usually a small capacitor with a tuned LC circuit, that delivers exactly 90 degrees at fc and a phase that shifts away from 90 degrees as the frequency changes. The multiplier output, after low-pass filtering, is proportional to the phase difference and therefore to the frequency deviation. Because it needs very little more than a multiplier, it is easy to place inside an IC.

PLL Demodulator

A phase-locked loop contains a phase detector, a loop filter and a VCO connected in a feedback loop. The phase detector compares the incoming FM signal with the VCO output. Its error signal is smoothed by the low-pass loop filter and fed back as the VCO control voltage. If the VCO lags the input, the voltage rises and speeds it up. If it leads, the voltage falls and slows it down.

When the loop is locked, the VCO tracks the instantaneous frequency of the input. To follow an FM signal, the control voltage has to move in step with the message, which means the control voltage itself is the demodulated output. In equation form, the control voltage is x(t) ≈ (kf / kvco) · m(t), where kf is the deviation constant of the incoming signal and kvco is the VCO sensitivity in Hz per volt.

The loop filter limits the control voltage to roughly the message bandwidth, which narrows the noise bandwidth. That gives the PLL a lower noise threshold than a slope or quadrature detector, and it needs few adjustments. Common ICs used in labs are the LM565 and the 4046 family.

Frequency Modulation and Demodulation Experiment on a Trainer Kit

A frequency modulation and demodulation trainer kit packages the modulator, demodulator, audio source and test points on one board, so students can generate and recover FM without building circuits from scratch. Lab manuals and kit catalogues often label the topic simply as frequency modulation demodulation, but the content is the same pair of processes. Kits differ in the details. Some use a 45 kHz or 100 kHz carrier with an XR2206 modulator and a PLL demodulator, and others use a 455 kHz carrier with varactor and reactance modulators. Always read the manual for your own board before choosing settings.

One widely used Indian example is the Scientech 2203 Frequency Modulation and Demodulation Techniques trainer. According to the manufacturer's datasheet, it has an on-board audio oscillator covering 300 Hz to 3.4 kHz with an output up to 10 Vpp. It provides three FM modulators (reactance and varactor types at 455 kHz, and an XR2206-based VCO with an adjustable carrier from 10 kHz to 200 kHz) and six demodulators: detuned resonant, quadrature, Foster-Seeley, ratio, and two PLL detectors based on the HEF4046 and LM565. A mixer lets noise be added to the FM signal before demodulation, and the board adds more than 35 test points and 12 switched faults for troubleshooting. That block-by-block layout maps almost exactly onto the theory above.

Aim and Apparatus

The aim is to generate an FM signal, measure its frequency deviation and modulation index, and recover the message with one or more demodulators. The apparatus is the trainer kit with its power supply, a dual-channel oscilloscope, patch cords and a frequency counter if the kit does not display frequency directly.

Procedure

Start with the modulator and set the audio oscillator to a convenient tone such as 1 kHz at low amplitude. Observe the unmodulated carrier first and note its frequency. This is fc.

Apply the message to the modulator input. Trigger the oscilloscope on the message and display the FM output on the second channel. Slowly raise the message amplitude and watch the cycles crowd and spread. Note that the envelope height does not change.

To find the sensitivity, apply a few DC levels to the VCO input, record the carrier frequency for each, and compute kf in Hz per volt from the slope. The deviation for any message amplitude is then Δf = kf · Am, and the modulation index is β = Δf / fm.

Connect the FM output to a demodulator, starting with the PLL. Observe the detector output and then the low-pass filtered output, and compare it with the original message in frequency and shape. Repeat with the quadrature and Foster-Seeley detectors and compare linearity and amplitude. Finally, inject noise through the mixer and see which detectors cope with it better, with and without the limiter.

Observation Table

The table below shows how to record and calculate results. The values are illustrative, assuming kf = 5 kHz/V and fm = 1 kHz, and are meant to show the method. Measured numbers from a real kit will differ.

Am (V)

Δf = kf·Am (kHz)

β = Δf/fm

Carson bandwidth 2(Δf + fm) (kHz)

1

5

5

12

2

10

10

22

4

20

20

42

With a spectrum analyser, the Bessel null method gives a second check. For a single tone, the carrier component vanishes from the spectrum at β of about 2.41, so the amplitude that produces that null gives Δf = 2.41 fm.

Applications of FM and Frequency Demodulation

FM broadcasting is the most visible application. In India, analog FM sits in the 88 to 108 MHz band. Akashvani (All India Radio) runs services such as FM Rainbow, FM Gold and Vividh Bharati, and TRAI's 2025 recommendations report that AIR has 591 stations with 754 transmitters, while 388 private FM radio channels operate across 113 cities. Planning margins go well beyond Carson's rule. TRAI's 2012 recommendation noted that adjacent carriers in a licence area were generally kept at least 800 kHz apart and proposed reducing that to 400 kHz, so current rules should be checked on the TRAI site.

Two-way and land mobile radio use narrowband FM. Wikipedia's summary of common deviations lists ±5 kHz for 15 to 25 kHz channel spacing and ±2.5 kHz for 3.75 to 12.5 kHz spacing, which keeps channels narrow and still gives acceptable voice quality.

The digital relative of FM is frequency shift keying (FSK), where the carrier jumps between discrete frequencies to represent bits. Frequency-modulated continuous-wave (FMCW) radar sweeps the carrier frequency linearly and measures distance from the frequency difference between the transmitted and returned signals. All of them need a frequency modulator at one end and a frequency demodulator at the other, so the circuits studied on a trainer kit have direct counterparts in practice.

Frequency division multiplexing is the idea that lets many FM stations share the same band, each on its own carrier. It is covered in a separate post on this blog about FDM.

Conclusion

Frequency modulation and demodulation follow a clean logic. The modulator turns message amplitude into frequency deviation, the channel carries a constant-amplitude wave that is largely immune to amplitude noise, and the demodulator turns frequency back into voltage. Everything else, from Carson's rule to the choice between a quadrature detector and a PLL, follows from that chain. The cost is bandwidth, and every FM system trades it for cleaner sound.

For exams, the high-value skills are reading β, Δf and fm from an equation, applying Carson's rule, and knowing how each parameter responds when amplitude or frequency changes. For lab work, they are measuring kf, setting a sensible message level, and following the frequency demodulation waveform from limiter to filter. Students who want to build that hands-on habit can work through a frequency modulation kit such as the Scientech 2203 using its labelled test points, and then check each result against the numericals above. Before sitting any exam, confirm the syllabus on the official GATE and recruitment notices for the current cycle.

FAQs

Q: What is the difference between frequency modulation and demodulation?

Frequency modulation is the transmitter-side process in which the message varies the carrier frequency. Demodulation is the receiver-side process that recovers the message from those frequency variations. In frequency modulation and demodulation as a pair, the modulator turns voltage into frequency, and the demodulator turns frequency back into voltage.

Q: Why is FM less affected by noise than AM?

Most noise adds to the amplitude of a signal. In FM, the information is carried by the frequency, so a limiter can clip away amplitude variations without disturbing the message. FM also gains in signal-to-noise ratio as the deviation increases, provided the carrier stays above the noise threshold, and this is paid for with extra bandwidth.

Q: Which frequency demodulation method is best?

There is no single best choice, but the PLL and quadrature detectors dominate modern integrated receivers because they are easy to build into an IC and need little adjustment. A PLL also tolerates weaker signals than most other detectors. Older Foster-Seeley and ratio detectors appear mostly in discrete designs and in exam questions.

Q: How does a PLL demodulate FM?

The PLL forces its VCO to follow the phase of the incoming FM signal. To do so, the control voltage must change in step with the instantaneous frequency of the input. That control voltage is therefore proportional to the message, and a low-pass filter delivers it as the demodulator output.