Frequency Modulation

Frequency modulation uses the instantaneous amplitude of a modulating signal (voice, music, data, etc.) to directly vary the frequency of a carrier signal. Modulation index, β, is used to describe the ratio of maximum frequency deviation of the carrier to the maximum frequency deviation of the modulating signal. The concept was pioneered by Edwin H. Armstrong in the late 1920s and patented in the early 1930s.

Depending on the modulation index chosen, the carrier and certain sideband frequencies may actually be suppressed. Zero crossings of the Bessel functions, J_n(β), occur where the corresponding sideband, n, disappears for a given modulation index, β. The composite spectrum for a single tone consists of lines at the carrier and upper and lower sidebands (of opposite phase), with amplitudes determined by the Bessel function values at those frequencies.

Technical Details

Frequency modulation (FM) is a widely used technique in RF communication, where the frequency of a carrier wave is varied by a modulating signal to encode information. A distinctive feature of FM is the behavior of its carrier amplitude, which can be suppressed under specific conditions determined by the modulation index. This suppression is mathematically tied to Bessel functions of the first kind, which govern the spectral composition of an FM signal. Unlike amplitude modulation (AM), where carrier suppression is achieved through external circuitry, FM carrier suppression occurs naturally as a function of the modulation index. This treatise examines the mechanisms of FM carrier suppression, its reliance on the modulation index, the role of Bessel functions, and its implications for RF system design.

Fundamentals of Frequency Modulation

In FM, a carrier signal c(t) = A_c * cos(2 * π * f_c * t) with amplitude A_c and frequency f_c is modulated by a signal m(t) = A_m * cos(2 * π * f_m * t) with amplitude A_m and frequency f_m. The instantaneous frequency deviates from f_c proportionally to m(t), resulting in:

s_FM(t) = A_c * cos[2 * π * f_c * t + β * sin(2 * π * f_m * t)]

where β = Δf / f_m is the modulation index, Δf = k_f * A_m is the peak frequency deviation, and k_f is the frequency deviation constant (in Hz/V). The modulation index β quantifies the extent of frequency variation relative to the modulating frequency, playing a central role in the FM spectrum and carrier behavior.

Bessel Function Dependency

The FM signal's frequency spectrum is derived by expanding the phase-modulated term using Bessel functions of the first kind. The expression becomes:

s_FM(t) = A_c * Σ [n = -∞ to ∞] J_n(β) * cos[2 * π * (f_c + n * f_m) * t]

where J_n(β) is the Bessel function of the first kind, order n, with argument β. The spectrum consists of:

• Carrier: A_c * J₀(β) at f_c (n = 0).
• Sidebands: A_c * J_n(β) at f_c ± n * f_m (n = 1, 2, ...), appearing as infinite pairs symmetric around f_c.

Key properties of Bessel functions relevant to FM include:

• J₀(0) = 1, J_n(0) = 0 (n ≠ 0): At β = 0, only the carrier exists.
• J_n(-β) = (-1)ⁿ * J_n(β): Symmetry ensures sideband pairs balance.
• Σ [n = -∞ to ∞] J_n(β)² = 1: Total power remains A_c² / 2, redistributing across components as β increases.

The carrier amplitude A_c * J₀(β) varies with β, unlike AM, where it is fixed unless externally suppressed. This variation enables natural carrier suppression in FM, as detailed in the following table of Bessel function values at key modulation indices where carrier and sideband nulls occur (color-coordinated):
 

This table highlights carrier nulls (J₀ = 0) at β = 2.4048, 5.5201, and 8.6537, and first sideband nulls: J₁ at 3.8317, J₂ at 5.1356, J₃ at 6.3802, J₄ at 7.5883, and J₅ at 8.7715.

Carrier Suppression Mechanism

Carrier suppression in FM occurs when the carrier amplitude drops to zero, determined solely by the modulation index β. The Bessel function J₀(β,0) oscillates and crosses zero at specific values; i.e., the carrier vanishes:

• First zero: β ≈ 2.4048
• Second zero: β ≈ 5.5201
• Third zero: β ≈ 8.6537
• Subsequent zeros follow at intervals of approximately π (e.g., 11.7915, 14.9309).

The total power A_c² / 2 redistributes to sidebands, with higher-order terms (e.g., J₃, J₄) gaining prominence as β increases. Unlike AM's DSB-SC, where suppression requires a balanced modulator, FM achieves this intrinsically through β.

Modulation Index and Spectral Behavior

The modulation index β dictates the number and amplitude of significant sidebands:

• Narrowband FM (NBFM): β < 0.3
  • J₀(0.3) ≈ 0.978, J₁(0.3) ≈ 0.148, J₂(0.3) ≈ 0.011 - mostly carrier, one sideband pair.
  • Bandwidth ≈ 2 * f_m, similar to AM.
• Wideband FM (WBFM): β > 1
  • β = 5: J₀(5) ≈ 0.1776, J₁(5) ≈ 0.3276, J₂(5) ≈ 0.0466 - multiple sidebands.
  • Bandwidth ≈ 2 * (Δf + f_m) (Carson's rule).
• Suppression Point: β = 2.4048
  • Carrier null, sidebands carry all power.

As β rises, the spectrum widens, and the carrier's relative contribution diminishes, hitting zero at Bessel nulls. This contrasts with AM, where sidebands scale linearly with m, and carrier suppression requires external intervention.

Example 1: Narrowband FM Radio (Two-Way Communication)

A walkie-talkie operates at f_c = 146 MHz, f_m = 3 kHz (voice), Δf = 0.9 kHz:

• β = Δf / f_m = 0.9 / 3 = 0.3
• Spectrum:
  • Carrier: A_c * J₀(0.3) ≈ 0.989 * A_c
  • Sidebands: f_c ± 3 kHz, A_c * J₁(0.3) ≈ 0.148 * A_c
  • Higher orders negligible
• Bandwidth: 2 * (0.9 + 3) = 7.8 kHz
• Power: Carrier ≈ 97.8%, sidebands ≈ 2.2%

No suppression occurs - carrier dominates, keeping the signal compact for efficient VHF comms.

Mathematical Analysis

Consider a practical example: f_c = 100 MHz, f_m = 5 kHz, Δf = 12.024 kHz, so β = Δf / f_m = 2.4048.

• Signal: s_FM(t) = A_c * cos[2 * π * 10⁸ * t + 2.4048 * sin(2 * π * 5000 * t)]
• Spectrum:
  • f_c = 100 MHz: A_c * J₀(2.4048) = 0
  • f_c ± 5 kHz: A_c * J₁(2.4048) ≈ 0.5192 * A_c
  • f_c ± 10 kHz: A_c * J₂(2.4048) ≈ 0.4318 * A_c
• Power: (0.5192² + 0.4318² + ...) * A_c² = A_c² / 2 (sum over all n)

The carrier vanishes, and energy redistributes, verifiable via Σ J_n(β)² = 1.

Example 2: Wideband FM Broadcast with Suppression

An FM radio station at f_c = 98.1 MHz, f_m = 15 kHz (stereo audio), Δf = 36.072 kHz:

• β = Δf / f_m = 36.072 / 15 = 2.4048
• Spectrum:
  • Carrier: A_c * J₀(2.4048) = 0
  • f_c ± 15 kHz: A_c * J₁(2.4048) ≈ 0.5192 * A_c
  • f_c ± 30 kHz: A_c * J₂(2.4048) ≈ 0.4318 * A_c
  • Up to J₀(2.4048,5) ≈ 0.0164 * A_c
• Bandwidth: 2 * (36.072 + 15) ≈ 102 kHz
• Power: Carrier = 0%, sidebands = 100%

Full carrier suppression occurs—ideal for testing Bessel nulls, though broadcast FM typically uses β ≈ 5 for richer audio.

Practical Implications

1. Bandwidth Control

   At β = 2.4048, bandwidth ≈ 2 * (12.024 + 5) = 34 kHz, wider than AM's 10 kHz for the same f_m. Higher β (e.g., 5.5201) increases bandwidth further, complicating suppression via filtering. The following table illustrates this for key suppression points:

   β	Δf (kHz)	fm (kHz)	Bandwidth (kHz)	Carrier Power (%)	Sideband Power (%)
   2.4048	12.024	5	34.05	0.0	100.0
   5.5201	27.601	5	65.20	0.0	100.0
   8.6537	43.269	5	96.54	0.0	100.0

2. Signal Design

   Broadcast FM (e.g., Δf = 75 kHz, f_m = 15 kHz, β = 5) rarely hits exact nulls, but carrier weakening enhances sideband richness. Narrowband communications (e.g., β = 0.5) avoid suppression for simplicity.

3. Interference

   Nulling the carrier at f_c reduces interference with co-channel signals but spreads energy, risking adjacent channel overlap.

4. Detection

   FM receivers (e.g., PLLs) extract m(t) from phase, not carrier amplitude - suppression doesn't distort data, unlike AM.

Comparison to AM

• AM: Carrier suppression (DSB-SC) removes A_c * cos(2 * π * f_c * t) via multiplication, leaving sidebands. Bessel zeros (e.g., J₀(2.4048) = 0) apply only in overmodulation analysis.
• FM: Suppression is inherent - no circuitry needed. J₀(β) = 0 shifts power naturally, but sidebands can't be independently suppressed without losing m(t).

FM carrier suppression is a direct consequence of the modulation index β and its interaction with Bessel functions. At β values like 2.4048, 5.5201, and beyond, J₀(β) = 0, nulling the carrier and redistributing power to an expanding array of sidebands. This natural phenomenon, rooted in the mathematics of J_n(β), contrasts with AM's engineered suppression, offering unique advantages and challenges in RF design. For engineers, understanding this dependency enables precise control of FM spectra, balancing bandwidth, power distribution, and system performance in applications from broadcasting to telemetry.