RF Fundamentals

Constructive Interference

Q: How much power gain does constructive interference give when combining N equal signals?

When N equal-amplitude signals add perfectly in phase, the voltages sum linearly so the total voltage is N times the single-element voltage. Power is proportional to voltage squared, so the combined power is N squared times the single-element power, an N-squared coherent gain. However, an N-element array driven from N independent amplifiers also delivers N times more total transmit power, so the antenna gain relative to a single element is N squared divided by N, equal to N, or 10 times log of N in dB. A 64-element array therefore provides about 18 dB of array gain over one element. The full N-squared scaling appears in the on-axis power density of the radiated beam, which is why a large coherent aperture concentrates energy so tightly along boresight.

Q: What phase error budget keeps a phased array near peak constructive combining?

The combined amplitude of two equal vectors is 2 times the cosine of half the phase difference, so a 30 degree phase error reduces the sum to cos(15 degrees) of the ideal, about 0.966, a loss of roughly 0.3 dB. Spread across a large array, random per-element phase errors of RMS value sigma reduce gain by approximately exp of minus sigma squared in radians. To hold beam-peak loss under about 0.5 dB, designers typically budget RMS phase error below 25 to 30 degrees, which is why 5 to 6 bit phase shifters with 5.6 degree least significant bits are common in millimeter-wave beamformers. Amplitude errors and element failures degrade sidelobes more than the main-beam peak.

Q: At what path length difference do two RF signals interfere constructively?

Two coherent signals add constructively when their path length difference is an integer number of wavelengths, delta r equals m times lambda for m equal to 0, 1, 2 and so on, which corresponds to a phase difference of an integer multiple of 360 degrees. At 28 GHz the free-space wavelength is about 10.7 mm, so a path difference of 10.7 mm, 21.4 mm, or 32.1 mm yields a constructive peak, while the half-wavelength points between them (5.35 mm, 16.05 mm) yield destructive nulls. In a multipath channel this spatial alternation produces standing-wave-like fading peaks and nulls spaced half a wavelength apart, which is the physical basis of small-scale fading and of why moving an antenna a few centimeters can recover a dropped link.

/kuhn-STRUHK-tiv in-ter-FEER-uhns/

When two or more coherent waves arrive with a phase difference near an integer multiple of 360°, their instantaneous fields add and the resultant amplitude rises toward the arithmetic sum of the contributors. For two equal RF tones this doubles the resultant voltage and quadruples the power density (both a +6 dB increase relative to one tone), and combining N equal signals yields an on-axis power density of N² times one element. This in-phase addition is the working principle behind phased-array beam formation, antenna directivity, and the bright peaks of multipath fading. The opposite case, where waves arrive 180° out of phase and cancel, is destructive interference.

Category: RF Fundamentals

Peak Condition: Δφ = m × 360°

N-element Gain: N² on-axis

How In-Phase Waves Reinforce

Interference is a direct consequence of the linear superposition principle for electromagnetic fields: at any point in space and time, the total field is the vector sum of the contributing fields. When two sinusoids of the same frequency meet, the magnitude of their sum depends entirely on the phase difference Δφ between them. At Δφ = 0 the crests align with crests and troughs with troughs, so the amplitudes add directly. As Δφ grows toward 180°, the sum shrinks, reaching complete cancellation for two equal waves. Constructive interference is simply the regime where Δφ is close enough to a multiple of 360° that the resultant exceeds either contributor.

In RF systems that phase difference comes from a path-length difference, an electrical delay, or a deliberate phase shift. Two coherent rays that travel paths differing by Δr reinforce when Δr equals an integer number of wavelengths. At 28 GHz, where the free-space wavelength is roughly 10.7 mm, a path difference of one wavelength produces a constructive peak while the half-wavelength point between adjacent peaks produces a deep null. This is why a stationary receiver in a reflective environment can sit on a peak or a fade depending on a few millimeters of position, and why diversity antennas are spaced to decorrelate those peaks.

The engineering payoff appears when reinforcement is made deliberate. Feeding an antenna aperture or an array so every element radiates in phase toward a chosen direction forces constructive addition along that bearing and partial cancellation elsewhere, concentrating energy into a narrow main beam. The same coherent combining at the receive side recovers weak signals from noise. Getting the phases right, and holding them stable across temperature and frequency, is the central design problem of beamforming hardware.

Superposition and the N-Squared Law

For two equal-amplitude waves the resultant follows a half-angle cosine: the sum amplitude is 2A·cos(Δφ/2). A 30° phase error therefore costs only cos(15°) ≈ 0.966 of the ideal voltage, about 0.3 dB, which is why finite-resolution phase shifters still combine efficiently. Scaling to N coherent elements, the on-axis voltage grows as N and the power density as N², while the array gain relative to one element (accounting for the N independent power sources) scales as N.

Distinguishing Array Gain From Aperture Power

A common source of confusion is whether N elements buy you N or N². The radiated on-boresight power density does scale as N² because the fields add coherently, but a transmit array also injects N times more total power than one element. The directivity (and the link-budget gain) is the ratio, N. Constructive interference accounts for the full N² field concentration; the extra factor of N over directivity is just the added source power.

Governing Relations

Two-wave superposition (equal amplitude A):
A_sum = 2A × cos(Δφ / 2)  →  maximum 2A at Δφ = 0, 2π, 4π

Constructive path-length condition:
Δr = mλ, Δφ = (2π / λ) × Δr = 2πm  (m = 0, 1, 2, …)

Coherent N-element combining:
V_tot = N × V₁, P_density ∝ N²P₁, Array gain ≈ N (linear) = 10·log₁₀(N) dB

Random phase-error gain loss:
G / G_ideal ≈ e^−σ_φ²  (σ_φ in radians)

Example: 64 elements → 10·log₁₀(64) ≈ 18 dB array gain; σ_φ = 25° (0.44 rad) → ≈ 0.8 dB loss.

Phase Difference vs. Combined Output

Phase Δφ	Path diff (28 GHz)	Sum of two equal waves	Power vs. one wave	Regime
0°	0 mm	2.00 A	+6.0 dB	Full constructive
30°	0.89 mm	1.93 A	+5.7 dB	Near constructive
60°	1.79 mm	1.73 A	+4.8 dB	Partial constructive
90°	2.68 mm	1.41 A	+3.0 dB	Quadrature (neutral)
120°	3.57 mm	1.00 A	0.0 dB	Partial destructive
180°	5.35 mm	0.00 A	−∞ dB (null)	Full destructive

Common Questions

Frequently Asked Questions

How much power gain does constructive interference give when combining N equal signals?

The in-phase voltages sum linearly, so total voltage is N×V₁ and on-axis power density scales as N². Because a transmit array also injects N times more total power, the directivity (link-budget) gain over one element is the ratio, N, or 10·log₁₀(N) dB. A 64-element array gives about 18 dB of array gain, while the field concentration on boresight is the full N².

What phase error budget keeps a phased array near peak constructive combining?

Two equal vectors sum to 2A·cos(Δφ/2), so a 30° error costs only about 0.3 dB. Across an array, random RMS phase error σ_φ reduces gain by roughly e^−σ². To hold beam-peak loss under ~0.5 dB, designers budget σ_φ below 25 to 30°, which is why 5 to 6 bit phase shifters with 5.6° LSBs are common in millimeter-wave beamformers.

At what path length difference do two RF signals interfere constructively?

Reinforcement occurs when the path difference is an integer number of wavelengths, Δr = mλ, equal to a phase difference of m×360°. At 28 GHz (λ ≈ 10.7 mm), peaks fall at 10.7, 21.4, and 32.1 mm while nulls sit half a wavelength between them. In a multipath channel this produces fading peaks and nulls spaced λ/2 apart, the basis of small-scale fading.

Related Terms

← Constrained Lens Contact →