How does a Butler matrix generate multiple beams?

A Butler matrix for N elements uses N/2 × log₂(N) hybrid couplers and fixed phase shifters arranged in log₂(N) stages. Each of the N input ports produces a unique linear phase gradient across the N output ports, which when connected to an N-element antenna array, forms a beam at a specific angle. For a 4×4 Butler matrix, 4 hybrid couplers and 2 fixed phase shifters create 4 simultaneous beams at angles of ±14.5° and ±48.6° from broadside (for half-wavelength element spacing). All beams exist simultaneously and are orthogonal, meaning they can carry independent signals without mutual interference.

What are the beam angles for a Butler matrix?

The beam angles for an N-element Butler matrix with half-wavelength element spacing are given by θ_n = arcsin((2n-1)/(N×2)) × λ/d, where n = ±1, ±2, ..., ±N/2. For a 4-element matrix (d = λ/2): beams at ±14.5° and ±48.6°. For an 8-element matrix: beams at ±7.2°, ±22.0°, ±38.7°, and ±61.0°. The beams are evenly spaced in sine space (not angle space), which means they crossover at the -3.9 dB points, providing continuous coverage without gaps. Increasing the element count narrows each beam and adds more beams to cover the same angular range.

How does a Butler matrix compare to digital beamforming?

A Butler matrix is purely passive and analog, requiring no power, no ADCs, and no digital processing. It generates all beams simultaneously with zero latency and handles the full antenna bandwidth. However, it produces only fixed beam positions and cannot adaptively null interferers or shape beam patterns. Digital beamforming digitizes each element's signal and computes arbitrary beam weights in software, offering unlimited beam steering, adaptive nulling, and per-beam amplitude control, but requiring N ADCs/DACs, significant processing power, and introducing quantization noise and processing latency. Many modern systems use a hybrid approach: a Butler matrix for coarse beam selection followed by digital fine-steering within each beam.

What is a Butler Matrix? | Beamforming Network Design

Definition & Architecture

A Butler matrix is a passive, reciprocal beamforming network that connects N input ports to N antenna elements through an arrangement of hybrid couplers and fixed phase shifters, producing N simultaneous, orthogonal beams pointing at predetermined angles. Invented by Jesse Butler in 1961, it is the RF analog of the Fast Fourier Transform (FFT): each input port produces a unique linear phase progression across the output ports, and the resulting beams are the spatial Fourier transform of these phase distributions.

An N-port Butler matrix requires N/2 × log₂(N) hybrid couplers (90° or 180°) and (N/2) × (log₂(N) − 1) fixed phase shifters, arranged in log₂(N) cascaded stages. For a 4×4 matrix: 4 hybrids and 2 phase shifters. For 8×8: 12 hybrids and 8 phase shifters. The matrix is lossless in theory (all input power reaches the antenna elements), making it far more efficient than a switched beam system using lossy power dividers and phase shifter networks.

Key Formulas

Beam Angles (N elements, d = λ/2):

θ_n = arcsin((2n − 1) / N) for n = ±1, ±2, ..., ±N/2

4×4: θ = ±14.5°, ±48.6°

8×8: θ = ±7.2°, ±22.0°, ±38.7°, ±61.0°

Component Count:

Hybrids = N/2 × log₂(N) | Phase shifters = N/2 × (log₂(N) − 1)

Beam Crossover Level: −3.9 dB (orthogonal beams in sine space)

Beamforming Network Comparison

Parameter	Butler Matrix	Blass Matrix	Rotman Lens	Digital BF
Beam Type	Fixed, orthogonal	Fixed, non-orthogonal	Fixed, wideband	Arbitrary, adaptive
Simultaneous Beams	N	M (M ≤ N)	M focal points	Unlimited
Bandwidth	10-20%	10-15%	Octave+	Full digitizer BW
Loss	Low (1-3 dB)	Moderate (3-6 dB)	Low (1-2 dB)	N/A (digital)
Complexity	Moderate	High (crossovers)	Low (planar)	Very High (N ADCs)
Adaptive Nulling	No	No	No	Yes
Power Required	None (passive)	None (passive)	None (passive)	High (DSP)
Typical Use	5G base station	Radar ESM	Wideband EW	AESA radar

Practical Application

A 5G massive MIMO base station uses an 8×8 Butler matrix at 3.5 GHz to create 8 fixed beams covering a 120° sector. Each beam has a 15° half-power beamwidth, and the beams cross over at −3.9 dB, providing continuous coverage. The Butler matrix is implemented in microstrip on a Rogers RO4003C substrate with eight broadside-coupled hybrid couplers and eight fixed-length delay lines as phase shifters. Total insertion loss is 1.8 dB. Users are assigned to beams based on their angular position, and the base station applies digital precoding within each beam for fine-grained multi-user MIMO. This hybrid analog-digital approach achieves 90% of full digital beamforming capacity with only 8 RF chains instead of 64.

Frequently Asked Questions

How does a Butler matrix generate beams?

N/2 × log₂(N) hybrids and fixed phase shifters create N unique linear phase gradients. Each input port maps to a beam angle. A 4×4 matrix uses 4 hybrids and 2 phase shifters for 4 simultaneous orthogonal beams.

What are the beam angles?

For d = λ/2: θ_n = arcsin((2n−1)/N). 4-element: ±14.5°, ±48.6°. 8-element: ±7.2°, ±22.0°, ±38.7°, ±61.0°. Beams cross at −3.9 dB for continuous coverage.

Butler matrix vs digital beamforming?

Butler: passive, zero latency, fixed beams, no adaptive nulling. Digital: arbitrary steering, adaptive nulling, but requires N ADCs and high processing power. Hybrid approach (Butler + digital precoding) balances performance and cost.

Butler Matrix