Clever Waveguide Trick Lets a Flat Antenna Array Fire Five Beams From Just Two Feeds

Engineers at the University of Quebec in Outaouais have built and tested a planar antenna that solves one of the quiet headaches of modern radio hardware: how to feed a large, highly directional array without burying the circuit board under a thicket of power dividers, coaxial routing, and precision-drilled metal walls. Their answer, published in Frontiers in Antennas and Propagation, is a five-element Yagi-Uda array driven by a substrate integrated waveguide operating in an unusually high-order mode, the TE50 mode, at 5.8 gigahertz. The fabricated prototype delivered a peak realized gain of 10.2 dBi with a measured radiation efficiency of 87.2 percent, and its measured behavior tracked the simulations closely enough to validate the entire concept from theory to anechoic chamber.

The device belongs to a family of structures known as substrate integrated waveguides, or SIWs. A conventional waveguide is a hollow metal pipe that guides electromagnetic energy with very low loss, but it is bulky, expensive, and awkward to merge with printed circuit boards. SIW technology, first formalized by Deslandes and Wu in the early 2000s, recreates the electromagnetic behavior of that metal pipe inside a flat dielectric substrate by stitching rows of metallized via holes between the top and bottom copper planes. The result is a waveguide that can be fabricated with the same printed circuit board processes used for any ordinary electronics, offering low profile, low weight, low cost, and seamless integration with microstrip circuitry.

Most SIW designs, however, operate in the fundamental mode, the simplest field pattern the structure supports, which has a single electric field maximum across the waveguide’s width. That simplicity becomes a liability when the goal is to illuminate a wide radiating aperture. To feed a five-element Yagi array from a fundamental-mode waveguide, a designer would need a multi-stage corporate power divider network, splitting the signal again and again to reach each element. Every extra splitter adds dielectric and conductor losses, introduces phase errors between branches, and complicates the layout. At higher frequencies the problem worsens, because dense via arrays with small diameters invite localized phase errors, radiation leakage, and drill overlap defects within the tight tolerances of standard PCB and photolithography processes.

The Canadian team’s insight was to skip the splitter entirely by exploiting a higher-order mode. In a waveguide whose width is scaled up, the electromagnetic field can arrange itself into multiple transverse maxima, much like the harmonics of a vibrating string. The TE10 mode has one field peak, the TE30 mode has three, and the TE50 mode has five, arranged with alternating polarity across the cavity width. The researchers quantified this trade-off carefully: a standard guide 25.4 millimeters wide offers broader fractional bandwidth but demands a complex feed; an intermediate 76.2-millimeter guide gives three peaks; and their proposed 127-millimeter-wide guide naturally establishes five spatially distributed electric field maxima, providing a direct five-point excitation for the radiating elements. The oversized cavity also reduced the conductor attenuation constant to 0.78 dB per meter, compared with 1.85 dB per meter for the fundamental-mode guide, because the wider structure carries the same power with lower current density on its walls.

Exciting a high-order mode cleanly is the hard part. An overmoded cavity supports many possible field patterns simultaneously, and any asymmetry in the feed can excite unwanted parasitic modes that scramble the radiation pattern. The team’s solution is an elegant two-step conversion. A microstrip T-type power divider on the upper substrate splits the incoming signal into two equal-amplitude, equal-phase branches, with tuning stubs compensating for reactive discontinuities at the junction. Each branch then feeds one of two longitudinal coupling slots placed symmetrically on the middle metal layer, spaced at exactly one-fifth of the SIW width. Rather than acting as five independent feeds, these two slots act as discrete spatial excitation sources positioned at the phase centers of the target mode’s electric field antinodes.

Once the energy couples into the SIW cavity, spatial modal synthesis takes over. The resonant boundary conditions imposed by the via walls enforce constructive interference at 5.8 gigahertz, transforming the two slot feeds into the required five-peak field profile with alternating polarities of plus, minus, plus, minus, plus. Simulated vertical electric field distributions confirmed exactly five distinct intensity peaks along the waveguide width, indicating high excitation purity and strong dominance of the desired TE50 mode over the lower-order TE10 and TE30 alternatives. Cutoff-frequency analysis using standard SIW equivalent-width formulas further supported the identification, since the simulated field pattern matched the characteristic five-maxima signature of the target mode and nothing else.

The radiating structure itself is a classic Yagi-Uda array, the venerable Japanese-invented design that remains a workhorse of directional communications nearly a century after its debut. The printed version was realized on Rogers RT/duroid 5880 substrate, a low-loss laminate with a relative permittivity of 2.2, a thickness of 1.575 millimeters, and a dielectric loss tangent of just 0.0018, metallized with 35-micrometer copper. A driven element is backed by a reflector and followed by five directors, with the first director placed 4 millimeters from the feed and a uniform 2.5-millimeter spacing between subsequent directors. The directors progressively channel the radiated energy into a narrow forward beam, and the five-point modal excitation ensures that all elements launch their surface waves with balanced amplitude and phase.

Design proceeded through a systematic four-step methodology. First, the SIW cavity width was calculated to support the TE50 mode at the target center frequency. Second, the T-divider arms and dual-slot geometry were optimized for strong spatial coupling and equal amplitude and phase excitation. Third, the Yagi element lengths and spacings were derived from surface-wave launch conditions. Finally, multi-objective parametric sweeps in the full-wave simulator ANSYS HFSS maximized gain and front-to-back ratio while keeping the reflection coefficient below minus 10 dB. The parametric study examined how the feed slot width governs coupling between the microstrip line and the radiating elements, how the via diameter sets the boundary conditions of the propagation structure, and how the slot length controls the electrical resonant path, with each parameter tuned to center the impedance match at 5.8 gigahertz.

The fabricated prototype then faced the ultimate test: measurement in an anechoic chamber. The reflection coefficient remained below minus 10 dB from 5.7 to 5.9 gigahertz, a fractional impedance bandwidth of 3.4 percent that fully covers the 5.8 GHz industrial, scientific, and medical band used by Wi-Fi, short-range radar, and industrial wireless systems. The peak realized gain reached 10.2 dBi at 5.8 gigahertz, and the radiation patterns showed a stable, broad main lobe with low side lobes in both planes, with 3 dB beamwidths of approximately 85 degrees in the E-plane and 90 degrees in the H-plane. Measured and simulated results agreed well, with only slight deviations attributed to fabrication tolerances and measurement uncertainties.

Benchmarked against recent state-of-the-art SIW-fed arrays and higher-order-mode antennas, the design holds its own while offering a structural advantage that matters in production. Compared with fundamental and intermediate higher-order-mode designs, it achieves competitive gain within a single-layer substrate, and by eliminating the dense internal via walls typically required for modal isolation or multi-layer stacking, it substantially relaxes fabrication tolerances and reduces cost. In an era when sub-6 GHz wireless links, radar sensors, and connected industrial equipment all crave compact antennas that beam energy where it is needed, a feed network that turns two slots into five perfectly balanced excitations is the kind of quiet engineering elegance that could spread quickly through the antenna design community. The authors indicate that future work will pursue bandwidth extension and array scalability, suggesting this TE50 approach may soon grow into even larger planar apertures.

Subject of Research: Design and experimental validation of a TE50-mode substrate integrated waveguide-fed Yagi-Uda antenna array operating at 5.8 GHz

Article Title: Design and fabrication of a TE50 mode SIW-fed Yagi antenna array at 5.8 GHz

Article References: Omrani, R., Louati, S., & Boutayeb, H. (2026). Design and fabrication of a TE50 mode SIW-fed Yagi antenna array at 5.8 GHz. Frontiers in Antennas and Propagation, 4, Article 1932708. https://doi.org/10.3389/fanpr.2026.1932708

Image Credits: AI Generated

DOI: 10.3389/fanpr.2026.1932708

Keywords: substrate integrated waveguide, TE50 mode, Yagi-Uda antenna array, 5.8 GHz, higher-order modes, mode conversion, directive radiation, microstrip power divider, antenna gain, PCB fabrication, ISM band, anechoic chamber measurement