Octagon, Square or Circle? Small Receive-Loop Geometry
Octagon, Square or Circle? Small Receive-Loop Geometry
A regular octagon encloses almost as much area as a circle made from the same conductor length, while keeping straight, measurable facets. That is a useful engineering compromise—not proof that eight sides produce better reception.
Why choose an octagon when a square is simpler and a circle encloses the most area? Start by stating the constraint. At equal perimeter, the octagon makes a large geometric step beyond the square and stops just short of the circle. Whether that matters at the receiver depends on the complete loop, front end, feedline and site.
Short answer: use a circle when maximum enclosed area per conductor length is the overriding constraint. Use a square when minimum part count or a rectangular mounting space dominates. Use an octagon when near-circular area and straight-part repeatability are worth four extra corners. None of those choices guarantees more SNR, a deeper null or better common-mode rejection.
Make the Shape Comparison Fair
Shape rankings change when the constraint changes. Equal perimeter is a useful comparison when conductor length, approximate mass and material use are being held broadly similar. For a regular polygon with n sides and perimeter P, its enclosed area is:
An = P2 / [4n tan(π/n)]A circle with the same perimeter encloses Acircle = P2 / 4π. The resulting geometric comparison is exact for ideal regular shapes:
| Shape | Area relative to circle | Area relative to square | What the number says |
|---|---|---|---|
| Circle | 100% | 127.3% | Maximum area for a given perimeter |
| Regular octagon | 94.8% | 120.7% | Near-circular area from eight straight sides |
| Square | 78.5% | 100% | Fewer sides and corners, with less enclosed area |
That 20.7% octagon-over-square result is not universal. If maximum width, height, bounding-box area, available tubing lengths or mounting points are held fixed instead, the dimensions and sometimes the ranking change. Always name the constraint before declaring a winner.
Area Sets One Small-Loop Voltage Term
For an electrically small loop in a magnetic field that is approximately uniform across its area, Faraday's law gives the open-circuit induced voltage magnitude approximately as:
|Voc| ≈ ω N A |B⊥|ω is angular frequency in radians per second, N is the number of turns, A is enclosed area in square metres and B⊥ is the magnetic-flux density normal to the loop. Under those stated conditions, induced voltage scales with area.
At equal perimeter, an ideal regular octagon therefore has 1.207 times the open-circuit voltage of a square in the same uniform field. Expressed as a voltage ratio, that is about 1.63 dB. It is about 0.46 dB below the equal-perimeter circle. These are geometry calculations—not measured receiver-SNR results.
More voltage is not automatically more SNR: wanted signals and external radio noise can both scale with effective area. Front-end noise, impedance, loss, strong-signal behaviour, balance and feedline pickup decide whether the geometric margin remains useful.
The Approximation Has Limits
The simple area law is most useful while the loop is electrically small and its current and incident field can be treated with a compact approximation. As perimeter becomes a larger fraction of wavelength, current magnitude and phase vary around the conductor. Parasitic capacitance, loop inductance, resonance, shield discontinuities, joints, loading and the amplifier input then become increasingly important.
At that point, comparing outlines alone is not enough. Model the complete conductors and feed structure, converge the mesh, and verify complex impedance, transfer response and pattern with measurements at declared reference planes.
Why Eight Sides Can Be a Sensible Stop
The geometric return diminishes as more sides are added. A regular 16-sided polygon encloses about 98.7% of the equal-perimeter circle's area—only about 4.1% more area than the octagon—while doubling the number of straight sections and angles.
That makes eight sides a defensible trade when a build needs straight members and known reference points. Equal facets can be cut and checked directly; opposite sides and vertices provide repeatable alignment references; and modular sections may simplify transport and service. Those are manufacturing and mechanical possibilities, not intrinsic RF performance claims.
Every extra joint is also a liability to control. Contact resistance, shield continuity, corrosion, water paths, alignment and structural loading can erase a small geometric advantage. A careful square or circle can outperform a badly assembled octagon.
What Side Count Does Not Guarantee
| Claim | Engineering boundary |
|---|---|
| An octagon encloses more area than a square | Yes at equal perimeter: 20.7% more for ideal regular shapes. |
| An octagon is more efficient than a circle | No geometric basis. The circle encloses more area at equal perimeter; realised efficiency depends on the complete system. |
| Eight sides create a deeper null | No. Null depth and direction depend on current balance, feedline common mode, nearby objects, arrival field and multipath. |
| Eight sides create higher CMRR | No. CMRR belongs to the complete differential system, including the sensor, front end, shield treatment, cables and installation. |
| An octagon is inherently wider-band | No. Transfer response depends on inductance, capacitance, loading, feedback, filtering and the receiving electronics. |
| A square loses current at its corners | Not merely because its bends are 90 degrees. Conductor dimensions, joints, shield continuity and loading need evidence. |
| Eight sides create eight pattern lobes | No. An electrically small planar loop retains the familiar magnetic-dipole pattern; polygon side count is not lobe count. |
Shielding and Balance Are Separate Questions
A shielded loop can reduce direct capacitive coupling into the inner receiving conductor when its shield topology and discontinuity are designed correctly. It does not become “magnetic only,” and an outline does not prove shielding performance.
Likewise, visual symmetry is not measured electrical balance. Unequal capacitance to the mast, enclosure, cable, ground or nearby conductors can convert common-mode energy into the differential signal that the receiver sees. A regular polygon may make dimensions easier to reproduce, but CMRR and null depth still require complete-system tests.
Compare Finished Antennas, Not Drawings
- Declare the constraint: equal perimeter, equal maximum width, equal height, equal area, equal mass or equal mounting envelope answer different questions.
- Record actual geometry: include conductor diameter, joints, shield gaps, feed structure, enclosure, support and cable route.
- Use a common reference plane: compare complex impedance and transfer response at the same calibrated terminals.
- Control the receiver: keep gain, attenuation, bandwidth, AGC and overload state fixed or recorded.
- Measure common-mode current: an apparent shape advantage can actually be a feedline or control-cable contribution.
- Restore the baseline: use A/B/B/A comparisons so changing propagation or local noise is less likely to masquerade as a geometry result.
- Match the claim to the measurement: impedance does not prove SNR, and a single null observation does not prove an entire azimuth pattern.
Bottom line: the circle is the equal-perimeter area champion. The octagon keeps 94.8% of that ideal area and can offer useful straight-part repeatability. The square remains the simplest regular straight-sided build. Choose the shape for the declared mechanical and electrical constraints, then measure the completed antenna before turning a geometry calculation into a reception claim.
Engineering References
- NIST Technical Note 1506 — Electromagnetic Theory of Reverberation Chambers (Appendix C develops the small-loop antenna response)
- NBS Scientific Paper 468 — Formulas and Tables for the Calculation of the Inductance of Coils of Polygonal Form
- NBS Technical Note 658 — Development of Electric and Magnetic Near-Field Probes
Mini-FAQ
- Does an octagon enclose more area than a square? Yes. For ideal regular shapes with equal perimeter, an octagon encloses 20.7% more area than a square.
- Is a circle theoretically better than an octagon? A circle encloses the maximum area for a fixed perimeter. A regular octagon retains 94.8% of that area and may be easier to construct from straight sections.
- Does the extra area guarantee better SNR? No. It can increase ideal open-circuit voltage, but wanted signal, external noise, front-end noise, loss, balance and common-mode pickup determine received SNR.
- Do square corners cause major RF loss? Not simply because they are right-angle bends. Conductor dimensions, joints, shield continuity, loading and frequency need to be evaluated.
- Does an octagon guarantee a deeper null or higher CMRR? No. Those are complete-system results governed by electrical balance, shield and feed design, cable common mode and the installed environment.
- How should two loop shapes be compared? State what is held equal, test at the same calibrated reference plane, control receiver state, measure cable current and use a restored-baseline field comparison.