Wire Antenna Simulator (NEC-2)
Solve your own wire antenna with the NEC-2 method of moments: dipoles, Yagis, verticals and loops, element by element. Radiation pattern, gain, impedance, VSWR, efficiency and current distribution, at one frequency or across a sweep, with a multi-objective optimiser on Pro.
A preset fills in the geometry below — every value stays editable.
| # | Element role | Feed point | Length (mm) | Boom position (mm) | Radius (mm) | Segments | |
|---|---|---|---|---|---|---|---|
| 1 | Driven element |
Drawn from the wires that will be solved, to one scale. The dot is the feed.
- Checking the geometry…
How It Works
This tool solves wire antennas with NEC-2, the Numerical Electromagnetics Code: a method-of-moments solver that divides every wire into short segments, solves for the current on each one, and computes the far field, the feed impedance and the losses from those currents. It solves your geometry — every element length, spacing, radius and segment — rather than a textbook figure for an antenna of the same kind, so mutual coupling between elements, ground reflections and conductor loss come out of the solve instead of being assumed.
Solving is a choice among four named modes. Instant is a closed-form textbook model of a preset antenna — a sanity check in milliseconds, and labelled as a model, never as a solve. Standard solves the geometry with NEC-2 at one frequency. Sweep solves it at every point of a band, so impedance, VSWR, gain and pattern are reported per frequency. Optimize (Pro and API) runs an NSGA-II search over the element lengths and spacings of a parallel-element array, trading gain, front-to-back ratio and match against each other, and returns the whole trade-off front with a recommended design.
Every geometry is checked before it is queued, against the limits within which NEC-2's thin-wire model is valid: no segment longer than a tenth of a wavelength, every segment at least eight times its wire's radius, no wire thicker than about 1/63 of a wavelength, wires that touch actually joined, and nothing below the ground plane. A geometry outside them is refused with the reason and a way to fix it — not solved into numbers that look fine and are not.
Every result checks itself. NEC-2 must account for all the power it is fed: what radiates plus what the conductors dissipate should equal the input, and a result that misses by more than 2% is marked. A feed resistance at or below zero, or an antenna radiating more than it is fed, is marked physically impossible. Each result also states the conductor it assumed (copper unless you choose otherwise — a perfect conductor overstates a small loop by several dB), the ground model, what the simulator filled in for you, and the request that reproduces the run.
Three ground models are available: free space, perfect ground (an exact image), and real ground with your chosen permittivity and conductivity, solved with NEC-2's Sommerfeld-integral method.
Methodology & References
References
- Numerical Electromagnetics Code (NEC) — Method of Moments — G. J. Burke and A. J. Poggio, Naval Ocean Systems Center TD 116 (1981) link
- Antenna Theory: Analysis and Design, 4th ed. — Constantine A. Balanis (2016)
- Three-Element Yagi Models: Standards of Comparison — L. B. Cebik, W4RNL link
- A Fast and Elitist Multiobjective Genetic Algorithm: NSGA-II — K. Deb, A. Pratap, S. Agarwal and T. Meyarivan, IEEE Trans. Evolutionary Computation 6(2), 2002
Powered by
Backend tests reproduce the NEC-2 User’s Guide Example 1 feed impedance to within 0.1%, the Balanis half-wave dipole directivity to within 0.05 dB, image-theory ground results, and L. B. Cebik’s published 3-element Yagi tables.
Related Calculators
Read the Guide
FAQ
How accurate is it?+
Within its thin-wire limits NEC-2 is the reference code for wire antennas, and this implementation is tested against published answers: the NEC-2 User’s Guide Example 1 feed impedance to within 0.1%, the half-wave dipole’s textbook directivity (Balanis) to within 0.05 dB, and L. B. Cebik’s NEC-4 reference 3-element Yagis to within 0.02 dB of gain and 0.03 Ω of feed resistance. What limits a real result is usually the model — how faithfully the wires describe the antenna — not the solver.
What does NEC-2 not model?+
This tool accepts bare round wires only. It does not model insulated wire, wires buried in the ground (NEC-2 cannot place a conductor below the ground surface; a buried radial field needs NEC-4), the feed line and balun (the feed is an ideal voltage source on one segment, so common-mode current on a coax shield is not included), or nearby objects such as masts and buildings unless you draw them as wires. Elements made from joined tubes of different diameters are a known NEC-2 weakness, which NEC-2 does not correct for. Geometries outside the thin-wire limits are refused rather than solved.
Which solve mode should I pick?+
Use Instant for a quick textbook figure for a preset. Use Standard to solve your own geometry at the frequency you care about, and Sweep when you need impedance, VSWR and gain across a band — for instance to find the 2:1 bandwidth. Use Optimize (Pro) when the element count is fixed and you want the lengths and spacings tuned.
How long does a simulation take?+
It depends on the number of unknowns — the total segment count — and the number of frequencies, and on the hardware your tier runs on. Each mode card quotes a runtime only once it has been measured on that hardware. A job over your tier’s budget is refused with the computed cost and a way to reduce it, rather than run for minutes and cut short.
What does the optimiser change, and what does it return?+
It varies each element’s length (by default within 15%) and each gap along the boom (within 50%), solving every candidate with NEC-2. Designs the self-check flags are never rewarded. It returns every design on the trade-off front — none beats another on gain, front-to-back and VSWR at once — a recommended design chosen by a stated rule, a comparison with the design you started from, and the seed, so the same job reproduces the same search.
Why is the power balance not checked over real ground?+
Over lossy soil some of the input power is absorbed by the ground, which is neither radiated nor conductor loss, so the balance is not expected to close. The result says so instead of reporting a check it did not make; the physical-bounds checks still run.
What does the gain number mean?+
Gain is in dBi — decibels relative to an isotropic radiator — and includes conductor loss, so it is never higher than directivity. A half-wave dipole in free space has about 2.15 dBi; a well-designed 3-element Yagi is typically 7–8 dBi.