The Limitations of NEC
NEC (Numerical Electromagnetic Code) is a powerful tool for antenna modelling. Hams use it to simulate radiation patterns, input impedance, gain, current distribution, and other characteristics of wire antennas. But it is crucial to understand what NEC is — and just as importantly, what it is not.
This article is mainly about HF antennas. That matters. At HF, the wavelengths are long, the antennas are often physically compromised, and the “antenna system” frequently includes much more than the visible radiator: coax shields, masts, railings, rigging, ground rods, counterpoise wires, rain gutters, concrete, trees, boats, balconies, and sometimes even the operator and shack wiring.
At higher frequencies, the environment still plays a role. VHF, UHF, and microwave antennas are also affected by nearby metal, masts, feedlines, buildings, vehicles, hands, and dielectric materials. But the practical modelling problem is often easier to define. The wavelength is shorter, ground planes and chokes are physically manageable, balanced structures are easier to build, and the boundary between “antenna” and “environment” is usually clearer.
HF is different. At HF, your installation often becomes part of the antenna.
NEC is not a magic box that can simulate the entire messy reality of your installation. It is based on mathematical approximations, idealisations, and simplifying assumptions. These work well in controlled scenarios, but they quickly become fragile in real-world environments like boats, balconies, RVs, rooftops, attics, or cluttered gardens.
NEC Is a Theoretical Tool
NEC solves electromagnetic problems using the Method of Moments (MoM), mainly for thin conducting wires in free space or over simplified ground. That makes it extremely useful for clean antenna problems, but it also means the model only knows what you explicitly describe — and only within the assumptions of the solver.
In practical ham use, NEC usually assumes:
- Wires are thin compared to wavelength
- Geometry is perfectly known and perfectly fixed
- Connections are electrically stable and repeatable
- The feedpoint is an ideal mathematical source
- The environment is either absent or highly simplified
- The feedline, shack, balun, choke, and operator do not exist unless modelled
- Ground is flat, uniform, and describable by simple parameters
- Nearby dielectric materials are ignored or heavily simplified
That is why NEC is so useful for comparing clean designs, but risky when used as a prediction machine for a messy HF installation.
| NEC Handles Well | NEC Misses or Simplifies |
|---|---|
| Thin wires in free space | Thick conductors, plates, hulls, curved rails, gutters, masts |
| Ideal bonds and joints | Corroded, loose, floating, intermittent, or lossy connections |
| Simple ground models | Layered soil, seawater over mud, concrete, rebar, roofs, decks |
| Ideal feedpoint excitation | Real coax, baluns, chokes, tuners, rig chassis, shack wiring |
| Balanced antennas in isolation | Common-mode current caused by asymmetry or poor feedline isolation |
| Far-field pattern of the model | Real installation pattern after unknown coupling and losses |
| Comparative lobes and nulls | Exact gain, exact null depth, exact takeoff angle in clutter |
| Simple verticals, dipoles, loops | Complex HF systems: boats, balconies, rooftops, attics, RVs |
| Known metallic structures if correctly modelled | Unknown conductors, floating metalwork, hidden wiring, wet surfaces |
| Clean conductor currents | Losses in traps, coils, ferrites, connectors, clamps, stainless parts |
Think of NEC as a clean lab tool. It sees the idealized geometry you draw. It does not automatically see the messy, lossy, unstable HF environment around your antenna.
Why HF Is the Difficult Case
At HF, antennas are often electrically large but physically inconvenient. A proper half-wave dipole for 80 meters is not something most people can install cleanly. Even on 40 or 20 meters, compromises are common: shortened wires, bent elements, low heights, nearby buildings, poor counterpoises, improvised supports, and coax routed through the near field.
This is where NEC can mislead. The model may show a beautiful balanced current distribution and a clean radiation pattern, while the real installation has RF current flowing on the outside of the coax shield, into a balcony rail, along a mast, through a tuner, or back into the shack.
In other words: at HF, the “antenna” is often not just the wire. The antenna is the wire plus the return path plus the feedline plus the nearby conductive and dielectric environment.
| HF Reality | Higher-Frequency Reality |
|---|---|
| Long wavelengths make full-size antennas difficult | Full-size antennas are physically smaller and easier to define |
| Counterpoises and return currents are often improvised | Ground planes and reference structures are often compact and repeatable |
| Coax shields easily become part of the antenna | Feedline isolation is often easier to implement and verify |
| Buildings, trees, masts, and boats are electrically significant | Environment still matters, but the affected zone is easier to identify |
| Balanced antennas are hard to keep balanced in practice | Mechanical symmetry and good choking are often easier to achieve |
| Ground is a major unknown | Free-space or controlled mounting assumptions are often closer to reality |
The physics does not change at higher frequencies. What changes is the practicality: it becomes easier to build, isolate, measure, and explain the antenna as a defined system.
What Happens on a Boat or Balcony
Many hams model their sailboat or balcony installs in 4NEC2, adding masts, rigging, railings, water, rooftop edges, and nearby metalwork. Then they trust the pattern as gospel. That is wishful thinking.
On a fiberglass boat, many parts are electrically ambiguous. The mast may or may not be bonded. Rigging may have intermittent connections. Stainless hardware may be lossy at RF. The engine block, seawater, guard rails, pushpit, pulpit, backstay, coax shield, and shore power wiring may all participate in unpredictable ways.
On a balcony, the situation is similar. The railing may be bonded to building steel — or not. The concrete may contain rebar. Nearby windows may have metallized coatings. Rainwater can change coupling. The coax may run along metalwork or through walls. The antenna may look like a simple wire in NEC, but in reality it is coupled to a large, unknown structure.
NEC cannot fix unknowns. If you do not know whether something is bonded, floating, lossy, resonant, wet, corroded, or connected through hidden wiring, the model cannot know either. Adding more “detail” to the model does not automatically make it more accurate. Sometimes it only makes the simulation look more convincing.
Don’t Ignore the Non-Metal Stuff
Another misconception is: “Only metal matters.” Wrong — especially at HF.
Most practical NEC models used by hams do not properly include nearby dielectrics such as trees, fiberglass, plastic tanks, wet ropes, wooden masts, roofing materials, walls, or concrete. But in practice, these things can matter. They can store electric-field energy, introduce loss, shift resonance, reduce bandwidth, and disturb the current distribution.
A dry support may be almost invisible. The same support when wet may become a lossy RF object. A fiberglass boat hull may not be metal, but it can still affect tuning. A tree close to the high-voltage end of a wire may change the feedpoint impedance. A plastic tank filled with water is not the same RF object as an empty plastic tank.
NEC only knows the electrical world you describe. Real HF installations contain many things you did not describe — and often cannot describe accurately.
NEC Can Only Show the Fields of the Model
It is better not to say that NEC “sees everything.” It does not. NEC calculates currents and fields for the geometry and assumptions inside the model. If a nearby structure is not included, it does not exist. If the material properties are wrong, the result is wrong. If a joint is assumed perfect but is actually corroded or floating, the model is lying politely.
NEC can model coupling between conductors that are included correctly. But that is the catch: they must be included correctly. In many HF installations, the most important coupling paths are the ones the operator does not know about: the outside of the coax, the metal balcony, a mast, rain gutters, rebar, shack wiring, a boat’s bonding system, or a random counterpoise path through the environment.
This is why a NEC far-field pattern may look clean while the real antenna behaves strangely. The model may show a symmetric dipole. The real world may be an off-center-fed, coax-radiating, building-coupled, lossy, distorted system.
Adding floating structures into NEC can make the model look more detailed but not necessarily more accurate. If the electrical state of a structure is unknown, “guessing it in” may be worse than leaving it out and treating the model as a clean baseline.
The Big Blind Spot: Feedlines and Common Mode
One of the biggest differences between NEC and reality is the feed system.
In a typical NEC model, the antenna is excited by an ideal source placed at the feedpoint. That source is perfectly defined. It does not have a coax shield. It does not have connector loss. It does not have an imperfect balun. It does not have a tuner, a rig chassis, or shack wiring attached to it. It simply injects voltage or current into the model exactly as requested.
That is not how most HF antennas are fed in the real world.
In reality, the coax shield can become part of the antenna. If the antenna is unbalanced, poorly choked, asymmetrical, too close to metal, or fed against an inadequate counterpoise, RF current may flow on the outside of the coax. That current can radiate, distort the pattern, change the impedance, increase noise pickup, cause RF in the shack, and make SWR readings misleading.
Common-mode current is usually not a problem in NEC because, in the model, there is often no common-mode path. The model does not automatically include the outside of the coax shield. It does not automatically include the shack. It does not automatically include the operator touching the microphone. It does not automatically include the random path back through a power supply, USB cable, rotator cable, or grounding wire.
| In NEC | In Real HF Installations |
|---|---|
| The source is mathematically ideal | The source includes coax, connectors, balun, choke, tuner, and rig |
| The feedpoint is perfectly balanced if defined that way | Small asymmetries can create common-mode current |
| No coax shield exists unless explicitly modelled | The coax shield can radiate and receive noise |
| No shack wiring exists | Power cables, USB cables, audio cables, and ground wires may carry RF |
| No imperfect balun or choke exists | Ferrites have impedance limits, loss, heating, and placement sensitivity |
| No operator exists | The operator, microphone, desk, and equipment can become part of the RF system |
A NEC model can show a perfectly balanced antenna even when the real installation is badly unbalanced. Common mode is not automatically “solved” in NEC — it is usually absent from the model.
Antennas in NEC Are Idealized
In NEC, antennas are beautifully obedient. Wires are straight. Segments are where you placed them. The feedpoint is exact. Connections are perfect. Conductors do not move in the wind. Rain does not change anything. A clamp is not slightly oxidized. A solder joint does not heat. A trap does not drift. A ferrite choke does not saturate. A PL-259 connector does not have water inside.
Real HF antennas are not like that.
A real antenna is a physical object in weather, mounted on imperfect supports, fed by imperfect cables, and surrounded by objects that may or may not be part of the RF system. This does not make NEC useless. It simply means the model is an ideal reference, not a guarantee.
When a NEC model says “gain = 6.1 dBi,” it really means: under the modelled assumptions, with the modelled geometry, over the modelled ground, with the modelled source, this is the calculated result. It does not mean your actual installation will produce exactly 6.1 dBi.
What NEC Cannot See at HF
The most dangerous modelling errors are not always inside the antenna geometry. They are often outside it. NEC may give a precise answer to an incomplete question.
| Effect | Why It Matters |
|---|---|
| Common-mode current on coax | Changes pattern, impedance, receive noise, and RF-in-shack behaviour |
| Lossy return paths | Reduces efficiency, especially for verticals and end-fed antennas |
| Uncertain bonding | Floating metal may couple unpredictably or become resonant |
| Real ground complexity | Soil moisture, layering, seawater, mud, concrete, and rebar affect results |
| Dielectric loading | Trees, wet wood, fiberglass, plastic, and walls can detune antennas |
| Balun and choke limitations | Real ferrites have finite impedance, loss, heating, and frequency dependence |
| Matching network losses | Coils, capacitors, traps, and tuners may dominate the efficiency loss |
| Hidden conductors | House wiring, rails, gutters, masts, and rigging may become part of the antenna |
| Mechanical movement | Wind, sag, wet ropes, and changing geometry shift impedance and pattern |
| Receive noise coupling | A model may show radiation gain but not local noise pickup through common mode |
At HF, the invisible part of the antenna system is often the part causing the problem.
How NEC Predictions Differ from Reality
In clean, controlled, simple wire-antenna cases, NEC can be impressively useful. But in messy HF installations, the absolute numbers can drift significantly. This is especially true for gain, front-to-back ratio, null depth, feedpoint impedance, and resonant frequency.
| Parameter | NEC Prediction | Typical Real-World Deviation |
|---|---|---|
| Forward Gain | +6 dBi ideal | Often 2–3 dB lower, sometimes more due to losses |
| Front-to-Back | 25 dB | Nulls may be 10–15 dB shallower |
| Elevation Angle | 20° | Can shift ±10–15° from ground and nearby objects |
| Resonant Frequency | 7.05 MHz | Can move ±50–150 kHz or more depending on surroundings |
| Bandwidth, 2:1 SWR | 300 kHz | Often 20–40% narrower in lossy or coupled installations |
| Feedpoint Impedance | 50 Ω resistive | May change strongly when feedline, height, ground, or clutter is included |
| Receive Noise | Usually not predicted | Can be dominated by common-mode pickup on feedline and shack wiring |
These are conservative, indicative values. Real-world HF installs often drift further depending on soil, clutter, feedline routing, losses, and bonding. NEC is a guide, not gospel.
So Why Use NEC At All?
Because it is still very useful — if you use it correctly.
NEC is excellent for understanding how a design should behave under clean conditions. It helps estimate resonance, impedance, current distribution, lobe direction, null placement, and relative performance. It is especially valuable when comparing one design against another using the same assumptions.
For example, NEC is very good at answering questions like:
- What happens if I make this dipole longer or shorter?
- Where are the current maxima and voltage maxima?
- How does height above ground change the elevation pattern?
- How does element spacing affect a beam?
- What happens to the pattern if I bend the ends?
- Where are the major lobes and nulls likely to be?
- How does one clean design compare with another clean design?
But NEC is weaker at answering questions like:
- Exactly how much gain will I get on my balcony?
- Will my coax shield radiate?
- How much RF will come back into the shack?
- How lossy is my real counterpoise?
- Is my mast bonded well enough?
- How much does the wet tree next to the wire detune the antenna?
- What is the real pattern of my boat installation?
Those questions require measurement, experimentation, and sometimes a healthy dose of humility.
What I Use Personally
I often use my own lightweight Python modelling code using NumPy and Method of Moments assumptions. It generally works in free space and ignores ground, clutter, dielectric loading, and real installation details. Yet the pattern shapes usually agree well with NEC for clean wire structures. The absolute dB values differ, but the comparative lobes, nulls, and behaviour trends are consistent.
That is the useful lesson: antenna modelling is about insight and trends, not chasing absolute decimals. Whether the model says 5.7 dBi or 6.2 dBi rarely changes the practical decision. What matters is understanding where the current flows, where the lobes go, where the nulls fall, and which design choices improve or worsen the behaviour.
For HF, the model is the beginning of the work — not the end. The real antenna is proven by measuring, choking, trimming, moving, observing, and iterating.
Takeaways for Hams
- This discussion is mainly about HF antennas, where the installation often becomes part of the antenna system
- NEC assumes idealized geometry, idealized sources, and simplified surroundings
- NEC is excellent for clean wire models and comparative design work
- NEC does not automatically include coax, common-mode current, shack wiring, baluns, chokes, or the operator
- A balanced NEC antenna may become unbalanced in the real world
- Common-mode current is usually absent from the model unless explicitly included
- NEC cannot reliably predict unknown bonding, corrosion, floating metal, or lossy joints
- Nearby dielectric materials can detune real HF antennas even when absent from the model
- At higher frequencies the environment still matters, but balance and isolation are often easier to achieve and explain
- Use NEC for comparisons and insight, not as final truth
- Expect at least ±2–3 dB deviation from reality in gain numbers for messy HF installations
- When in doubt: measure → choke → tweak → observe → iterate
When modelling HF antennas on boats, balconies, attics, rooftops, RVs, or compromised gardens: NEC is an approximation, a starting point — not a guarantee. The model sees the idealized antenna you describe. The real world includes feedlines, common-mode currents, lossy return paths, dielectric loading, floating conductors, and hidden coupling. The real work is measuring and adjusting on site.
Mini-FAQ
- Is this mainly about HF? — Yes. The problem is most severe at HF because the antenna, feedline, counterpoise, ground, and surroundings often become one coupled RF system.
- Can NEC model my balcony or boat antenna exactly? — No. It cannot account for unknown bonds, clutter, dielectric loading, feedline common mode, and lossy real-world connections unless they are known and modelled correctly.
- Is NEC useless? — Not at all. It is very useful for baseline design, understanding current distribution, and comparing clean antenna models.
- Does NEC show common-mode current? — Only if the common-mode path is included in the model. In many simple NEC models, the coax shield and shack do not exist, so common mode cannot appear.
- Are higher frequencies immune to the environment? — No. VHF, UHF, and microwave antennas are also affected by surroundings, but the antenna system is often easier to define, balance, isolate, and measure.
- How accurate are NEC gain numbers? — In clean models they are useful. In messy HF installations they can easily be off by 2–3 dB or more.
- What should hams trust more? — NEC for shapes, trends, and comparisons; real measurements for truth.
Interested in more technical content? Subscribe to our updates for deep-dive RF articles and lab notes.
Questions or experiences to share? Contact RF.Guru.