Antenna Modeling Software: EZNEC, MMANA-GAL, and 4NEC2 Compared

I'm going to save you the weeks of confusion I went through: if you want to design or optimize an antenna, model it in software first. Full stop. The days of cut-and-try antenna building ended when these tools became freely available.

But which one? There are three serious antenna modeling programs in the ham radio world. All use the NEC (Numerical Electromagnetics Code) engine underneath. They differ in interface, features, and workflow.

The Quick Answer

• Just getting started? MMANA-GAL. Free, intuitive, gets you results in minutes.
• Serious antenna designer? EZNEC. Industry standard, best documentation, $89 full version.
• Power user who wants full NEC control? 4NEC2. Free, powerful, ugly interface.

Now let me explain why.

MMANA-GAL

Price: Free Platform: Windows (runs in Wine on Linux/Mac) NEC Engine: Built-in MININEC + NEC2 option

This is where I tell most people to start. The interface is dated (looks like Windows 98) but the workflow is dead simple:

1. Define wires by coordinate points
2. Set frequency
3. Click calculate
4. See pattern, gain, impedance, SWR

The built-in antenna library has hundreds of designs — dipoles, Yagis, verticals, loops, quads. Open one, adjust dimensions, recalculate. You can have a working model in under five minutes without reading a manual.

What it does well:

• Yagi optimization (the built-in optimizer is excellent for Yagi design)
• Quick "what if" experiments
• 3D pattern visualization
• SWR sweep across a frequency range
• Element taper correction (real aluminum tubing dimensions)

Limitations:

• Can't model complex ground systems accurately
• Wire-only — no surface modeling
• Limited to about 500 segments before it gets slow
• The MININEC engine has known accuracy issues near ground; switch to NEC2 for verticals and ground-mounted antennas
• Documentation is... sparse. Some of it is in Japanese (original developer is JE3HHT)

My workflow: I prototype almost everything in MMANA-GAL first. Get rough dimensions, check that the concept works, optimize element spacing. Then if I need more accuracy (ground effects, radial systems, nearby structures), I move to EZNEC.

EZNEC

Price: Demo free (20 segments), EZNEC+ $89, EZNEC Pro/4 $500+ Platform: Windows NEC Engine: NEC-2 (standard), NEC-4 (Pro version)

The standard in amateur antenna modeling. Written by W7EL (Roy Lewallen), who's been developing it since the early 1990s. The documentation alone is worth studying — Roy explains antenna concepts better than most textbooks.

What makes it worth $89:

• Accurate ground modeling (Sommerfeld-Norton ground). Real-world ground effects on pattern and gain are modeled correctly.
• Handles radial systems properly — essential for vertical antenna design
• Excellent source documentation — every feature has a detailed explanation of what it does and why
• Near-field analysis (useful for RF safety compliance)
• Very stable — I've never had it crash on a valid model
• Standard file format that others can read

The interface: Utilitarian. Table-based wire entry. Not pretty but efficient once you learn it. You enter wires as endpoints with coordinates, set the number of segments, define sources and loads, and run.

The demo version limits you to 20 segments total. That's enough to model a simple dipole or 3-element Yagi, but not much more. Enough to decide if you like the workflow before buying.

When EZNEC is the right choice:

• Modeling ground-mounted verticals with radial systems
• Any antenna where height above ground significantly affects performance (all of them, but especially low dipoles)
• When you need publishable results that others will trust
• RF exposure analysis for compliance
• When you want to actually understand what the model is telling you (the documentation teaches this)

I'll get hate for this, but the free demo is borderline useless for real design work. 20 segments forces such coarse modeling that results are approximate at best. If you're serious about antenna work, spend the $89 for EZNEC+. It's cheaper than one roll of coax.

4NEC2

Price: Free (donationware) Platform: Windows NEC Engine: NEC-2 (full implementation)

4NEC2 is what happens when an engineer builds a tool for engineers. It's powerful. It exposes virtually every NEC-2 capability. And the learning curve is steep.

What sets it apart:

• Full access to all NEC-2 card types (the raw commands that control the simulation)
• Geometry builder with variables and equations — you can parameterize designs
• Built-in optimizer that can sweep multiple variables simultaneously
• Frequency sweep with animation — watch the pattern morph across a band
• Supports transmission line models (simulate matching networks, phasing lines)
• Far/near field calculations
• Free. Completely free.

The catch: The interface is not intuitive. At all. You'll spend your first hour figuring out where things are. The menus are dense. The help is a reference, not a tutorial. Several windows open simultaneously and interact with each other in non-obvious ways.

I spent three evenings just learning how to properly enter a simple dipole, run a sweep, and export results. Not because it's complicated in concept — but because the UI fights you until you learn its patterns.

When 4NEC2 is the right choice:

• You want full NEC-2 power without paying
• You're modeling complex phased arrays with transmission line feeds
• You need parametric sweeps ("show me gain vs. boom length from 3 feet to 12 feet in 6-inch steps")
• You're comfortable reading NEC card documentation directly
• You want to model matching networks as part of the simulation

My honest opinion: If you can get past the interface, 4NEC2 is more capable than EZNEC+ (not Pro) in some areas — particularly transmission line modeling and parametric optimization. But the EZNEC documentation and interface make it vastly more productive for typical ham antenna work. You'll get accurate results faster in EZNEC even though 4NEC2 can technically do more.

NEC Basics You Need to Know (Regardless of Tool)

All three programs run on the same physics engine. Some universal rules:

Segments per half-wavelength: Use at least 10-11 segments per half-wavelength of wire. Fewer = inaccurate. More = slower computation for minimal accuracy gain. The sweet spot is 11-21 segments per half-wave.

Segment length ratios: Adjacent segments shouldn't differ in length by more than about 2:1. Abrupt changes create numerical errors. Taper gradually.

Wire radius matters: Thicker wires have lower impedance and wider bandwidth. The model needs accurate wire diameters — don't just guess.

Ground type: Free space, perfect ground, and real ground give vastly different results. Use real ground with appropriate conductivity for your location. Average soil: conductivity 0.005 S/m, dielectric constant 13.

Convergence check: Run your model, then double the segments and run again. If results change significantly, your original segmentation was too coarse. If they're nearly identical, you're fine.

My Recommendation

Start with MMANA-GAL because it's free and fast. Build a few Yagis, model a dipole at various heights, get comfortable with what antenna models tell you.

If you get serious about antenna design — especially verticals, low antennas, or anything where ground effects matter — buy EZNEC+. The $89 is the best investment in your antenna education.

Keep 4NEC2 installed for when you need its unique capabilities (parametric sweeps, transmission line modeling). Learn it gradually. Don't try to make it your primary tool from day one unless you enjoy suffering.

And whatever you use: model first, build second. A $0 simulation that shows your antenna is resonant at 14.2 MHz saves you from cutting aluminum for an antenna that ends up resonant at 13.8 MHz. Ask me how I know.