Designing antennas for 4NEC2 modelling

Part #8 of the "Roger Writes" series - November 2023

This page explains the steps to design a simple antenna, in NEC2 format (using variables, so for 4NEC2). See 4NEC2 antenna models for lots of examples (aka starting points)

A brief history of mathematical antenna modelling

BRACT - Early 1970's
- Used to analyse antennas consisting of driven thin wires in free space
AMP - Mid 1970's (AMP = Antenna Modelling Program)
- Revised to improve model input
- Allowed for the effects of ground planes
AMP2 - Late 1970's
- Adds calculations for extended surfaces like reflectors
NEC - Early 1980's (NEC = Numerical Electromagnetics Code)
- Better simulation of currents along wires, and at junctions between them
- Better modelling for wires with multiple connections
- Better handling for wires with different diameters
- (Closed source, no longer available)
NEC2 - Mid 1980's
- Better accuracy
- (Open sourced, so used as the basis for many modelling packages)
NEC3 - Late 1980's
- Uses the Sommerfeld model to properly model wires buried in or close to the ground
- (Closed source, no longer available)
NEC4 - 1990's
- Improvements for small antennas (UHF/Microwave)
- (Closed source, executable available for license from Lawrence Livermore National Laboratory)
NEC5 - 2000's
- Improvements for wires and surfaces with mixed potentials
- (Closed source, executable available for license from Lawrence Livermore National Laboratory)

Note: MiniNEC is an independent package based on the concepts of NEC. The source code for improved versions has been released, MiniNEC2 (1984) and MiniNEC3 (1986).

Modelling packages

These are essentially graphical wrappers around the NEC (Or MiniNEC) engines. Generally they allow:

Simple NEC file entry (via graphical, or syntax assisted editors)
Command generation (SWR, Near-field, Far-field, fix/range of frequencies)
Graphical display of results

The most widely used packages for Windows are:

EZNEC - Now free (https://eznec.com/)
- Uses the (free) NEC2 engine - but has options to use the licensed NEC4 and NEC5 engines
- Has a graphics entry editor and a basic syntax assisted editor
- Stores files in a propitiatory .ez format complete with settings
- Has good graphics display options, including things like antenna voltages and currents
- No expressions allowed in calculations
- Has very limited optimiser
MMANA-GAL - Basic version is Free (http://gal-ana.de/basicmm/en/)
- Uses the (free) MiniNEC3 engine - best performance
- Basic syntax assisted editor with live graphical viewer
- Stores files in a propitiatory text based format complete with settings
- Has good graphics display options, including things like antenna voltages and currents
- Allows limited expressions in calculations: +, -, *, /, ()
- Has limited optimiser
4NEC2 - Free (https://www.qsl.net/4nec2/)
- Uses the (free) NEC2 engine - but can be (not easily) modified to use the licensed NEC5 engine.
- Has a graphics entry editor and several syntax assisted editors
- Stores files in the basic NEC2 format, and uses global settings for some things (like target impedance)
- Has graphics display options, good for antenna designs, and near-field/far-field, but not so good for wire voltage/current
- Upto 64 variables (includes the pre-defined ones, that can be deleted)
- Expressions and operators allowed in calculations, variable assignments, and wire definitions:
  - Expressions: +, -, *, /, ^, ()
  - Operators: sin, cos, tan, atn, sqr, exp, log, log10, abs, sgn, int, mod
- Powerful optimiser effective on even partially parametrised designs

Main limitations of antenna modelling

These are the main limitations of the NEC engines, (similar for most antenna modelling):

Wires must be segmented, more segments = slower (2x segments = 4x slower)
Segments (including feedpoint) must be at least 1/1000^th wavelength long, should be less than 1/10^th wavelength long
Limitations on wire intersections (segment size, wire diameter, and acute angle limits)
Parallel wires must have similar (sometimes very similar) segment lengths/segmentation
Auto segmentation options don't always work!
Various ground models are available, but the better ones are much slower.
Modelling capacitors and inductors/traps is limited (uses modified for segments)
Defining antennas with multiple feedpoints, with different phasing/loading is complicated (EX lines with delays are complicated)
Modelling helix and arc wires is extremely complex (have to be defined, then moved within the model)

The good news: Many errors are really warnings, and often the result is "close enough" as a starting point when making an antenna.

Basic design / NEC file commands

The core algorithms only model straight line segments, although helix, arcs, and wires can be entered, they get split into straight segments. (Note: By default the dimensions are in meters.)
Here are a few basic commands:

Command	Description
CM	Full line comment
CE	End of initial comment block
SY	Variables definition (4nec2 only)
GW	Wire Specification
GE	End of Geometry Input
GN	Ground parameters
GS	Scale Structure Dimensions (if you really want to use feet instead of meters add a "GS 0.3048" to your file)
EX	Excitation - usually via one of the defined wire segments
FR	Frequency - frequency for calculations and plots
EN	End of input
'	Adds comments to the end of a line.

In the NEC file format, wires are defined one at a time, with a GW line:
GW {ID} {SEGS} {X1} {Y1} {Z1} {X2} {Y2} {Z2} {RAD}

{ID}: Wire identifier/tag (integer, 0=special case)
{SEGS}: Number of segments to split this wire into (integer)
{X1} {Y1} {Z1}: Start point of the wire in 3D space
{X2} {Y2} {Z2}: End point of the wire in 3D space
{RAD}: Radius of the wire

Examples:

GW 99 1 -0.04 0 8 0.04 0 8 0.0005: GW = Wire
Tag 99
With 1 segment
From (-0.04,0.0,8.0)
To (0.04,0.0,8.0)
Radius of 0.0005 = 1mm diameter
Description: A small wire 8m up, and 4cm either side of the centre point
GW 10 45 -0.04 0 8 -9.72 0 5 0.0005: GW = Wire
Tag 10
With 45 segments
From (-0.04,0.0,8.0)
To (-9.72,0.0,5.0)
Radius of 0.0005 = 1mm diameter
Description: A long diagonal wire starting from the left side of tag 99, at 8 meters up, ending nearly 10meters away, 5 meters up
GW 11 45 0.04 0 8 9.72 0 5 0.0005: GW = Wire
Tag 11
With 45 segments
From (0.04,0.0,8.0)
To (9.72,0.0,5.0)
Radius of 0.0005 = 1mm diameter
Description: A long diagonal wire starting from the right side of tag 99, at 8 meters up, ending nearly 10meters away, 5 meters up

If we plot these out we get a very recognisable shape:

Example

Why variables

While we can sketch out an antenna, and work out where the 3D coordinates should be by hand (or spreadsheet). Every time something changes, we have to do the calculations over (and over) again. Allowing variables is a major advantage of 4NEC2 over the EZNEC and MMANA packages.
Using variables, and the built in calculations, allows a bunch of input parameters to define the whole antenna, then uses the computer to do the calculations!
If you really want the pretty graphics and other features of the other package, then from 4NEC2 you can select Calculate|NEC Output-data|Use original file, and take the 4nec2\out\{whatever}.inp file, then import this raw nec2 file into EZNEC and MMANA, and display it from there.

Start of file

Let's start our parametrised inverted V design, with something about the antenna:

Command	Comment
CM	Classic Inverted V for 40m, taller centre pole shorter outer poles
CM	Example by Roger Walker, GW6HRU
CE

Defines/Calculations

Now define some values, and do some calculations:

Command	Variable	Value/Calc	Comment
CM These are the input values
SY	freq =	7.100	'Frequency (in MHz) - Usually easier to work with than wavelength
SY	midheight =	8.0	'Midpoint height (in meters)
SY	endheight =	5.0	'End height (in meters)
SY	feedwidth =	0.08	'Feed point width (in meters)
SY	radius =	0.0005	'Wire radius (in meters)
SY	factor =	0.98	'Velocity and ground proximity factor, start with 98%
CM These are the calculated values
SY	wavelen =	299.792458/freq	'Wavelength of freq (in freespace)
SY	wirelen =	factor*wavelen/4	'Length of each wire (a quarter wave, with velocity/ground factor)
SY	halfwidth =	sqr(wirelen^2 - (midheight-endheight)^2)	'Pythagoras: Diagonal² = Width² + Height² So: Width=√(Diagonal² - Height²)

Wires

Using these variables in the wire commands we get:

Command	Tag	Num segs	Start		End		Radius	Comment
GW	99	1	-feedwidth/2	midheight	feedwidth/2	midheight	radius	'Feed point
GW	10	45	-feedwidth/2	midheight	-halfwidth	endheight	radius	'Left diagonal wire
GW	11	45	feedwidth/2	midheight	halfwidth	endheight	radius	'Right diagonal wire

Admin commands

Now we add a few extra commands, to create a complete .nec file:

Command	Parameters	Comment
GE	1	'End of geomentry, 1=Use Real Ground
GN	2 0 0 0 4 0.003	'Ground parameters, 2=Moderate, 4=dielectric constant, 0.003=Conductivity
EX	0 99 1 0 1 0 0	'Tag=99, segment=1, Load 0=Voltage source
FR	0 0 0 0 7.100 0	'Default frequency for modelling=7.100
EN

What can we do with it?

Putting the above lines in a basic .nec file allows us to do some simple analysis in 4NEC2 (start 4NEC2, then File|Open 4nec2 in/out file).
Starting with an SWR frequency sweep of this file (Calculate|NEC output-data|Frequency sweep), we can see that our initial velocity estimate is not so great. For our "perfect" antenna model, we get:

Because we have defined this antenna in terms of variables, we can easily change the them, 4NEC2 will automatically modify the antenna design, and we can re-simulate.
We could guess different velocity factors, homing in on a better value, or we can let the computer do the work. If we ask it to optimise the "factor" variable, giving SWR 100% weighting factor, for 7.1MHz (Calculate|Start Optimiser|Select "factor"|Start), we get:

This finds a factor of 0.9654 (so 96.54%), if we use that, and re-run the SWR sweep, we get:

We can now use these modified values to view a number of parameters, like gain:

Or a Smith chart:

So we now have a better estimate for the design. To find the calculated lengths we can click on the wires in the graphical editor, and it will show the wire lengths. Using them we can go and make something!
If you want to see how close the simulations can be to physical antennas, see my writeup on building a 3-element cubical quad for 2m.

The Roger Writes series

I research / dabble with lots of things, and figured that if I write my notes here, I can quickly reference them, also, sometimes, they are useful to others!
Here is what I have so far:

Portable LiFePo₄ battery bank (RW#19 - January 2026)
Upgrading the caravan battery to LiFePo₄ (RW#18 - July 2025)
Replacing the rear motor mount and rear pads/discs on a Mitsubishi Outlander PHEV (RW#17 - March 2025)
A simple 15m/21MHz inverted-V antenna (RW#16 - February 2025)
Outside Tree Christmas light project (RW#15 - December 2024)
Fitting a dashcam to a BMW i3 (RW#14 - October 2024)
Cosplay Arduino programming (RW#13 - June 2024)
ATU-100 modifications (RW#12 - February 2024)
Garage door open warning (RW#11 - January 2024)
Horizontal Halos and Vs (RW#10 - December 2023)
Faking Dell PSUs (RW#9 - December 2023)
Designing NEC2 antennas (RW#8 - November 2023)
4NEC2 antenna models (RW#7 - October 2023)
3-element cubical quad for 2m (RW#6 - September 2023)
Cobwebb Antenna dimensions (RW#5 - August 2023)
WS2812 and similar 'Smart' LEDs (RW#4 - January 2023)
Investigation and details of the AFHDS-2A protocol (RW#3 - June 2022)
Changes we've made to our caravan (RW#2 - January 2022)
Basic solar/battery/off-grid calculations (RW#1 - January 2022)
Credit card size walking robot aka Monster-Chan (RW#0 - Originally 2016 - Moved here 2022)
Why you should avoid certain number of drives with ZFS (aka why isn't my ZFS/raidz array the size I expected)
Simple way to connect a Servo to a PC (RS232 port)
Using ZFS with FreeBSD

Homepage.
This page was lasted updated on Tuesday, 28-Nov-2023 14:57:39 GMT