The source is what drives the antenna. Think of it as the feed point of the antenna. It’s where we apply either a SOURCE of voltage or current to drive the antenna.

In EZNEC, sources are configured in the source window

Here is the source I specified for the Inverted L antenna

FIRST … I specified only 1 source

SECOND … I specified the wire position as Wire #1, at 0% from END 1 (E1), in other words, at the bottom of the antenna. The program put the source as close to what I specified as possible. NEC always puts the source in the middle of a segment, so the ‘Actual Pos.’ or actual position is 12.5% from E1 in Segment #1.

THIRD … I specified an amplitude of ‘1’ and phase ‘0’ degrees. One is always a good starting point for an investigation. Phase is useful when there is more than one source in the model and you need to establish the relative phase between them.

FOURTH … The type of source is ‘I’ a current source. This means that a CONSTANT 1 amp is applied to the feed point and everything else works from there. If you’re applying a constant voltage, you specify it as a ‘V’ voltage source. When the source is applied across a junction, you specific a ‘SPLIT’ source (SI or SV).

Categories: Antenna Modeling
3 Comments »

Just for fun, another look at the Delta Loop with the Antenna Model program is interesting because it uses MiniNEC instead of NEC as a computational engine:

Compare this to the NEC model and see what you think.

Categories: Antenna Modeling
No Comments »

After some discussion, we refined the description of the antenna to this:

—————– Message ——————————–

But there is also a problem with this description. Do you see it? Let’s look at the height above ground segment by segment:

—————– Message ——————————–

Let me step back and clarify the point “every bit of information is critical to get a good model”. Especially at this stage it’s true, but we might find that we can compromise later. The reason is very simple. Radiation is caused by current flow in the antenna wire. The important parts of the wire for the remote pattern will be where current is highest. A section where current is low or zero will have little or no impact on the pattern and so inaccuracies at this point will not have a measurable impact on the result. Once you’ve been modeling for a while, you’ll begin to have a feeling for where these places are, but any time you’re working with an unfamiliar antenna or an unusual setup for antenna that you may have worked with in a different environment you’ll need to do some preliminary modeling to understand where the current is high and where it’s not.

I always try to be careful with initial setup, looking for ways to confirm my setup and getting the numbers right. Once I can create a model that accurately reflects what I find, then I know I’ve got something to experiment with. When I’m working with someone else on a model, I try to not impose an interpretation without confirmation because as you can see from this simple example, it’s VERY easy to get the numbers wrong, and until I know just where the current is high, I don’t know which errors will be significant to the result.

From the reconstruction above … I entered the wires like this:



You can see from this where the current is high and where it’s lower. This will change when I add the radials, but it’s a good start.

You’ll find that modelers differ on what they consider important at various stages and the order of their steps varies with the person. I’m sure Cebik would look at it differently if he were able to comment. However, no matter how you actually use the outputs and what order to do the steps, you still face the same problems and will eventually get to the same results.

Categories: Antenna Modeling
8 Comments »

Our next step with the Delta loop is to get the geometry correct. A bit more clarification got this out:

—————– Message ——————————–

(0, 0, 16.4)
(0, -5.9, 26.6)
(0, 5.9, 26.6)

With these figures in hand, we could rebuild the model

—————– Message ——————————–

The Delta Loop is an interesting case to model because we ordinarily put the source in the middle of a wire for a dipole (exactly what we’ve been doing on our 40 Meter Dipole model), but here we have a case where we’re modeling it at a vertex which introduces some potential problems.

We have three ways we can model this:

1. Put the source on a wire next to the vertex and as close to the vertex as possible
2. Put the source across the vertex, partly on one side and partly on the other (EZNEC has a convenient way to do this when describing the type of source)
3. Put a short wire where the vertex is to connect the two wires that come to the vertex and put the source on this short wire.

For the purpose of this model, I chose to use the 2nd option and model it with what EZNEC calls a ’split source’. So here’s EZNEC’s summary of the model description at this point:

                      EZNEC+ ver. 4.0

10 m Delta Loop Array ()                1/13/2009     7:31:10 PM

         --------------- ANTENNA DESCRIPTION ---------------

Frequency = 28 MHz
Wire Loss: Copper -- Resistivity = 1.74E-08 ohm-m, Rel. Perm. = 1

              --------------- WIRES ---------------

No.          End 1   Coord. (ft)              End 2   Coord. (ft)       Dia (in)  Segs  Insulation
      Conn.      X       Y       Z     Conn.      X       Y       Z                   Diel C  Thk(in)
1      W3E2        0,      0,   16.4    W2E1        0,   -5.9,   26.6       #12   40       1        0
2      W1E2        0,   -5.9,   26.6    W3E1        0,    5.9,   26.6       #12   40       1        0
3      W2E2        0,    5.9,   26.6    W1E1        0,      0,   16.4       #12   40       1        0

Total Segments: 120

              -------------- SOURCES --------------

No.      Specified Pos.     Actual Pos.      Amplitude    Phase    Type
       Wire #  % From E1  % From E1  Seg       (V/A)     (deg.)
1       1         0.00       1.25    1        1           0         SV

No loads specified

No transmission lines specified

Ground type is Real, High-Accuracy

              --------------- MEDIA ---------------

No.    Cond.    Diel. Const.  Height    R Coord.
       (S/m)                  (ft)      (ft)
1      0.005     13           0         0       

------------------------------------------------------------------------

Categories: Antenna Modeling
No Comments »

Another recent question to the Antenna Modeling discussion group asked about a Delta Loop:

—————– Message ——————————–

When you’re building a model, one of the first things you need to do is be specific about the geometry. You can talk in general about Delta Loop patterns, but if you want to build a model, you’ve got to be specific:

—————– Message ——————————–

Once we have some numbers, we can start to build a model. Problem is, not everything is specified and as you can see, some of the numbers can be used in a way that is consistent with what’s been said, but is not correct. To illustrate this, I tried to be as simplistic as possible and modeled the antenna with a 40 foot section between two trees and a feed point 16.4 feet above the ground:

—————– Message ——————————–

In the next post, the author has corrected me and we redo the model.

Categories: Antenna Modeling
No Comments »

Picking up from the question posted earlier, here are my original thoughts on the posting with some expansion on some points to be aware of:

1) surroundings will have an effect on the antenna pattern. Even the tree you’re running through will have an effect which is likely varying as it’s conductivity varies over time. Is there anything conductive which isn’t mentioned in the description? That would impact the actual pattern of the antenna. At least initially, I’d build the model showing only the antenna wire and the radials, then see how close that brings us to what you can actually measure.

2) NEC2 cannot model wires on or under the ground. The radials will have to be modeled slightly above ground which will affect the accuracy of the model. NEC4 could handle that, but that’s not readily available. This will lead to some inaccuracy, but usually at a level small enough to be unimportant.

3) the feed line and chokes can be modeled at best approximately. One way to handle them is to treat them as a separate modeling problem generating a voltage or current at the output of the feedline, which is then put into the antenna model for that frequency. Trying to model the feedline in EZNEC probably won’t be terribly accurate.

Antenna Modeling will not give you details because the assumptions made in building any real model will always compromise the details at some level. I think our first effort is to see if we can reproduce the SWR as you’ve measured it. That also raises a point:

4) where did you measure the swr? Was it measured in the shack (I expect so) and if so, what you measure there will be affected by the coax you’re using to feed the antenna, how long it is, and so forth. You can think of a coax as an impedance transformer. We’d need to know the length of the coax and where the coils are in the line (exact measurements) in order to model it.

The very first step now is to get the geometry of the wires laid out correctly. I’ve put a layout diagram at

As a starting point. This has several problems:

1) I don’t know exactly where the feed point is, where is the start for this antenna? ON the ground? 1 foot above? 10 feet above?

2) All of the wires are in the same azimuth plane in my diagram. After the 8 foot vertical line, a plane is defined by the 8 feet of wire angled up at 45 degrees. On the next length of wire up to the tree, does it lie in the same plane as the first two pieces or is it angled? Same thing for the horizontal section from the tree to whatever other support it has at the end. You didn’t say anything about the horizontal support. What is it?

As a start, I set up the preliminary geometry in EZNEC. You can see the wires plotted on an XYZ coordinate axis here:

The wires are all lined up. Are they really? For this example, I set the feed point at 10 feet above the ground, assumed that the only thing at 90 feet was the last wire (#4 in the model), so the antenna REACHES 90 feet at the tree. That’s a lot of assuming!

3) How are the radials laid out relative to the antenna itself.

If you’ve been following my sample discussion about modeling on my blog, you’ll know that we have to be able to choose the X, Y, & Z position of the end point on each side of each wire. My diagram shows all the wires lying in the same plane, That’s probably not true, but how much difference is there?

4) I’d assume that the feed point is near the shack (house?), if so what’s around there? Metal gutters? Metal fences? Anything conductive can potentially impact the antenna.

5) the ground lines are tied to some metal gutters and a fence. We need to position them in the model using wires to approximate their locations. Their impact may be minimal or it could be that we could simply lump their effect into the model, but unless we know the geometry, we won’t know how to account for them.

ANY model is an approximation. The closer you want the model to reflect the real situation, the more careful you need to be about the geometry. We can probably get approximate answers pretty easily by ignoring most of the points I’ve made above, but we don’t know whether they’re important until we understand them.

Categories: Antenna Modeling
No Comments »

This question was posted to the AntModel Antenna Modeling list on yahoogroups. I’ve started to discuss it there, but I’ll pick it up here as well because I can embed the graphics here.

————— Question ————————-

I’ll post the first part of an answer tomorrow

CHANGES: Reformatted the email to better point out that it IS an email.

Categories: Antenna Modeling
No Comments »

Coordinate Systems - Our 40 Meter Dipole

Let’s go back to our 40 meter dipole and look at the actual numbers we used to describe the wires. We started by computing the length of the antenna we wanted to erect from the standard formula for a dipole:

dipole length (feet) = 468 / frequency (Mhz)

= 468 / 7.15

= 65.454545 feet =~ 65.45 feet

and from that we came up with this entry for a single wire:

so how did we get from 65.45 feet to coordinates

(X = 0, Y = -32.73, Z = 70)
(X = 0. Y = 32.73, Z = 70)

Why this set of numbers and not some other?

If the answer isn’t clear to you, don’t worry. If you’re not used to working coordinate system problems it can be a little hard to see the logic behind the choices, but once you understand the why, it becomes pretty easy to make choices like that yourself.

We chose these actual numbers for a lot of reasons, let’s see if we can get them all down:

1. To simplify our lives - we wanted to make the model as easy as possible to understand, so we want to set the antenna as simply as possible. We need a wire that’s approximately 65.45 feet long. We can put it anywhere in the coordinate system. For example, we could run the same wire along the X instead of the Y axis:

   (X = -32.73, Y = 0, Z = 70)
   (X = 32.73, Y = 0, Z = 70)

   You could certainly model it this way, but by CONVENTION, we usually want to see the pattern maximum along the X axis. We already know that the maximum radiation for a dipole is at right angles to the wire, so to get the maximum along the X axis, we lay out the wire along the Y axis.

2. We want to work with a horizontal antenna -
3. To be ready to work with the antenna above ground - We could easily have modeled the antenna at z = 0 for both the start and the end points

When we come back to this, we’ll take a look at some other ways we can model this wire and how we can transform one into another.

Categories: Antenna Modeling
2 Comments »

Coordinate Systems

Let’s take some time and go back and look in more detail at the setup for our 40M Dipole. So far, we’ve seen that we can set up the model in a wide variety of software and even use the original NEC program if we’d like to produce the same results at this very simplistic level. Before we get too far though, we need to step back and look at the details to make sure we get them right.

In this installment, we’re going to take a quick look at coordinate systems. In the future, we’ll take apart the setup parameters one-by-one and see what they mean.

Antenna Geometry

In order to model an antenna, we have to have a way to specify it in space. The standard way to do this is to specify the end points of each wire in the model.

IMPORTANT POINT: All of the antenna models we’re going to deal with make a simplifying assumption that we need to understand: ANY antenna can be reduced to a collection of short, straight wires. A straight wire is simpler to model mathematically than a curved wire. If we make our straight wires short enough, we can approximate a curved wire good enough to get useable answers.

First, consider a point in two dimensions, say the surface of a table.

We specify the point by giving two coordinates. For all of our work, we’ll use straight line measurements (linear measurements). So we pick a reference point, called the ‘origin’ and we assign it coordinates X=0 and Y=0. we measure from this position along arbitrary axes at right angles to each other, X along the ‘X’ axis and Y along the ‘Y’ axis. In our example below, we measure 2 to the right and 3 up (X = 2, Y = 3) for the point in the upper right. The upper left point is 3 to the left and 1 up (X = -2, Y = 1). It’s simple and easy to work with.

For a real antenna, we need to place it in three dimensions, so we add a Z coordinate which is perpendicular to the X & Y coordinates.

We normally assume that the X & Y axes are both along the surface of the earth and Z is straight up. In ‘free space’ (far away from the Earth if you like), there is no surface to refer to, so X, Y, and Z are chosen to make the model convenient to work with.

By specifying the X, Y, and Z coordinates for each end of a straight wire we have completely specified where the wire is. If the Earth is the X-Y plane, we’re exactly specified where the wire is relative to the Earth.

Consider two points in our 3D system. Point P is at X=3, Y=0, and Z=5 (also expressed by convention as 3,0,5). if the measurements are in feet, then this means from the origin, we are 3 feet in the positive X direction, zero feel in the Y direction, and 5 feet up along the Z axis like this:

Every wire in our models will be specified by it’s end points as X, Y, and Z. In our next installment, we’ll talk a bit about doing more than just defining the points.

————– A small digression ————–

There are lots of ways to specify the coordinates of a point. In the Spherical Coordinate system, we measure the distance directly from the origin to the point called R which is known as the radius. We measure the angle from the X axis called phi ‘Φ’ and from the Z axis called theta ‘Θ’.

Why bother with different coordinate systems? Primarily for convenience. Some problems are more naturally expressed in one coordinate system and not so conveniently in others.

For example, consider an isotropic source radiating at the origin. Over time, electromagnetic waves will propagate straight outwards form the course. If we’re following any single ray, it would be along the radius line and Θ and Φ will remain constant, so a spherical coordinate system would be a natural way to specify the situation. Mathematicians and Physicists frequently select a coordinate system like this to simplify the mathematics.

For our work in Antenna Modeling, we’ll consistently use the X,Y,Z system (also called the Cartesian or rectangular coordinate system). You should be aware that other systems may be in use for special purposes in material you pick up to read.

Categories: Antenna Modeling
1 Comment »

I use the linux system for a lot of serious work because it is a very friendly environment for scientists and engineers. I can get nearly any language I may need and if I’ve got source code, I can compile nearly any program I might like to have.

The downside of working on Linux is that software like EZNEC and NECWin isn’t available at affordable prices. As such, I generally work on Linux with the nec2c. My standard platform for such work has been Ubuntu 6, but I’ll be upgrading to Ubuntu 8 soon. I run Ubuntu on several platforms:

1. I run it in a virtual window on my Macbook Pro and run it there

2. I have a dedicated PC behind my desk which boots to Linux and is my harder work processor.

I picked up my copy from the Unofficial NEC Archives which has a lot of very useful NEC stuff there. There is a compiled version of nec2c available there that runs on Debian based systems (such as Ubuntu), so it’s worth checking out. I like to recompile from source to make sure I know how.

Another useful web site for Amateur Radio tools for Linux is the Linux Amateur Radio Software Database where you can find nec2c and programs to do most anything you’re interested in doing on your computer. While you’re there, take a look around at some of the other packages available for Linux.

Categories: Antenna Modeling
4 Comments »