[Top] [All Lists]

[TowerTalk] Ant Modeling (long)

To: <>
Subject: [TowerTalk] Ant Modeling (long)
From: (David Robbins)
Date: Fri, 18 Jan 2002 01:01:27 -0000
You are obviously right.  No one has built and measured ALL the possible
combinations of antennas and surroundings so they can say for sure that
the simulation numbers are absolutely right in all cases.

However.  They have evaluated the basic code and methods used (at least
for NEC which is the core of most of the better programs I think) in
enough cases to know that it provides good results.  Will any program
accurately predict your antenna performance half over your roof with
real ground slopes and other nearby junk, no.  Can you accurately
measure the performance of your antenna in its current surroundings, no.
So of what value would either of those things be in your case?
Absolutely none!

So, why bother putting all the money, time, testing, etc into developing
programs that can't predict real world performance?  Because its
infinitely cheaper and faster than building and measuring all the
possible variations, tabulating them, finding the trends, and optimizing
from that, then retesting, etc, etc, etc, ad nauseum.  But what good is
a computer model if it doesn't represent the real world?  Well, several
things, as long as you keep two things in mind... 

First.  You only compare modeled antennas with other modeled antennas.
Now why is this valid?  As long as you are using the same software and
modeling antennas that behave reasonably like each other the comparisons
are normally very good.  They can be used to optimize elements, evaluate
changes in height, and test other changes as long as you stay within the
limitations of the models.

Second.  You don't compare one modeled antenna to some other range
tested antenna.  Unless you have carefully calibrated the range to a
common antenna that is similar to the one being tested the results just
won't compare.  All the real world things affect the range test and you
just can't model them all accurately enough to do a valid comparison.

So what does this tell you?  First, you can't compare the performance
figures quoted by one person or company to those quoted by anyone else
unless you know that they are using the same type of software and the
same set of assumptions.  Second, you can't predict how an antenna will
work when hanging half over your house.  Third, antenna modeling is good
if you are comparing designs or trying to optimize a particular design
as long as you are careful to stay within the limits of the software you
are using.

David Robbins K1TTT
AR-Cluster node: 145.69MHz or telnet://

-----Original Message-----
[] On Behalf Of Tom Osborne
Sent: Thursday, January 17, 2002 23:05
To: TowerTalk
Subject: [TowerTalk] Ant Modeling (long)

Hi All

In looking over previous posts, there seems to be an awful lot of
credence to computer modeling of antennas.
I've seen things like "A three element beam with lengths of (n)
and spacing of (n) will give 7.13 dbd and 18.34 db FB.  How do we
know this?  Has someone built and tested every conceivable
spacing, length, height and checked them with a field strength
meter.  How about 4 elements on a 32 foot boom.  There are
hundreds of different configurations available.  Then after that
is figured out, What if the boom diameter is changed?  Do all the
readings change again.  Then how about raising it up 5 feet.  Do
all the readings change again?  It seems to me that there are
thousands of different configurations when you figure boom size,
element size, taper, element spacing, element length, ground
below the antenna, terrain, etc.  Surely each configuration
hasn't been verified manually.  But if I put in "2 elements on 20
on a 12 foot boom", I'd probably get a gain of 5.34 dbd and FB of
12 db or something like that.  It seems to me that in order to
say that such and such an antenna with a certain length of
elements and height has a certain amount of gain, you'd have to
measure 360 degrees around the antenna for side lobes, rear
lobes, etc, the do the same thing from 0 degrees to 90 degrees
360 degrees around the antenna to check the take-off angle.  
Where does the data that we take so religiously come from?  I
know we can model a 3 or 4 element antenna and move each element
a couple of inches or so and see different amounts of gain and
If these antennas are modeled at VHF frequencies, it would be
hard to make the same assumptions at HF.  It seems as there are
just too many variables to take the data as absolute.  If I
wanted to model my 20 meter beam, how would I take into effect
the fact that half of it overhangs my shack roof and the other
half is directly over one of my 40 meter verticals, and about 20
feet away in a tree is my 80 meter inverted "V."  Plus the ground
slopes to the N.E. at about a 25 degree angle for a half mile or
so.  The hill in back goes up another couple of hundred feet. 
How in the world would I ever be able to take into effect
everything that effects the antenna.  It just seems to me it's
hard to put in certain dimensions and get an absolute answer.  73

Where do you get ICE bandpass filters & beverage matching boxes?  The
same place that pays for the hosting of this list:  The eHam Store.
Order online at

FAQ on WWW:     
Administrative requests:

Where do you get ICE bandpass filters & beverage matching boxes?  The
same place that pays for the hosting of this list:  The eHam Store.
Order online at

FAQ on WWW:     
Administrative requests:

<Prev in Thread] Current Thread [Next in Thread>