[TowerTalk] How accurate is EZNEC?

Tom Rauch W8JI@contesting.com
Thu, 6 Sep 2001 07:19:59 -0400

Hi Richard,

> question the accuracy of such programs.  Such as: can EZNEC or similar
> modeling programs accurately determine which quad array: 1/4 waves,
> 1/2 waves, or 5/8 waves... will outperform the other in real life? 

EZNEC, like any other modeling program, is a tool. The results we 
get with any tool depend on how well we understand how to use 
the tool. This especially includes modeling phased systems, where 
the user has a choice of both current or voltage ratios and phase. 
The user is the biggest limitation, and persons with deep 
understanding of how systems actually work can model "new" or 
"unconventional"  ideas faster.  

For example, people often try to model elements that are not fed at 
current maximas in phased arrays the same way they do current 
fed elements, using current sources not placed at the current 
maxima in the element operating with well known current and 
phase ratios. Doing so results in incorrect current and phase 
distribution in arrays, and the conclusion often is the array "won't 
work" when the modeller is actually at fault.

This common modeling error appears mostly with shunt-fed 
elements, or non-1/4wl verticals in phased arrays, and leads to the 
old wive's tales that "shunt fed towers can't be phased" or that 
"1/2wl verticals won't phase correctly". Even though nothing is 
further from the truth. I phased a shunt-fed 1/4wl tower against a 
dissimilar height series-fed T antenna and got well over 20dB F/B. I 
can also model that system and it works.  

Models by nature, are shortcuts but virtually none of the short-
cutting is in the antenna calculations of NEC based programs. 
There are shortcuts in ground calculations because EZNEC and 
other NEC based programs treat the ground as a homogeneous 
medium. In truth, soil characteristics vary widely with depth and 

Whenever there is a shortcut, there is potential for notable error. I 
do not have great confidence in the ability to accurately calculate 
loss of conductors very near earth. Some real world data seems to 
bear that out, although data is sparse.

> Again, I am not against using EZNEC for a general idea of how an
> antenna array might work.  But I have a serious problem believing
> EZNEC has all the answers and much prefer to hear or read about actual
> side-by-side antenna comparisons.  I believe it's the difference
> between "projected performance" and "actual real life 

There are far more errors in using S meters, and making actual FS 
measurements, than most people realize or are willing to admit. 
Too many errors to talk about even.

No wonder some think they have supergain antennas that break 
the rules of physics. It is often difficult to accurately know if one 
antenna is actually better than another, let alone determine what 
the dB difference is!

My favorite example is a test I did on 75 meters between a G5RV 
and a dipole. The dipole won easily in dozens of on the air reports 
when I announced which antenna was which during the test. 
Problem was, I wasn't changing anything at all except what I told 
the other guy. When a few people couldn't see the change on an S 
meter, they told me the dipole audio was better!


73, Tom W8JI

List Sponsored by AN Wireless:  AN Wireless handles Rohn tower systems,
Trylon Titan towers, coax, hardline and more. Also check out our self
supporting towers up to 96 feet for under $1500!!  http://www.anwireless.com

FAQ on WWW:               http://www.contesting.com/FAQ/towertalk
Submissions:              towertalk@contesting.com
Administrative requests:  towertalk-REQUEST@contesting.com
Problems:                 owner-towertalk@contesting.com