On Nov 15, 6:23*am, Art Unwin wrote:
Cebic found when comparing different style programs that some behaved
well in certain circumstance where others did not. Yet all antenna
programs
are based on the use of Maxwells equations where all programs should
have the same results, after all Maxwells equations are exact and not
fudged. One of the reasons is that since Maxwells laws are exact
radiators used must be resonant at repeatable points designated as a
period.
* *Fact is that most users use fractional wavelength designs, usually
a half wavelength, that is not resonant at repeatable points where
the area around the datum line of a sine wave is never equal when
generated around a tank circuit.
* * The reason for this is "voltage over shoot" which gets smaller
with every cycle but never disappears. Thus when programs are used
based on fractional wavelength radiators the results will never show
100% accountability and in fact efficiencies derived will be in the
order of 92%!
* If the radiator is of a wavelength then one is not using a "fudge"
figure
in the calculations and *then becomes possible to attain total
accountability with efficiency of 100%. regardles of what type program
is used.
* *If one is to use exact equations, as are Maxwell equations, then
one must also use measurements that are also exact and repeatable and
that is definitely not fractional wavelengths!
*What one gains from this aproach is that any radiator of any shape,
*size or elevation can provide figures in the order of 100% as long as
the radiator is a multiple of a wavelength where it is *resonant at
exact and repeatable measurements.
If anybody can give pointers that refute the accuracy of the above I
would be very interested in hearing them

the key is that while all the programs are based on maxwell's
equations, it is impossible to implement maxwell's equations with 100%
accuracy on a digital computer. this is true of any and all
simulation and modeling programs for electrical or mechanical design.
all such programs make approximations and take shortcuts to reduce
calculation time while maintaining some minimum level of accuracy and
precision. it is important to understand the assumptions and
simplifications that have been made in order to make proper use of the
programs. typical traps in antenna simulations are that they don't
like very small or very large length/diameter ratios... so using them
for extrement long or short wires or very fat or very thin wires may
produce results that aren't realistic. many of them also don't like
very small spacing between wires, this is where most optimizer
programs fall apart, they start moving wires close together and get
strange results like super gain or unrealizable narrow beam patterns,
often accompanied by a very low feedpoint impedance.

most reputable programs like NEC have been validated very diligently
over many years and their accuracy is well documented... as are the
restrictions and assumptions that apply, but you have to read ALL the
documentation, not just the quick start guide. Other programs like
mininec, ao, yo, yagimax, and others make even more simplifications
and therefore added restrictions so they can run on a desktop
relatively quickly. unfortunately they don't always document the
limitations as well as the professional level products. after all the
professionals have millions of dollars riding on the accuracy of
designs, hams have only pennies, so it just doesn't pay to write lots
of documentation or do lots of testing that won't be read for ham
users.

so, while all the programs must be based on the same equations, the
results they generate, especially in the fringe cases, may be vastly
different. remember two maxims... 'garbage in - garbage out', and
'you get what you pay for'.