On Feb 14, 12:13 am, Roy Lewallen wrote:
snip

The short answer is that I don't think anyone really knows. I'm
convinced that the program accurately calculates the field from the
antenna and environment specified by the model. But there are some
pretty significant ways in which the model doesn't represent reality.
EZNEC uses the NEC ground model which is highly simplified - its ground
is perfectly flat, homogeneous to an infinite depth, and infinite in
extent. Real ground is curved and stratified with many layers of
sometimes highly differing conductivity and permittivity. Besides the
deficiency of the ground models, there might be some interesting
phenomena like ground wave energy following the ground for a while, then
launching some distance from the antenna. This wouldn't be modeled
properly by EZNEC or NEC. And although polarization is rotated during
ionospheric propagation, maybe there's some inherent advantage to
launching a vertically polarized signal. EZNEC and NEC make no attempt
at modeling propagation. Anecdotal evidence seems to find more of a
disparity between model results and observations at low frequencies (80
meters and below) than higher frequencies. Whether this is due to the
greater ground skin depth at lower frequencies, different propagation
effects, or maybe just the vagaries of anecdotal reporting, is something
I don't think anyone knows.

Roy Lewallen, W7EL

Thanks Roy, but darn, I was hoping that was not the answer. I was
hoping this was a subject someone had worked through definitively. Oh
well. What counts is how well an antenna works, not what calculations
show.

At no point did I think that EZNEC and NEC2 were busted. The output I
saw matched expectations arrived at from digging through text books
and scientific papers. I was eventually looking for a clue as to what
the NEC2 algorithms might be missing. I found the "leaky ground wave"
thing for lower frequencies an intriguing idea. I expect that above
about 10 MHz, where ground wave propagation becomes a fairly minor
consideration, NEC2 should provide a fairly accurate prediction of
vertical antenna performance.

Gary - N0GW