It sounds like you might have made the mistake of connecting a wire
directly to Sommerfeld or reflection coefficient ground. Doing this with
NEC-2 (or EZNEC) produces a resistance of unpredictable and meaningless
value at the connection point, lowering the indicated field strength by
an unpredictable amount. (EZNEC gives you a warning message when you try
to do this.)

EZNEC provides an option not available in NEC-2, a "MININEC-type"
ground. This functions as a perfect ground when calculating impedances
and currents, but uses the user-specified ground constants (conductivity
and dielectric constant) when calculating the pattern. It simulates an
antenna with a lossless ground system, allowing you to separately see
the effect of ground conductivity on the pattern without the magnitude
of the field being affected by changes in the ground system loss. The
best you can do with either EZNEC or NEC-2 if you want to include ground
system loss is to include radial wires just above the ground in the
model and connect the vertical to them. Then, however, any differences
you see will hold only for that particular ground system -- and, the
above-ground approximation isn't a terrifically accurate representation
of a buried system.

Using the MININEC-type ground with EZNEC (and only 10 segments, so this
can easily be done with the demo program) and starting with the Vert1.ez
example file, the gain of a resonant (~0.24 wavelength) high vertical at
7 MHz with "average" ground is -0.0 dBi at an elevation angle of 26
degrees. Changing the height to 0.625 wavelengths (easily done by first
changing Units to Wavelengths) produces a maximum gain of 1.19 dBi at 15
degrees elevation angle. The 1/4 wave trace protrudes outside the 5/8
wave trace only from about 25 to 41 degree elevation.

But more interesting is the gain difference at various low elevation
angles. The comparison is easily done with EZNEC v. 4.0 by saving the
trace from one antenna, then superimposing that pattern on the pattern
of the second antenna. By clicking the name of the superimposed pattern
in the 2D plot window, a new entry appears in the data box showing the
difference between the two at the angle of the cursor.

It turns out that the 5/8 wave really shines at really low angles when
the ground is poor, but isn't so impressive when the ground is very good
-- at least at 7 MHz. Over average ground, the gain difference is at or
just above 3 dB up to about 10 degrees. (My explanation of the reason
for the difference over real ground was overly simplistic. I apologize.)
Above 10 degrees, the difference decreases. Over poor ground, the gain
difference is about 4.5 dB up to 5 degrees, and over 4 at 10. So if you
have poor ground, you can really benefit from a higher radiator. Over
very good ground, though, the difference is about 2 dB up to 5 degrees
elevation, only 1.2 at 10 degrees, and less than a dB at 12 degrees and
above. So it might or might not be worthwhile to extend the height of a
tower for that amount of benefit.

Those figures depend on frequency, too, and the pattern shape varies
considerably with frequency and ground characteristics. So modeling the
particular situation would be a good idea before doing any expensive and
extensive tower lengthening. In all the cases I looked at, however, the
5/8 wave vertical did show some gain over a quarter wave vertical up to
at least 14 degrees. Whether the difference is worth the added height is
up to the individual.

Roy Lewallen, W7EL

Richard Fry wrote:
__________________

I investigated your concept statements using NEC-2 models of 1/4-wave
and 5/8-wave verticals in the 40m band (7.3MHz), working against the
same infinite ground plane of "Average" parameters.

* The 5/8-wave vertical has a peak gain of 0.2dBi,
16 degrees above the horizon.

* The 1/4-wave vertical has a peak gain of -6.4dBi,
26 degrees above the horizon, and its entire radiation
envelope is always within that of the 5/8-wave.

I don't know which range of elevation angles is considered most useful
for skywave paths on 40m, but it would appear that with equal tx power,
a 5/8-wave vertical always will have a usefully better skywave than a
1/4-wave vertical over a typical ground plane -- and probably by more
than 3dB.

If you could check my conclusions on this I'd be grateful.

RF