David J Windisch wrote:
. . .
Modeling it with EZNEC shows that, at lower frequencies, it can have more
radiation at lower angles than some other simple antennas at the same height
above ground, and that shifting the feedpoint away from the center of the
phasing line produces a uni-directional pattern.
. . .

I don't think you'll see both those effects (lower radiation angle and
unidirectional pattern) at the same time.

You'll find that the elevation pattern of small to moderate sized Yagis,
log periodic antennas, and most other horizontal arrays is essentially
the same as that of a dipole, at least at the lower elevation angles.
The reason is that most horizontal arrays have a broad free space
pattern in the vertical plane -- not that much different than a dipole,
in the forward array direction. The W8JK is unusual in that respect. Its
free space vertical pattern is noticeably narrower than most other
moderately sized horizontal arrays, so it has a narrower elevation
pattern when mounted over ground, too. The net result is a greater
concentration of radiation at lower angles.

Unfortunately, the cost of this narrower vertical pattern is a
bidirectional horizontal pattern, resulting in half the applied power
being wasted unless you happen to be talking to two oppositely-located
people at once. So for a given power input, the gain is 3 dB less than a
unidirectional antenna with the same lobe width, or the same as a
unidirectional antenna with a considerably wider lobe. The only way to
fairly compare the gains is by modeling and looking at the field
strength of the antennas at the azimuth and elevation angle of interest.
(Loss must be realistically included in the models, too.) It's entirely
possible that the advantage of concentration of radiation at lower
angles is more than offset by the lower gain due to its bidirectionality
and/or loss (see the next paragraph). Of course, you can change the feed
as Dave suggests to get a unidirectional pattern. But then your free
space elevation pattern becomes a broad cardioid, and you end up with a
dipole-like elevation pattern over ground.

Among the advantages of the W8JK are simplicity and the fact that the
pattern changes very little over very wide bandwidths. On the down side
are the bidirectionality and, particularly when spacing is close, a low
feedpoint resistance and rapid change in feedpoint impedance with
frequency. The feedpoint resistance can be low enough that wire loss
becomes a concern, and if not, matching network and perhaps even
feedline loss can become significant. You're almost certain to be
running a very high SWR on the feedline, so very low loss line is a must.

Roy Lewallen, W7EL