On Wed, 22 Nov 2006 18:59:31 -0500, "C. J. Clegg"
wrote:

On Wed, 22 Nov 2006 15:15:35 -0800, Richard Clark wrote:

Perhaps you would care
to elaborate how the simplicity of two extra wires has been trumped.

I'll try, though from the tone of your message it sounds like your mind is
made up. :-)

My mind is made up? I've repeatedly wondered why you have approached
this with a defeatist attitude.

We will be operating on many different frequencies across the range of 4
to 9.

Again, this has been apparent from the beginning.

I don't even know (yet) how many different frequencies will be in
use (they won't tell me).

That doesn't matter all that much, except to anticipate failure.

So I can envision many pairs of dipole elements, each cut for a certain
frequency in the range, and laid out like the spokes of a wheel.

If you re-read my posting, I've done nothing more complex than to add
TWO more wires. ALE may easily jump between 200 frequencies, but
there is absolutely nothing about that which demands a resonant
frequency for each of them.

I have plenty of land here but I don't have ready supports for that kind
of an array.

That's what I mean by "impractical".

Then the solution is not impractical by any definition, you are simply
over embroidering the problem with a slavish interpretation of
necessity. One pair of wires cut to a low end, one pair of wires to a
high end, both pairs fed at the same point. It may take as many as
four pairs (I doubt it), but to abstract this wildly to 200 goes
beyond the pale when a skeleton biconical could easily accomplish this
with flat response (over a much larger bandwidth) with only 16 pairs
of equal sized wires.

This cage monopole:
http://home.comcast.net/~kb7qhc/ante.../Cage/cage.htm
is flat over 5 Ham bands.

This discone:
http://www.qsl.net/kb7qhc/antenna/Discone/discone.htm
operates flat over more than two octaves of bandwidth. With scaling,
I can count at least 55 discrete frequencies that would fall into the
2:1 mismatch region - and this say nothing of those frequencies
between them, nor of their end points which could be made to span 4 to
9 MHz.

Yes, a lot of wire, but use less wire for a rougher approximation. If
you are looking at an arbitrary 50% efficiency, a 5:1 circle
encompasses a lot more points for less wire.

73's
Richard Clark, KB7QHC