At 10:10 AM 5/9/2005, Al Williams wrote:
>I am trying to understand why the so-called reflector has to be longer
>than the driven element and the so-called director has to be shorter than
>the driven element for the standard configuration Yagi. I have been
>reading antenna books and also trying to rationalize what is going
>on. Secondly, I am trying to rationalize why the longer element is called
>a reflector and the shorter element a director. I posted this to
>Towertalk perhaps two years ago and only got simplistic (?)
>answers. Hopefully some of the great gurus here can provide an explanation.
>
>I think that I understand the fundamental principle of the Yagi i.e. the
>induced currents in the parasitic elements cause a re-radiation in all
>directions. For compass bearing points around the array, the currents in
>all of the elements are vectorially added. In some directions the vectors
>will add, but in other directions the vectors will subtract.
Yes, and the phase of the current in the element is determined by the
mutual impedance, which in turn is determined by the spacing and the
length. So, you've got a phased array, where most of the elements in the
array are excited by mutual coupling, rather than explicitly driven through
some sort of feed network.
>
>What bothers me is that the currents at the end of the elements are low so
>I would think that the ends would not have much effect on the pattern?
That's true. You can make a directional antenna with identical length
elements, and make the phase shifts with reactive networks at the feed
points, for instance. The difference in length is mostly to adjust the
reactive component for that element (and, to a smaller degree, to change
the amount of coupling).
There's two things going on in the classical analysis of a phased
array. You have the "element pattern" (assumed identical for all elements)
that gets multiplied by the "array pattern" (which is independent of the
elements, and is assumed with isotropic elements) to create the overall
antenna pattern.
If you built a direction antenna with very short elements, but with the
same relative phases and amplitudes as the antenna with long elemenets,
you'd get the directivity contribution from the phased array, but not the
directivity inherent in the elements. (That is, the 2dB dipole directivity)
>I think that I have the answer for my second question. For a resonant
>driven element, a longer parasitic element will have vectors on the side
>facing the driven element that will add to the vectors from the driven
>element. Hence the "reflector" appears to reflect the wave. For a
>shorter element the vectors on the side away from the driven element will
>add to the vectors from the driven element. Hence the "director" appears
>to direct the wave.
yes.. and, also there's a historical precedent... the first directive
antennas using this sort of scheme were dipoles in front of a screen, where
it's obvious that the screen is reflecting the wave. Turns out that a
single wire (of the appropriate length) works almost as well as the whole
conductive screen. That wire became known as a reflector.
So, then, (and here it gets real fuzzy, etymologically), it would be
logical to refer to elements on the "other" side as directors.
As for why you see one reflector and multiple directors, I suspect that it
has to do with practical antenna designs. You want to minimize the total
material in the antenna. The driven element is usually close to a half
wavelength (so that it's non-reactive). You could put the parasitic
elements either longer or shorter, and shorter is cheaper, less wind drag,
etc. One can, in fact, design antennas of arbitrary numbers of elements
with the driven element placed in one place or another, as long as the
phases and amplitudes work out. (An interesting experiment is to take an
optimizing antenna design program, a fairly vanilla 3 element beam, and try
driving each of the 3 elements, and let it optimize. You wind up with very
similar results, performance wise.)
Also, bear in mind that tradition plays a strong part in all this. Until
recently, analysis of multi-element parasitic arrays was a real pain in the
rear, so people tended to just use "proven designs" (such as those
developed by NBS) or modify existing designs. Those very early designs
were built by very much tedious hand experimentation on antenna ranges and
very much tedious analysis. If you're going to go to a LOT of work, for a
design that's going to be reproduced a lot of times, minimizing cost, mass,
etc. is worthwhile (leading to the many short director design, as opposed
to the many long reflectors designs).
The math for analysis of multi-element parasitic arrays wasn't really
developed until 40s, 50s, & 60s, long after Yagi and Uda wrote their papers
in the late 20s. As Thiele puts it, "Much is known about the Yagi-Uda
array, but that has been due mostly to experimental data rather than a
method of theoretical investigation." And it wasn't until computers made
the "method of moments" practical (say, the 1970s) that there was much
detailed analysis of Yagis. Harrington's paper on Matrix Methods was
published in 1967, and Thiele's paper on "Analysis of Yagi-Uda type
antennas" wasn't published until 1969.
And, even with all the sophisticated analysis, you're still driven towards
a one reflector, multiple director configuration, because that's cheapest,
lightest, and smallest, for a given performance.
_______________________________________________
See: http://www.mscomputer.com for "Self Supporting Towers", "Wireless Weather
Stations", and lot's more. Call Toll Free, 1-800-333-9041 with any questions
and ask for Sherman, W2FLA.
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|