Yes, thankyou
Art
Cecil Moore wrote:
art wrote:
Cecil, I can't get this on my computor just a mix of letters here and
there Any suggestions

Is your email address in your heading correct?
If so, I'll email the file to you.
--
73, Cecil http://www.w5dxp.com

O.K. I believe i left things as a Gaussian field with what can look
like a bunch of pencils tightly held together inside a arbitary
boundary
which is very closely shaped like the periphery of the bunch of
pencils.
Now each of these pencils represent a charge running along a resonant
element
and at a specific point in time each of these resonant elements have a
like charge
as all the other resonant elements together with a direction arrow
representing the phase of the charge at a particular point in time. Now
all these charges are in equilibrium
so when the cluster is excited with a time varying field each charge is
going to seek the easiest way thru the arbitary boundary but the
question is where? It would apear that the arbitary boundary is strong
every where around the periphery of the charges. Ideally we would like
to chose fore ourselves where the leakage points occur so that we know
which directing the radiating vectors are. If we look closely at the
phases of the individual charges you will not that there are two
pointing to the same direction which ideally is where we want the main
lobe to be so we have to make a weakness in the wall or arbitary border
to have a weakness where one of the vectors need to breach the border.
To do this we insert a detuned element near the point we want the
vector to escape . The vector that we are aiding to escape is the
summattion of the radiation that normally would radiate out of the rear
of the array but we are trying to harvest this radiation to complement
the other forward vector in a similar way that a stack antenna does but
ofcourse we only have one feed point.
When we feed one of the enclosed resonant elements the charges will
breach the boundary in two places in a common direction and when
breaching the border the Eand H field transformation starts. I.e. the
directive vectors both point the same way until the internal charges
are depleted./ Sounds perfect....... except for one thing, we don't
want to put a lot of elements up into the air as a cluster infact four
or five will do us fine.
Thus with just five charges within the gaussian field it can be seen
that the strength of the border is very shaky and tho we still have two
vectors pointing the same way some of the reaward radiation is going to
escape tho some will make it to form a second vector.
As soon as the gaussian field is excited with a time varying field
there will be two vectors emerjing in the predesigned direction i.e.
forward and there will be a scattering of leakage vectors to the rear
all of whichare off to generate a radiative field in their exit
direction.
So instead of a yagi type array that is bound to a common plane with
detuned elements
that are coupled to arrive at a final direction and at the same time
generating E/H fields
the Gaussian style really has a single dipole represented by the
Gaussian cluster which has predesigned directional abilities at the
instance of applying a time variant field.
Of course one can decide to add detuned elements around the array but
basically the arrays work is already done.
So what has this Gaussian array done for us? If you look in the antenna
handbook you will find a whole chapter that laments about the fact that
compromises must be made with Yagi designs and even if you arranged for
maximum gain over a stretch of frequency something comes along that
takes some of that gain away from you.The interesting point about the
Gaussian array is that the gain curve is relatively flat across the
band in question and more important the SWR curve is pretty much a
mirror of the gain curve, something I suppose
like a band pass filter. Another point to notice is that the placement
of elements in such an array is one of choice and not limited to a
single plane as per the yagi and also not limited in the individual
direction of the element cluster.
For a simple illustration of a Gaussian antenna we will place five
elements parallel to each other and on the same mast. Since it is
difficult for most programs to come up with thye resonant length of the
element cluster I will provide a sample which really is only using two
of the X Y Z dimensions of the computor program.

Center Frequency of use 14.25 MHz
element length height

1 209.46 927.1 feed element
2 198.25 973.97
3 172.78 822.86
4 219.83 964.4
5 185.53 922.4

Note you can readily see which element is detuned but note the varience
of the other elements in length where each is resonant!

Element diameters were 1.3 inches tapered and all dimensions are in
inches
The above will give you something to play with and I believe EZNEC is
available for sample purposes.
When you make this type of array it is not only different horizontal
and vertical patterns you can get but also circular and with field of
different shapes. Bearing in mind that the element placements are
random try the above array but swinging each element around
progressively
to form a star shape when viewed from above tho ofcourse element
lengths will have to change to assume equilibrium status. For the more
adventurous you can have an arrangement similar to a yagi that
stretches out horizontaly though not necessarily on a single plane or
if your program is limited in size just play around with two or three
elements.
As you know I use the AO program by Beasely which really is not tied in
anyway solely to a yagi design and will infact determine element
lengths ,positions etc simply by asking for maximum gain or whatever
and ofcourse start at 50 ohms impedance since unlike the Yagi you wont
find yourself with terrific gain but only a few ohms impedance.
So have at it and play with it a bit to see exactly what such an array
can do
Again I use the AO set up which makes the above antennas a breeze to
design and I haven't begun to probe all the variations

Art KB9MZ...........XG

snip

"art" wrote in message
ps.com...
technobabble snipped
Center Frequency of use 14.25 MHz
element length height

1 209.46 927.1 feed element
2 198.25 973.97
3 172.78 822.86
4 219.83 964.4
5 185.53 922.4

ok, if i am reading this right the elements are stacked vertically above and
below the fed element?
assuming the distances are in inches, your element 5 is about 5" from the
fed element? and elements 2 and 4 are about 8" apart??

assuming the above is correct, you have probably rediscovered the hazards of
runaway optimizations. aka: how to get super gain without really trying!
you will most likely find that the bandwidths are unusually narrow, and the
feed point impedance is extremely low or highly reactive... all signs of a
design that can't be fed without extremely high losses... and would only
work on one frequency, if at all.

On Wed, 6 Dec 2006 22:19:01 -0000, "Dave" wrote:

you have probably rediscovered the hazards of
runaway optimizations. aka: how to get super gain without really trying!

Hi Dave,

Super gain? I improved the gain and the feed Z by throwing away 4
wires. I hope "AO program by Beasely" is not slurred by this example.

Center Frequency of use 14.25 MHz
element length height

1 209.46 927.1 feed element
2 198.25 973.97
3 172.78 822.86
4 219.83 964.4
5 185.53 922.4
specification for a classical waste of time. [However, it is
certainly the poster child for first establishing a reference against
which improvement is measured.]

73's
Richard Clark, KB7QHC