In W8JI's very rural area minimum main lobe beamwidth is more valuable to
him than minimum sidelobes. Tom's choice of 330 foot broadside spacing in
a very rural area with little or no local RFI makes lots of sense.
I actually use a little more than 330 ft now in that array, and considerably
more spacing in broadside Beverage arrays.
Half-wave spacing forms two nulls at ground level, directly of the sides of
the array. You have ONE null on each side at ground level. Most people fail
to realize the side nulls at half wave spacing are at ground level only. It
is a less than perfectly deep null at wave angles above earth level.
Wider spacing, compared to 1/2 wave spacing, forms a null cone. This
provides **two nulls** at ground level off each of the sides (now you double
the chances of groundwave noise being in a null). Even more important, the
side null forms a cone that reaches maximum elevation off the side.
This means directly off the side of an array, WIDER spacing gives a deeper
null. This is contarary to what most people assume, because they look at the
null as a ground wave side problem. People tend to overlook the fact that
distant signals and noise come from angles higher than zero degrees
elevation.
They also overlook the fact that a wider spacing provides four groundwave
nulls (two on each side), instead of just two nulls (one on each side) of
1/2 wave spacing.
I can't think of a single case, besides a single groundwave noise source
directly off the side, where 1/2 wave spacing would be an advantage.
<<<
If you look at end-fire cell patterns, you will find closer spacing gives
more directivity. It is easier to have a wider null area with somewhat
closer spacing.>>>
Right.
End fire spacing has essentially no effect on beamwidth and sidelobe
levels, so Tom's choice of 70 foot end-fire spacing makes little if any
measurable difference compared to 130 foot end-fire spacing. Larger
end-fire spacing (up to 1/4 wavelength) is somewhat more forgiving of
phasing errors and mismatched signal levels and slightly more
efficient.>>>
Actually there is another very common myth or mistaken assumption about
phasing. We assume phase difference between elements should be 180-s where s
is electrical degrees spacing. We assume with quarter wave (90 degree)
spacing we want 180-90 = 90 degree phase difference, or with 45 degree
spacing we want 180-45 = 135 degree phasing. Once again, despite being in
countless articles and books, this is almost always NOT optimum. The only
case where it is optimum is where we want a single zero wave angle null
directly off the back.
This is almost never the case, because as with 1/2 wave broadside spacing,
this forms only ONE null at zero elevation. By increasing phase delay we
split the back null into TWO nulls at zero angle, and have a null cone that
is elevated directly off the back about the same as the zero elevation nulls
are angled off the side.
It's really pretty silly, when we think about it for a while, to design
skywave systems or noise rejection systems that have a single groundwave
null peak. What almost any case demands is a null covering the widest
possible area, and to absolutely be above zero elevation. This not only does
a better job of notching unwanted signals and noise from skywave, it also
doubles the null area on the ground for local noise.
If I had multiple groundwave noise sources, I would not use 1/2 wave
broadside spacing. I would use wider spacing.
If I had skywave side signals to reject, I would not use 1/2 wave broadside
spacing.
The only case I would use 1/2 wave spacing would be ground wave noise
exactly off the sides, and nowhere else.
If I had multiple groundwave noise sources, I would not use 180-s phasing or
1/4 wave endfire spacing. I would use narrower spacing and use a phase delay
of some number greater than 180 - s.
If I had skywave signals or noise to reject, I would not use 180-s phasing
or 1/4 wave endfire spacing. I would use narrower spacing and use a phase
delay of some number greater than 180 - s.
The only case I would use 180-s phasing would be a ground wave signal
exactly in line with the element's endfire spacing. That would be a single
signal condition.
This about the problem in three dimensions, and it is easy to see why 1/2
wave broadside and a 90 deg space 90 deg lag system is almost never the best
choice. Even my 45 degree spaced elements are not delayed 135 degrees, but
rather some larger delay amount. This actually increases null depth and the
statistical likelihood something unwanted is in a deep null.
73 Tom
_________________
Topband Reflector
|