[TowerTalk] 4-square questions

[jimlux]

On 3/6/17 5:16 AM, Bob Shohet, KQ2M wrote:
> Hi,
>
> I have a few questions that I have not seen addressed anywhere and I
> would appreciate thoughts and comments that any of you are willing to
> share.
>
> I currently use a 4-square for 80 meters (with 3/16 wavelength
> spacing – approx. 52’ between the 63’ long elements and ~ 73.5’
> across the “square” of the array) with three elevated radials for
> each of the four legs.  The feed points are about 12 feet above
> ground and the tops of all 4 elements are attached to ropes that go
> to the trees through which the pull ropes pass.  Two elements are
> almost completely vertical but the other two are vertical for ~ 50’
> and horizontal for the remainder and are attached to the pull ropes
> with cable ties.  (The elements lengths are shortened ~ 3% for mutual
> coupling plus another ~ 2% due to the PVC jacket on the element
> wire)
>
> Directly behind the 4-square the ground is flat and is considered to
> be wetlands.  On BOTH sides (running North to South) and in FRONT of
> the 4-square (to the North), there is a natural ledge formation (I am
> on a hill top) that is  ~ 12’ feet high and runs for ~ 400’ before
> dropping back to ground level.   Assume that the feed points in the
> “front” of the array (to the North) are about 10’ – 20’ from the
> ledge on either side.
>
> A simple “picture” of the elements of the array and ledge is as
> follows: . . . N O R T H
>
> LEDGE                                           LEDGE Continues
> Continues
>
>
> L       NW                           NE         L E
> E D                                                      D G
> G E                                                      E SW
> SE
>
>
> FLAT and wetlands ...........................
>
>
> S O U T H . . . .
>
> Assume that the top of the ledge is ~ 12’ high and that the feed
> points are also ~ 12’ off of flat ground and are looking at the top
> of the ledge.
>
> Putting aside questions of the phasing system (I use the Comtek box)
>
> 1)  Are the minimum wave angles of the transmitted and received
> signals primarily determined by the height above ground at the feed
> point?  At the top of each element?  Or some complex calculation of
> the average height along the entire length of the vertical legs?  I
> am sure that having part of two elements horizontal also affects the
> wave angles.
>

The latter, some complex calculation, the vertical pattern is affected 
by the "ground plane" many wavelengths out.  You'd have to model it to 
see what's really going on. And unfortunately, for a vertically 
polarized antenna, you can't use HFTA (because HFTA makes use of the 
fact that for H-pol, you can consider the earth as a perfect reflector 
for the most part)

You might be able to get a qualitative answer (see how much it changes) 
by using the NEC "ground cliff" mechanism
http://www.nec2.org/part_3/cards/gn.html

Set up just one vertical element over flat ground, then change it to a 
cliff/ledge and see what happens.


> 2) If the feed point is at the same vertical height as the top of the
> ledge when the array is pointed towards the ledge, is the minimum
> wave angle affected?
>
> 3) Is the minimum wave angle increased if the feed point of the
> vertical legs is below the top of the ledge?  Is it a straight
> geometry calculation – that is if the feed point were as 0 feet and
> the direction of the array was aimed pointed straight at the 12’ high
> ledge from a distance of 20’, would the effective wave angle be ~ 35
> degrees?  Or, in fact, is the minimum wave angle considerably lower
> since the broadside of the NW – SE legs are effectively radiating to
> the NE and therefore is about an additional 37’ away from the ledge
> than the NE element?

Not really. Consider an antenna as being broken up into a bunch of 
smaller antennas, each with different current and phase.  The far field 
is the sum of the radiation pattern from each of the little sub 
antennas.  This is what NEC does, it calculates the current (mag and 
phase) in each segment, then calculates the far field for each segment, 
then sums all the segments together.

With vertical antennas and uneven ground, it is almost never a simple 
geometry thing.   with horizontal antennas, all the little 
"sub-antennas" are the same height above the ground, so the calculation 
is the same for all of them. Furthermore, for Hpol, the reflection from 
the ground for almost any angle is like a perfect mirror, so geometry works.



>
> 4) I would rather have a ground mounted 4-square with a ground screen
> of about 120 radials and with all four elements being fully vertical,
> but I am concerned about effects of the ledge on the wave angles when
> the array is “pointed” at the the ledge.

To my mind, the big advantage of directional antennas on HF is not so 
much forward gain, but the fact that they have nulls that can be steered 
to block undesired signals.

You can have fairly big phasing errors and the forward gain doesn't 
change much (tenths of a dB), but a phasing error can kill the null 
depth. I suspect that this is why some people swear by 4-squares and 
others swear at them.

Unfortunately for you, your questions are one that can really be 
answered two ways: experiment (build it and see how it works) or some 
sort of finite element modeling.  And the FEM is going to be tricky if 
you want to stick with NEC, because it really doesn't deal with the 
uneven ground surface well.

As mentioned above, you can fool with NEC and use the "second ground 
surface-cliff" and model it to see.   One thing to bear in mind is that 
you can run two models and kind of combine the outputs.

Say you have your antenna system at the north end of a N-S ridge, so you 
have a cliff on 3 sides (W,N,E) and flat ground to the south. You run 
the NEC model with flat ground, and you can get the pattern to the 
south.  Then run it with the cliff (on all sides) and you get the 
pattern to the West, North, and East.

It's not perfect, but you'll at least know if it's terrible or not.

Bear in mind that the cliff ground in NEC does NOT factor into the 
element interaction or current calculations, it's far field only.  So 
the input impedances, etc. are all calculated as if you're on flat ground.

>
> I would appreciate any thoughts/comments that you might have.  Thank
> you in advance for your interest and cooperation!
>
> 73
>
>
> Bob KQ2M _______________________________________________
>
>
>
> _______________________________________________ TowerTalk mailing
> list TowerTalk at contesting.com
> http://lists.contesting.com/mailman/listinfo/towertalk
>