Towertalk
[Top] [All Lists]

Re: [TowerTalk] 40-2CD´s distance stack

To: cx6vm.jorge@adinet.com.uy
Subject: Re: [TowerTalk] 40-2CD´s distance stack
From: jimlux <jimlux@earthlink.net>
Date: Fri, 19 Nov 2010 09:33:14 -0800
List-post: <towertalk@contesting.com">mailto:towertalk@contesting.com>
Jorge Diez - CX6VM wrote:
> Hello,
> 
>  
> 
> What´s the minimum distance between antennas to start playing with HFTA for
> a 3 stack of 40-2CD´s?
> 
>  
> 
> HFTA is doing well with a distance of 33 ft and also 16 ft!  but this is too
> close.
> 
>  




HFTA doesn't take into account interactions in the near field between 
the antennas.  It just looks at the far field pattern, assuming a 
particular pattern for each of the antennas in the stack.

As to where the assumption of "antennas are isolated from each other" 
starts to break down?  The classic "it depends".. certainly, anything 
with a half wavelength or so will have some coupling.  Antennas with a 
lot of gain have a lot more stored energy in their near field, which 
makes them more susceptible to interaction.

The only way to know for sure is to model it in something like NEC.  It 
doesn't have to be a really accurate model (e.g. you don't need to know 
taper schedules, boom corrections, etc.).. what you are looking for is 
the magnitude of change when you put the second antenna in.  Model it 
with single conductors of some reasonable compromise diameter..  Run the 
pattern at some frequency you're interested in (you don't really care 
whether the antenna has good vswr, at this stage)..

You DO want to check the pattern and make sure that the F/B and F/R 
ratios are in the ballpark you expect (they'll be different, because 
your model is pretty sketchy, but it should be close)..  You could 
change the modeling frequency up and down to get where you want (or 
scale all the element lengths by some small fraction).. the idea is to 
get relative element currents about the same as they are in the real 
antenna.

That way, the magnitude and phase of the near field being modeled is 
going to be comparable to what the real antenna has.

Now, throw in the second antenna, and compare the patterns.  If the 
side/back lobes change by a bunch, bingo, you've got interaction, and 
then you can decide if you want to improve the accuracy of your model, 
or approach it a different way.

Or, you can look at the element currents and see if they change a lot. 
You can look just at the current at the center of the element (put 
excitation sources at the center of each element, but set to very small 
voltages, and you don't have to hunt through the segment currents to 
find it.. it will be in the table of feedpoints)

In particular, look for significant changes in the phase of the element 
current, especially if that element is carrying a current that is, say, 
more than 20% of the driven element.

Small changes in phases and magnitudes have almost no effect on the 
forward gain, but will have a much bigger effect on side and back lobe 
levels.  Turning a -20dB null into a -10dB null doesn't take much power 
in the wrong place.  That same 10% of the total power is only 0.5 dB in 
the forward direction.


The other thing to do with a NEC model is to not drive the second 
antenna (i.e. set its feedpoint excitation to 1 millivolt or something 
small) and only excite the first antenna.  Then see how much current is 
induced in the second antenna's feedpoint. If there's no interaction, 
then there won't be any current.  If the current is <10% of the current 
in the primary antenna, then you're looking at any "errors" from the 
second antenna being 20dB down, and probably negligible.
_______________________________________________



_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk

<Prev in Thread] Current Thread [Next in Thread>