[TowerTalk] modeling interaction

[Eric Scace K3NA]

Hi Dick --

    Some thoughts:

1.  Simply deleting the source from the unused antenna changes that 
element from a dipole to a continuous piece of metal -- which is NOT 
what you have.

    Rather than deleting the source, you must leave the source present.

    Now you have to decide what impedance the unused piece of coax 
places across the drive point of that antenna at the model frequency. 
(A piece of coax running into the shack, not connected to any 
transmitter or receiver, presents some impedance across the drive point 
of the antenna.) Since this is not an actual system you can measure yet, 
you don't know what this impedance will be.  That may be an issue for 
some lengths of coax.  And it may be different when the antenna is 
unused vs. connected to a receiver (i.e., terminated in a nominal 50 + 
j0 ohm).

    For playing around, you could assume no transmission line is 
connected to the unused antenna.  In this case you can specify that NO 
current is applied to the unused antenna's drivepoint.

    However, beware of an odd twist in NEC code.  If you specify ZERO 
drive current (or voltage), NEC assumes the drive current should be 1A 
at 0°!  (The explanation for this is too tedious to bother with in a 
short note).

    So, the correct way to specify zero current (or voltage) is to 
actually specify a very small value.  I have found that NEC4 will accept 
1e-10 (e.g., 100 nanoamps) is about the smallest value which the 
calculating engine will accept.  I have not tested NEC2 engines in this 
regard.

    Remember, in a real implementation, the feedline (with its balun) 
will put a parallel impedance on the unused driven element that will 
have some impact on the response of the unused antenna.  A proper study 
will need to take this into account.

2.  I assume you are looking at two slices through the driven antenna's 
pattern: an azimuth slice at some important elevation angle, and an 
elevation slice through the main beam.

    In severe interactions, you will see a change in the main beam's 
gain or pattern shape.  This disruption will clearly show up in a change 
of impedance for the driven antenna.

    But many interactions will not show up as a significant change in 
driven antenna feedpoint impedance.  Instead, the main beam changes 
slightly (perhaps a dB) and the lost power shows up in new minor lobes. 
  These minor lobes can occur anywhere -- quite possible in directions 
that you do not see in those two slices through the antenna's pattern. 
Those minor lobes open up the antenna to receiving more QRN and QRM from 
directions which are not of interest to the operator.  To find these, a 
more comprehensive plot of the entire 3-D antenna pattern must be 
examined.  One way to do this has been outlined in my article series in 
the National Contest Journal (parts of which are posted on the NCJ website).

73,
    -- Eric

on 05 Jun 21 15:27 Dick Green WC1M said the following:
> Hi. I'm trying to use EZNEC 4.0 to determine whether there would be any
> interaction between a 2-el 40m beam at the top of a tower and various beams
> placed lower on the tower. For example, I've tried putting a 40-2CD at 105'
> over a 3-stack of 4-el 15m monobanders at 90'/60'/30'. The model showed no
> significant difference in the 15m pattern with or without the 40m beam, and
> no significant difference in the 40m pattern with or without the stack. Same
> story for 3-stacks on 20m and 10m at the same heights. I even tried putting
> a 5-el 15m beam at 100', only 5 feet from the 40m beam, and got no changes
> in pattern for either antenna. Seems to me that there would be some
> interaction between 40m and 15m antennas spaced that close. Is that
> incorrect, or is there a problem with my model?
>  
> FYI, I'm deleting the source for the antenna that's not on the band in
> question. In other words, if I'm checking the pattern for the 15m stack, I
> leave its sources intact and either delete the source for the 40m beam or
> set the V to 0 (results are the same either way.) If I'm checking the
> pattern for 40m, I leave its source intact and delete the sources for the
> 3-stack. Is this the right way to do it? The only difference from reality is
> that I'm not including booms and transmission lines for any of the antennas.
>  
> If it appears to the experts that I'm doing things correctly, can anyone
> give me a good example of two antennas that should show some interaction
> when placed in close proximity? I'd like to model a case of known
> interaction so I can make sure I'm doing this right.
>  
> 73, Dick WC1M