Re: NEC-LIST: How do I simulate a CARDIOID Array?

Martin Berger wrote:
>I am experiencing problems in simulating a cardioid array at
>300MHz. The array consists of two vertical ½ lambda dipoles,
>stacked horizontally with a separation of ¼ lambda. The dipoles
>are fed with equal magnitude and 90 degrees phase difference.
>
>These are the requirements for a cardioid pattern as detailed in
>Antennas by J D Kraus 2nd ed. pp.452. Included on pp.452. are 30 polar
>plots for various dipole spacing and phase differences. They were all
>simulated and gave expected results (OK) or wrong results (X):
>
>
> Phase Difference
>
> Spacing 0 45 90 135 180
> ------------------------------
> 1/8 lambda OK X X X OK
> 1/4 lambda OK X X X OK
> 3/8 lambda OK X X X OK
> 1/2 lambda OK OK OK OK OK
> 5/8 lambda OK OK OK OK OK
> 1 lambda OK OK OK OK OK
> ------------------------------
>
>
>The NEC code is:
>
>CM
>CE
>GW 1 9 0 0 -.207 0 0 .207 .02
>GW 2 9 0 .25 -.207 0 .25 .207 .02
>GS 0 0 1
>GE 0
>EX 0 1 5 0 1 0
>EX 0 2 5 0 0 1
>FR 0 9 0 0 280 5
>RP 0 1 360 1000 90 0 1 1
>EN
>
>It would be very much appreciated if anyone can suggest what is going
>wrong here....

Martin --

1. Your main problem is that you excited your array with voltage
sources rather than current sources. The impedances seen by the two
sources differ in the "asymmetrical" cases wherein the phase
difference between the sources is not zero or 180 degrees. Therefore
the currents at the two feedpoints are not at all what you intended or
expected. I.e., their magnitudes are not equal and their phases do
not differ by 45, 90, or 135 degrees as you think they do.

To solve this problem, change the sources to *current* sources. Since
(I think) some or all versions of NEC allow you to specify multiple
voltage sources but not multiple current sources, I suggest inserting
a large resistive "load" in series with each source, in the center
segment of each dipole. A value of R = 1.E6 ought to do it. A more
elegant (and much less lossy!) solution is to put a quarter-wavelength
transmission line between each voltage source and its feedpoint. The
transmission line transforms the source impedance of the source from
zero to infinity.

2. Even with correct feedpoint currents you will find that the
radiation pattern differs slightly from what you expect, because the
distribution of current *along* one wire will not be exactly like the
distribution in the other wire. This occurs because of coupling
between the wires, and the effect is therefore greatest for closest
spacing. You can reduce substantially the magnitude of this effect by
making the wires thinner in their middles and thicker toward their
ends. This inductive center-loading and capacitive end-loading also
reduces the self-resonant wire length. BTW, I've never heard of this
solution from anyone else; it's something that I "invented" myself. I
strongly suspect that it's described somewhere in the literature, and
would like to know where. Anyone?!

3. Roy Lewallen wrote a nice article on the topic of feeding cardioid
and similar array antennas, which appeared in one of the ARRL Antenna
Compendium volumes a couple of years ago, IIRC. Perhaps someone can
provide a more precise citation.

73 de Chuck W1HIS
Received on Sat Mar 28 1998 - 12:18:55 EST