[NLRS] [BC'ers] Fwd: Re: Anybody have experience with multi-band feeds?

[Dr. Gerald N. Johnson]

As the parasitic element gets close to the driven element, they tend to 
act as one.

Modeling with NEC2 fails with small spacings like would be a parallel 
wire transmission line. It fails to show the effects of close spacing on 
the coupling, much like the log formula for parallel line impedance 
fails for impedances under about 275 ohms because it fails to account 
for non uniform charge distribution around the wire that comes from the 
electric field being higher where the surfaces are closer. The invcosh 
formula takes that field difference into account. Unfortunately ARRL HQ 
insists you can't make parallel wire line with lower than 87 ohms Z 
because the simple log formula shows that impedance when the center to 
center spacing is one wire diameter, e.g. the wire are touching.

MININEC does better for small spacings, but it has a region closer 
spaced the NEC2 where it fails too. Probably from the same effect, its 
known, Cebik did an article comparing them in QEX, May June 1988 
beginning on page 3 (to 16). And the programmers of MININEC have an 
article in the same issue beginning on page 17. These are not available 
from ARRL archives so I have scans at:
http://www.geraldj.networkiowa.com/papers/nec/Cebik4s.pdf and
http://www.geraldj.networkiowa.com/papers/nec/Mininecs.pdf
because my paper for Aurora on closely spaced stacked yagis will spend
considerable time comparing NEC2 and MININEC performance for closely
spaced wires.

Part of the modeling difficulty is that the Method of Moments treats a 
wire as having infinitesimal diameter when computing fields, it only 
takes wire diameter into effect when computing resistance and reactance. 
And as was shown back in 1938 that wire proximity affects the charge 
distribution and so the capacitance between the wires neglecting that 
makes the coupling lower than it should be.

Back to the closely spaced antennas, a W8JK antenna, two radiators fed 
out of phase so that there can be considerable circulating currents 
(which increases the radiation) between them the gain increases as 
spacing gets shorter until the resistive losses of the radiators and the 
connecting feeder take over to limit the increase in circulating current 
and when really close the radiation cancels because its getting to be 
close to 180 degrees out of phase at all distances and angles. But 
that's not a parasitic array.

The charts that show gain vs yagi boom length tend to start at nearly 
one wavelength boom length and even at 1 wavelength there is a tendency 
to turn down more rapidly than the general line of the curve. There is a 
tendency to rise more slowly for very long yagis as well showing there 
is an optimum range for that near doubling of gain with doubling length. 
K1FO and DL6WU designs for long yagis are not intended to work for short 
yagis. But I'll have an example at Aurora.

The patch antenna works on the edge current of the patch and replaces a 
gaggle of linear radiators arranged along a pole by radiators at the 
edges of the patch so while the length in the prime direction vector 
gets very small, the area perpendicular to that is much larger than that 
of a yagi. E.g. the wind load is higher from the planar patch.

One good reference I'm reading is Kraus "Antennas and Propagation," 4th 
edition printed only for southern Asia. It has a portion on propagation 
added for that part of the world to Kraus "Antennas for all 
applications" 3rd Edition for the rest of the world. It does't mince 
mathematics. But much can be skipped. Its available for purchase on line 
at quite reasonable prices shipped from India. And delivery was 
surprisingly quick. Its about 900 pages so its not a one evening or one 
month read.

73, Jerry, K0CQ



On 3/21/2011 7:04 AM, Marciniak, Ed wrote:
>
>
> What's the limit as spacing tends to zero on a dipole with a reflector (or ground plane)? With one director and one reflector, what is the limit?
>
> What is the gain of a patch antenna less than the thickness of an aluminum rod?
>
> 6db at zero length isn't that irrational, though a bit optimistic?
>
>