The problem of modeling linear loads in NEC (-2 or -4) is multi-faceted.
First, both versions of NEC have difficulties with angular junctions
involving materials with different diameters. Note that tapered diameter
correction features work only for linear elements, and then only for a
limited frequecy span each side of resonance. The angular junctions are
ruled out.
The second problems is close-spaced wires of different diameters: NEC
reacts badly to these, ordinarily giving results that inflate the gain and
deflate the source impedance. If one has the patience, one can check the
validity of the model by summing all of the points in the sphere (or
hemisphere) which should average to the value for an isotropic source.
Most implementations of NEC do not make this check convenient.
Linear loads at the center of an element, when equispaced from the main
element, act as shorted transmission line stubs. However, when not
equispaced and/or when placed further out on the element, these loads do
not have equal currents on each parallel wire, and thus depart from stub
performance, Hence, they cannot be easily modeled either as loads or as
transmission lines--that is, the results cannot be automatically accounted
as accurate to the particular antenna being modeled.
Linear loads can be modeled in MININEC, since it does not suffer either of
the problems noted. However, for the corners, MININEC requires either the
use of very short segments (that often carry the model beyond segmentation
limits) or of length tapering toward to corners. Free space models using
the symmetry feature of AO might well be possible if the length tapering
features is implemented. I do not know how well the "New MININEC" of
Rockway and Logan would handle this problem, since I do not have a
functional copy. NEC4WIN from Orion has reprogramed the core for Windows
and removed the segment limitation from its MININEC 3.13, but I have not
tried this problem in that program.
Physical modeling of linear loads is possible, but the extra work of
finding a valid model and a satisfactory program persists.
It is possible to model the general principle by modeling the elements as
continuous elements without loads, using heavy segmentation densities, and
deriving the Leeson monotaper equivalents. Then using the same diameter
material, one can create linear loads of parallel shorted stubs and bring
the array to desired characteristics (max gain, max F-B, etc.). However,
these substitute models are not necessary replicas of some particular
design, but only indicators of the sort of performance possible with
linear loading. They are also useful for studying the currents on linear
loading lines in various configurations (straight down, folded back
toward center, folded outward toward ends, etc.).
Hope these notes are useful. NEC-4 limitations are documented in a piece
I did for QEX some time back. NEC-2 limitations are generally more severe
for the same phenomena.
-73-
LB, W4RNL
L. B. Cebik, W4RNL /\ /\ * / / / (Off)(423) 974-7215
1434 High Mesa Drive / \/ \/\ ----/\--- (Hm) (423) 938-6335
Knoxville, Tennessee /\ \ \ \ / / || / (FAX)(423) 974-3509
37938-4443 USA / \ \ \ \ || cebik@utk.edu
URL: http://web.utk.edu/~cebik/radio.html
--
FAQ on WWW: http://www.contesting.com/towertalkfaq.html
Submissions: towertalk@contesting.com
Administrative requests: towertalk-REQUEST@contesting.com
Problems: owner-towertalk@contesting.com
Search: http://www.contesting.com/km9p/search.htm
|