I responded directly to Dave, but it appears that a more general interest
in this feature of NEC programs exists, so will pass along a short note
here.
EZNEC and similar NEC-2 and NEC-4 implementations have a TL (transmission
line) input card. They permit the modeling of mathematical transmission
lines. That is, the lines will show the correct progression of impedance,
etc. transformation as lines or as capacitive/inductive stubs
(open/shorted), but are not accounted within NEC as physical elements of
the antenna structure. Therefore, mutual impedance and currents that
result in fields are not calculated for the transmission lines entries.
For calculations where transmission lines do not in the real world form a
significant part of the radiating structure, as with loading stubs, NEC
will produce correct results--that is, results which permit antenna
construction within the normal variables of terrain and ground clutter.
And performance will be correctly predicted, again, within those same
limits inherent to all modeling.
To check modeling results within the model structure, you can calculate
the stub load (capacitive or inductive reactance) and use a fixed
reactance load (or a component specified as L or C) and then replace it in
the same location with a transmission line stub. The results should be
close enough together to call the difference insignificant.
You may specify line or stub length either as a physical length, using the
transmission line Zo and velocity factor, or you may specify it as a
length in electrical degrees using the Zo alone.
I have run a large number of phased antennas, phased arrays, and stub
systems and have found the system to be completely accurate (although it
does use ideal/lossless lines). For example, in some notes I did on using
quarter wavelength 75-ohm matching section, sometimes with other cable
lengths between the antenna and the matching section, NEC results
tabulated exactly with independent calculations and Smith chart
calculations. Incidentally, if joining cable sections remotely from the
antenna, simple add a very short wire at any distance from the antenna
with only 1 segment. Connect the cable ends to this segment. A remote
segment of this sort can also be a feedpoint (Source) for the antenna,
giving an estimate of operating point impedance/SWR. Since the lines are
only mathematical, you can place this short segment anywhere in the
universe without affecting calculations. The transmission line length
itself will be as specified in the TL card/table.
Among the antenna types on which these lines have been successfully used
are loaded reflector verticals and horizontal parasitic arrays, ZL and
related antennas, collinear arrays, bi-directional wire beams, etc.
The transmission line system will not successfully model transmission
lines used as part of the actual radiator, such as bazookas. To model
these antennas, you need to create physical parallel wires having the same
characteristic impedance by virtue of wire diameter and spacing as the
transmission line used. If velocity factor plays a role in these
antennas, the model will not be exact, since the VF of such modeled lines
will by 1.0 instead of the actual line VF. In some antennas that have
been developed experimentally rather than from theory upward, it is not
always clear whether or not VF plays a role--hence, some incertainty about
the relationship of the model to reality.
I hope these notes are useful.
-73-
LB, W4RNL
--
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
|