Re: NEC-LIST: end caps on wires

Re: End caps on wires

NEC-4 includes an approximation of wire end caps. This is needed,
since the electric field is set to zero at match points along the wire
axis. It is not correct to say that the field is zero there unless
the wire is closed at the ends and source. Actually the field would
be very nearly zero on the axis at points more than a few radii from
the ends of a small diameter tube, but end caps help the numerical
solution near the ends when the segment length to diameter ratio is
small. Otherwise you get rapid oscillations of the current. That
happens anyway, but the end cap approximation delays the problem to
smaller length to radius ratio.

In NEC-2 the boundary condition is matched at points along the wire
surface, so this condition is correct whether the ends are closed or
open. You need to use the Extended Thin-Wire Kernel (EK) to
approximate matching the field on the surface due to a current on the
surface. NEC-2, like NEC-4, sets the current at the wire end to a
slightly non-zero value to allow for the current that would flow onto
an end cap (or back on the inside of the open tube). However, the
field due to current on the end cap is not computed.

To see the effect of end caps, you can model a thick wire as a cage of
thin wires. Something like 12 wires is probably enough, and you can
use symmetry (GR). The radius of each wire should be the radius of
the thick wire divided by the number of wires, so that the "equal area
rule" for a wire mesh is satisfied. The end cap can then be
represented by a disk of radial wires connecting the outer cage wires
to the axis. If you taper down to very small segment lengths near the
end and on the end cap the solution will approximate the singularity
in charge at the edge.

There probably will not be much difference between a wire with open or
closed ends. Brown and Woodward ("Experimentally Determined Impedance
Characteristics of Cylindrical Antennas", Proc. I.R.E., April, 1945,
pp. 257-262) were unable to detect any significant difference even
on very thick wires.

Jerry Burke
LLNL
Received on Wed Mar 04 1998 - 09:31:44 EST