NEC-LIST: NEC4D and buried antennas

The ground treatment in NEC-4 is the same as that in NEC-3 and the
insulated wire treatment is the same as in NEC-3I, so they have been
around for a while. Solutions with these codes for buried wires have
been validated for a number of cases.

As George Hagn said, he has obtained experimental validation for a
buried insulated wire. We have also checked against an analytic
solution by J. Wait for the propagation constant on an infinite
insulated wire below the ground surface. As the wire approaches the
interface from below, the phase and attenuation constants undergo a
characteristic change, and NEC results were in very good agreement
with Wait's solution.

The impedance of monopoles on buried radial-wire screens have been
checked against the approximate solutions such as Wait and Pope (Appl.
Sci. Res. Sec. B, Vol. 4, pp. 177-195, 1954) and Collin and Zucker,
"Antenna Theory, Pt. II." Those results use the surface impedance of
the ground in parallel with the wire screen and a compensation-theorem
solution. The input impedance from the NEC solution was in good
agreement with these approximations for cases where the approximations
were expected to be valid.

A useful self-consistence check can be obtained by setting the ground
conductivity to zero and integrating the "average gain" around the
sphere above and in the ground. Average gain should equal 1.0 if
there are no ohmic losses, since total radiated power must equal input
power. We got 0.99 on one recent check for a thick monopole on 113
buried thin radials, which seems very good. However, there is a bug
in NEC-4 for reflection of the radiated field from a buried wire down
into a dielectric ground (theta > 90 deg., conductivity = 0). That
should not affect many practical calculations, but must be fixed
before this average gain check can be used. The fix will be included
in the July issue of the ACES Newsletter and is attached below.

There are also demonstrated cases where NEC is not accurate for
antennas near ground. Small loops near ground are to be avoided. As
a loop becomes electrically small the NEC impedance matrix becomes
increasingly ill-conditioned. This can be a problem without ground,
but with the limited accuracy of the table-lookup used for the field
due to ground, the error can easily overwhelm the solution. A loop
smaller than resonant is bad, and the minimum size will depend on the
distance of the loop from the ground. The problem is obvious when
input resistance goes negative, but I would not trust any results for
a small loop near ground.

Making accurate measurements with ground can be difficult, as George
indicated. Measuring ground parameters is difficult. If a wire is
buried without insulation the contact with the ground may be less than
perfect, especially if the wire has corroded. At the 1991 IEEE APS
Conference a paper was presented showing the measured attenuation and
phase along a wire above ground. It was claimed that the measured
results agreed with R.W.P. King's formulas, while NEC was in
disagreement. However, further investigation showed that if King's
result was evaluated accurately it was in agreement with NEC and with
a solution published by Wait. King's results, derived from
transmission line theory, are usually accurate for highly conducting
ground, but not for low sigma and epsilon, while Wait's solution has
no such restriction. Varying parameters in Wait's result and NEC
indicated that no choice of sigma and epsilon of the ground could give
quite as high an attenuation as had been measured. Maybe the
difference was due to ground conductivity varying with depth, which
cannot be modeled in NEC, or due to some other experimental factors.

We are always interested in hearing about any new validation or
anti-validation that people get.

Jerry Burke
LLNL

============================================================
Correction to subroutine ZYSURF for NEC4D.FOR
For NEC4S.For use REAL and AIMAG in place of DREAL and DIMAG
============================================================
C
C COMPUTE SURFACE IMPEDANCE AND ADMITTANCE OF GROUND
C
      CTHM=SQRT(1.-(ETAX/ETA)**2*(1.-CTH*CTH))
      IF(CTH.LT.0.)CTHM=-CTHM !add
      IF(ABS(DREAL(CTHM)).LT.ABS(DIMAG(CTHM))*1.E-6)THEN !add
         IF(CTH*DIMAG(CTHM).GE.0.)CTHM=-CTHM !add
      END IF !add
      YSTE=CTHM/ETAX
      ZSTM=CTHM*ETAX
Received on Thu Jun 18 1998 - 10:03:18 EDT