NEC-LIST: LPDA NEC-2 feed problem

I've been having trouble modeling a log periodic antenna with NEC-2. The
main problem seems to be in applying excitation to the model. In the model
which follows, feed needs to be applied across the transmission line at the
short-element end of the antenna. Experiments have shown that NEC-2 begins
giving inaccurate results when the source is placed on a wire shorter than
about 0.02 wavelength. In this model, the transmission line conductor
spacing is approximately 0.002 wavelength, making violation of the
empirical rule necessary. The problem also seems aggravated by NEC-2's
problem with connected dissimilar diameter conductors; in the model, the
transmission line is #12 wire and the elements 0.5" diameter. (For those
who have EZNEC, it's example file LOGPER.EZ, furnished with the program.)

By making small changes in the feed structure, large changes in the
reported gain result. Curiously, the pattern doesn't change but just gets
smaller. The only factor which can cause this in practice is loss, yet
there is no loss mechanism (load or wire loss) in the model.

The following model extends the transmission line a short distance and has
two sources, one placed in each conductor of the extension. This gives
reasonably good results with NEC-4, particularly if the diameter of the
extension wires is increased to the 0.5" diameter of the elements. However,
it makes the results worse when modeling with NEC-2. Here are some results
with various feed structures:

Gain in dBi for NEC-4D/NEC-2S -- see notes below for structure description.
  1. 3.47/2.55
  2. 4.13/3.89
  3. 4.34/4.16
  4. 2.76/3.15
  5. 2.80/2.03
  6. 3.61/2.72

1. As in the model below: #12 wire transmission line extended, 2 sources.
2. Same as 1 but extension 0.5" diameter.
3. Single 0.5" dia wire across trans. line, 0.5" dia, 1 segment, 1 source.
4. Same as 3 but 3 segments.
5. Same as 4 but #12 wire.
6. Same as 3 but #12 wire.

So by making relatively minor changes in the feed structure, the gain can
be made to vary from 2.03 to 4.16 dBi with NEC-2, and 2.76 to 4.34 with
double-precision NEC-4.

I also modeled this antenna with MININEC, which is much more tolerant of
source placement. For all the cases above, the gain varied only from 4.875
dBi (case 2) to 4.9 dBi (case 1). In the absence of better information, I
trust this result more than the NEC-2 or -4 results.

Does anyone know of a method of modeling the feed to this antenna which
will result in a reasonably accurate gain with NEC-2? It's also fairly
important to maintain a reasonably accurate indication of the feedpoint
impedance. A log periodic antenna using transmission line models in place
of the #12 wires in this example gives somewhat better results. However, it
may not accurately represent an antenna physically like the one modeled
here, so I'd like to stick with this basic antenna model.

Any light anyone could shed on this would be greatly appreciated. Thanks!

Roy Lewallen

CM 17-10m Log Per - ARRL Ant Book
CE
GW 1,8,0.,-4.1518,-.0304,0.,0.,-.0304,.0159
GW 2,8,0.,0.,.03048,0.,4.15188,.03048,.0159
GW 3,8,.99644,-3.3213,.03048,.99644,0.,.03048,.0127
GW 4,8,.99644,0.,-.0304,.99644,3.3213,-.0304,.0127
GW 5,8,1.79375,-2.6573,-.0304,1.79375,0.,-.0304,.00953
GW 6,8,1.79375,0.,.03048,1.79375,2.65735,.03048,.00953
GW 7,8,2.43129,-2.1257,.03048,2.43129,0.,.03048,.00794
GW 8,8,2.43129,0.,-.0304,2.43129,2.12573,-.0304,.00794
GW 9,8,2.94157,-1.7005,-.0304,2.94157,0.,-.0304,.00635
GW 10,8,2.94157,0.,.03048,2.94157,1.70053,.03048,.00635
GW 11,3,0.,0.,-.0304,.99644,0.,-.0304,.00103
GW 12,3,.99644,0.,-.0304,1.79375,0.,-.0304,.00103
GW 13,3,1.79375,0.,-.0304,2.43129,0.,-.0304,.00103
GW 14,3,2.43129,0.,-.0304,2.94157,0.,-.0304,.00103
GW 15,3,0.,0.,.03048,.99644,0.,.03048,.00103
GW 16,3,.99644,0.,.03048,1.79375,0.,.03048,.00103
GW 17,3,1.79375,0.,.03048,2.43129,0.,.03048,.00103
GW 18,3,2.43129,0.,.03048,2.94157,0.,.03048,.00103
GW 19,1,3.048,0.,-.0304,3.048,0.,.03048,.00103
GW 20,1,0.,0.,-.0304,-.1524,0.,-.0304,.00103
GW 21,1,0.,0.,.03048,-.1524,0.,.03048,.00103
GW 22,1,-.1524,0.,-.0304,-.1524,0.,.03048,.00103
GW 23,3,2.94157,0.,-.0304,3.048,0.,-.0304,.00103
GW 24,3,2.94157,0.,.03048,3.048,0.,.03048,.00103
GE 0
FR 0,1,0,0,21.
GN -1
EX 0,23,2,0,1.414214,0.
EX 0,24,2,0,-1.414214,0.
RP 0,1,361,1000,90.,0.,0.,1.,0.
EN
Received on Mon Jul 14 1997 - 10:31:42 EDT