NEC-LIST: LPDA NEC-2 feed problem

Re: NEC-LIST LPDA NEC-2 feed problem

It does not look like the source model is the primary problem with your
LP. I would excite it with a source on the end segment, or maybe put
three segments on the end and the source on the middle one. However,
that gave a result very close to yours from 14 to 44 MHz. You need to
compute average gain in a situation like this to have a reference of how
well the model is doing. For your original model excited on the center
of three segments on the end of the transmission line the NEC4D results
were:
Frequency Gain Average gain Corrected gain
  20 MHz 3.36 -1.24 4.60
  22 3.53 -1.29 4.82
  30 4.06 -1.42 5.48
  40 3.43 -1.40 4.83

Running it with two and four times as many segments on each wire
produced a somewhat worse result. For 4x resolution:

  22 2.92 -1.95 4.87

This indicates a probable problem with thick wires and short segments.
With three segments on each wire of the short feed transmission line
the center of the first segment on the line is about 2.8 times the
element radius from the axis of the element. While this does not
trigger warnings in NEC-4, it is close enough to degrade accuracy.
Changing the feed transmission line to one segment gave better results:

Frequency Gain Average gain Corrected gain
  20 3.88 -0.73 4.61
  22 4.05 -0.77 4.82
  30 4.57 -0.91 5.48
  40 3.92 -0.91 4.83

The best result seemed to be with no feed transmission line, and the
source on a single segment across the gap of the first element with
the same radius as the element (your case 3). That yielded:

Frequency Gain Average gain Corrected gain
  20 4.28 -0.32 4.60
  22 4.44 -0.38 4.82
  30 5.00 -0.49 5.49
  40 4.50 -0.33 4.83

I don't know what you could do with NEC-2 on this problem, but it would
certainly be necessary to compute average gain and correct the maximum
gains. With the short-circuit termination the transmission line has a
large standing wave at low frequencies, and that may contribute to the
modeling problems. I don't know of any limitation on the length of a
source segment as long as it has equal length segments on each side.
With NEC-2 it probably should be run in double precision, while with
NEC-4 single precision may be OK.

Jerry Burke
LLNL
Received on Tue Jul 15 1997 - 11:02:03 EDT