Pete and all,
Here is what I have developed for K9AY Loop modeling... Remember, although
the antenna was "discovered" by modeling, all the hard data on appropriate
terminating resistance and pattern shape were obtained through extensive
listening tests on the first unit. I have not found a model that does not
need to be "tweaked" to match the observations.
Method #1

I model the loop exactly as it is constructed, connected to "Real" ground
using the SommerfieldNorton "High Accuracy" ground type. This is NOT a
recommended procedure for NEC, because the currents in/out of ground can be
unpredictable. However, using the segmentation on the list below and the
"Average" ground conductivity, I get modeled results that match the observed
performance.
Because this model goes against NEC modeling guidlines, I rarely publicize
it. It suits my purposes, within strict limits based on LOTS of experience
with this particular antenna type. I believe it is an acceptable model 
just as with Beverages and EWEs when they are modeled with connections to
High Accuracy Ground.
(Dimensions in feet)
Wire 1 0,0,0 0,15,4 #14 9 segments
Wire 2 0,15,4 0,0,25 #14 13 segments
Wire 3 0,0,25 0,15,4 #14 13 segments
Wire 4 0,15,4 0,0,0 #14 9 segments
Ground  "REAL" "HIGH ACCURACY" "AVERAGE" (.005, 13)
Source: Wire 1, 0%
Load: Wire 4, 100%, 400 ohms
Freq: 1.83 MHz
Method #2

Instead of the two long wires used by W7EL, a model using "High Accuracy"
ground can be created using (8) radials 30 to 50 feet, less than one foot
above ground. Length doesn't matter too much, as long as the radials enclose
an area about twice the footprint of the antenna. Be sure the radial field
include two that are exactly in line with the loop.
This model requires a terminating resistance much higher than what was
empirically determined, but it "follows the rules" and the pattern shape is
in the right ballpark. I never use this model  it was developed as a
"first cut" at making a model that met NEC usage guidelines. It is possible
that this is an accurate model for a loop installed over radials, but that's
not how I make them.
Method #3

I have not yet thoroughly tested this modeling method, but it gives
consistent results and appears to match realworld behavior.
1. Use the REAL/MININEC ground
2. Raise loop 1 foot above ground
3. Add a wire from ground to the junction of Wires 1 and 4 (which were
connected to Ground in the above wire list)
4. Feed/terminate the loop at the same points as above  inner ends (first
pulse) of the semihorizontal bottom wires.
5. Add a 150 ohm resistive load in the new wire between ground and the loop
feed/term junction.
MININEC ground models antenna impedance as if connected to perfect ground. I
believe the resistor in the vertical wire represents the reduced currents
in/out of lossy ground. 150 ohms is a "trialanderror" value, and the right
value probably varies with local ground conditions. In any case, the pattern
shape is right, dBi gain and feedpoint impedance are pretty close to the
High Accuracy Ground model I've been using.
73, Gary
K9AY
_______________________________________________
Antennaware mailing list
Antennaware@contesting.com
http://lists.contesting.com/mailman/listinfo/antennaware
