NEC-LIST: Back messages on EX5.

Sorry Malcolm et. al.,

Here are L.B. Cebik's last three messages, most recent
first. I believe I've already posted all of mine to the
list.

Sincerely,
Doug Miron

Doug,

The model that you sent is sensitive to changes in the
segment length of the
loading hats. As an experiment, I increased the number of
segments per
radial or spoke to 26 from the original 6. The source
resistance does not
change much, but the reactance is about +j18. The AGT is
now about 1.01.
The AGT will not change very fast in this type of model,
because the end
spoke loads are largely self-cancelling with respect to
fields and
interaction. But each spoke shows the appropriate current
curve (divided by
4 relative to the last center section segment).

None of this resolves your EX0 vs. EX5 question, but I
thought the model
properties might be of interest to you in your analysis.
Here is the
revised model that I used to look at the spoke currents--at
the increased
segmentation per spoke.
=======
CM End-loaded vertical in free space.
CE
GW 1 11 0 0 0 0 0 4 0.004
GW 2 26 0 0 0 20 0 0 0.008
GW 3 26 0 0 0 1.22515e-15 20 0 0.008
GW 4 26 0 0 0 -20 2.4503e-15 0 0.008
GW 5 26 0 0 0 -3.67545e-15 -20 0 0.008
GW 6 26 0 0 4 20 0 4 0.008
GW 7 26 0 0 4 1.22515e-15 20 4 0.008
GW 8 26 0 0 4 -20 2.4503e-15 4 0.008
GW 9 26 0 0 4 -3.67545e-15 -20 4 0.008
GE 0
EX 0 1 6 1 316 0 1
FR 0 1 0 0 2.9372 0.0
PT 0 1 1 11
XQ
EX 0 1 6 1 316 0 1
FR 0 1 0 0 2.9372 0.0
PT 0 6 1 26
XQ
RP 0 46 46 1002 0 0 2 2
EN
===========
I do find a very high correlation in this model among the
corrected AGT
source resistance values using either EX0 or EX5. My EX5
line is
EX 5 1 6 00 316 0
with the center element (Tag 1) reduced to 10 segments to
place the bicone
source at the very center of the antenna (the junction of
segment 5 with
segment 6).

NEC-2 AGT Reported R Corrected R X
EX5 .93467 1.37926 1.289153 21.0202
EX0 1.01105 1.27421 1.28829 18.8761

NEC-4 AGT Reported R Corrected R X
EX5 .931615 1.38464 1.2899514 24.1024
EX0 1.00765 1.27922 1.2889006 21.8010

This was done on GNEC 1.6.2 for reference, although we have
found that
different CPUs may shuffle values very slightly within the
same program
(applies also to EZNEC Pro/4). I also note without
explanation the 3-Ohm
shift in reactance between NEC-2 and NEC-4. More important,
I believe, is
the coincidence of corrected source resistance values
between EX0 and
EX5--within under 0.1% of each other within each core.

Hope something here is useful. If not, simply delete.

Regards,

LB

Doug,

I do not have an answer for either question--the departure
of NEC from
constant current calculations or the fact that the AGT
correction on
feedpoint resistance does becomes less accurate for
feedpoint impedances
containing higher reactive components. For longer antennas,
my suspicion is
that the conditions that yield the original inaccurate gain
report tend to
change the NEC-calculated operating conditions enough to
yield a different
resistance and reactance. Gain is less sensitive than the
source conditions
and perhaps the drift with deviant or non-ideal AGT scores
affects gain less
than it does the source impedance. While this condition
seems to be the
case for larger antenna structures, it might not be
applicable to small
constant current antenna structures.

Regards,

LB

Doug,

Actually, both results are the same in terms of a corrected
feedpoint
resistance. The average gain times the reported source
resistance = the
corrected source resistance--a fact that holds true only so
long as the
reactive component of the source impedance is low.v 5.47 *
.952 = 5.21
Ohms, and 6.15 * .848 = 5.21. Taking 10 log (AGT) gives a
value in dB that
tells us how high (>0 dB) or low (<0 dB) the reported gain
value is--again,
within reasonable valuesof AGT. Add or substract the amount
low or high to
the gain report for a corrected gan figure.

Regards,

LB
============================================================
=============
L. B. Cebik, W4RNL /\ * / / / tel:
865-938-6335
1434 High Mesa Drive /\ / \/\ ----)(---
http://www.cebik.com
Knoxville, TN 37938-4443 / / / \ / / || / e-mail:
cebik_at_cebik.com
USA / / / \ ||
W4RNL_at_ARRL.org

-- 
The NEC-List mailing list
NEC-List_at_robomod.net
http://www.robomod.net/mailman/listinfo/nec-list