On 5/3/20 3:49 PM, Wes wrote:
There's some misinformation here. Rather than reinventing the wheel, I
will suggest reading the EZNEC help section: "Stepped Diameter
Correction Overview".
Wes N7WS
On 5/3/2020 12:19 PM, jimlux wrote:
Excellent point - tapered means "conical section" (i.e. smoothly
varying radius), and most antennas are actually stepped.
In all the NEC engines, there's no problem connecting a "wire" with
radius X to a wire with radius Y, especially if they're in a straight
line, and the segment length isn't less than the diameter.
\
I stand corrected on the ability of NEC2 to handle the steps..
from that manual:
"NEC-2 is known to be inaccurate in modeling connected wires having
different diameters. (Note: This is sometimes called being "tapered". It
shouldn't be confused with EZNEC's segment length tapering.) The problem
gets worse as the segments near the junctions get shorter and as the
diameter difference gets larger."
..
It goes on to describe the well known Leeson method for creating a
constant diameter element that models the stepped diameter.
..
" The accuracy of NEC-2 is worst in the presence of stepped
diameters when the segment length/diameter ratio is small. "
...
"The inaccuracy of NEC-2 in the presence of large steps in diameter
typically shows up as an incorrect reactance"
---
this is all correct - and the reactance is important because the
relative phasing of the elements is what gives the antenna decent
side/backlobe performance
I guess the problem I have is that this is pretty fuzzy in terms of
"what is large or small" in this context, and what the source of the
modeling error is.
NEC-4 is substantially improved in terms of the "basis functions" for
the current in a segment, so it handles short segments better, and steps
better. Although there are things it can't do - I need to recover my
password so I can go read
"NEC-4.1: Limitations of Importance to Hams" in QEX
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|