NEC-LIST: Model-Based Parameter Estimation

Duncan, Juergen,

Thank you for your turn-around-response.

I am aware of Ed's papers on the subject of model-based parameter
estimation, and work is still going on in this area. I see that Ed is
chairing a session at the ACES'99 Symposium on the subject. At the
moment all I can do is a step-by-step analysis.

When I was writing my paper entitled "Revisiting wire aerials at the
turn of the century", Proc. of 4th International Conference on
Antennas and EM Theory, Xi'an, China, August 19-22, 1997, pp.73-78,
Jerry Burke did an analysis for me using NEC-4 that showed fine
structure detail (impedance vs. frequency), "fine" wrt to the
frequency step he used; for comparison with my step-by-step analysis
using MININEC. MININEC agreed almost exactly with NEC-4, excepting in
places where impedance changed rapidly with frequency.

The model-based parameter estimation program used by Jerry certainly
isolated the frequency-band(s) of interest, if ones interest is
interpretation of this fine structure detail. Whether a step-by-step
analysis would have brought NEC-4 into agreement with MININEC is yet
to be determined --- if I do look at this detail.

The oscillatory change of impedance vs frequency, confined to a narrow
range of frequencies, is *very* large for a 3-wire fan, smaller for a
5-wire fan, very small for a 12 -wire fan, and more less invisable for
the 54-wire fan.

It was for this reason I wanted a small frequency step for the 54-wire
fan.

MININEC agreed almost exactly (for a 3-wire fan) with the
impedance-frequency plot determined experimentally, I repeat - almost
exactly. The agreement was not so good as the number of wires
increase --- but I must admit that we goofed (partly) --- since to
support many wires we used a horizontal insulated rod. We realized
this mistake, and for the 3-wire fan we removed the rod --- which
brought experiment into agreement with theory.

I started out my study with the fan wires open, since the
published-and-republished photograph of Marconi's fan antenna shows
clearly 12-wires, with a blob at the end of each wire. But we have
now concluded that this is not a real photograph. It is a heavily
retouched photograph of Marconi's 400-wire-cone antenna, which
collapsed before he could use it. The artist painted out 18- of the
20-masts, and drew in the 12-wire fan. According to history (the
written word) Marconi's antenna was a 54 wire fan, and he certainly
did not have 54-insulators to connect the wires to the triatic.

Rossi Giuseppe and Kok Chen have expressed interest in my files. My
files are EZNEC-Files, which I could send to anyone interested (by
e-mail) --- but, the antenna models are pretty simple: the source is
on a short wire feeding the multi-wire fan. I have models for 3-, 5-,
7-, 12-, 24- and 54-wire fans with a no-sag "catinerary" wire. I also
have a 24-wire fan with a catenary, having a reasonable sag.

73, Jack

PS

I forgot one important point for anyone who models multi-wire fans ---
where the source on a single wire feeding many wires. The treatment
of the charge distribution at a junction is a problem with NEC-2; but
NEC-4 and MININEC should do it right . The condition used to
determine the distribution of charge at the junction for NEC-2
considers only the radius of each wire. NEC-4 uses a quasistatic
approximation taking account of neighboring wires, so should do it
better [Burke, personal communications, 1996].

_____________________________________________
John S. (Jack) Belrose, PhD Cantab, VE2CV
Senior Radioscientist
Radio Sciences Branch
Communications Research Centre
PO Box 11490 Stn. H
OTTAWA ON K2H 8S2
CANADA
TEL 613-998-2779
FAX 613-998-4077
e-mail <john.belrose_at_crc.ca>
_____________________________________________
Received on Thu Feb 04 1999 - 18:07:42 EST