The shortfall of any software is that it is bound to the assumptions or rules
of the programmer(s). Yes, we have some wonderful antenna modeling software,
but as Jerry states, the only real data is empirical. It's got to be measured.
Good software (not all of it is) is a great place to start, but they the model
has to be built, tested, tweaked and tested again. It can be an endless cycle.
We've seen the same thing to some extent when we model an HF antenna at VHF for
convenience. Not all the laws of the real world apply themselves to RF that is
removed from the desired frequency by an order of 10 or more.
So, don't shun good tools like NEC and MINNINEC, but understand the caveat that
not everything that looks like a duck is indeed a duck
Dave
WA3MKB
p.s. Besides, the greatest fun can be found in building and toying with
something new. I have learned a lot from even my dumbest ideas (proven to be so
when built and tested).
-------------- Original message --------------
From: "Dr. Gerald N. Johnson" <geraldj@storm.weather.net>
> On Thu, 2008-06-05 at 03:41 -0700, Jim WA9YSD wrote:
> > It is a wounder that these antenna companies has not made an HF yagi
> > with a folded dipole or a DB antennas as the driven element. They all
> > ways come up with some thing strange. Must be too expensive. The DB
> > and FD has the best coupling factors that equals 1 or 0.9. A dipole is 0.1
> > .
> > Now one would conclude that if you had wide spaced elements on a yagi
> > that the DB and FD would be used. You think?
>
> Some VHF and UHF beams have used a folded dipole driven element. Many
> use a T match or some other scheme. The choice is generally based on the
> loading of the driven element by the directors and reflector. The best
> condition is when the folded dipole has a feed Z of 200 ohms so a 4:1
> coaxial balun can be used. That only works when the director/reflector
> coupling gives a dipole center Z of 50 ohms. In most computer optimized
> yagis, the dipole Z is lower or much lower than 50 ohms and the 4:1
> transformation doesn't result in a convenient Z at the folded dipole,
> like 50 ohms or 200 ohms. There the T, Gamma, Betta, or some form of
> stub match is often used to step up the impedance to something
> conveniently matched to a 50 ohm feed line.
>
> Then for a multiband beam, I don't know of a folded multiband driven
> element while traps work on a dipole.
>
> There is no such difference in coupling between a folded or straight
> dipole. And no good mechanical way to make a double bazooka to be
> supported only in the middle. Besides the double bazooka is a sham
> antenna anyway. It doesn't do anything for radiation or reception with a
> decent receiver.
> >
> > Strip the braid off of some RG8X and slide the center piece down the inside
> > of a driven element of you favorite yagi and tune the driven element and
> > see
> what happens.
> >
> That's a feed that works, but isn't necessarily weatherproof or broad
> banded.
>
> > So many people have computers these days, and antenna programs, that
> > they just do not go out side and experiment. The Wheel has already
> > been invented folks. It just needs a new groove to make it work better.
>
> Sometimes the new wheel has been found to be faulty. I know I've used
> EZNEC to invent a new antenna, but found the computation to be
> inconsistent and the freshly built antenna to work far poorer on the
> antenna range than the most optimistic computation. These programs
> handle bends poorly, and bends that aren't at right angles worse than
> that. The computation results depend the number of segments per piece of
> a bent or straight element.
>
> On segmentation, the originators of the NEC and MININEC programs
> recommend starting with a low number of segments and note the results,
> then increase the number of segments keeping track of the results, and
> hope for the results to plateau. The expectation is that the results are
> best along that plateau (gain or f/b vs number of segments). But I have
> found that the plateau isn't always asymptotic to a great result as the
> number of segments are increased further that the result wanders off
> from that plateau. Yes, EZNEC and ELNEC compute minimum and maximum
> segment numbers, but they don't take into account complications of the
> antenna structure for all possible cases. Using NEC and MININEC is more
> off an art form than an absolute computation. The results computed
> aren't for sure good until the antenna has been built and properly
> measured and even then there can be a dB or two of uncertainty.
>
> Programs like YAGIOPT are based on simpler computations than NEC and
> work only fair for single band yagis. Their computation is based on a
> formula in Kraus (W8JK) original antenna text book that is only a good
> approximation for PARALLEL elements very close to HALF WAVE RESONANT.
> The program and the formula will give an answer for an interlaced two
> band beam, but it won't work that way on the antenna range because that
> answer is WRONG because the second band elements aren't near to half
> wave resonant on the band being computed. But the simpler computations
> are fast because they are simple.
>
> There are other fundamental zits in the NEC family, but I'll leave those
> to other times because I appear to be alone in the world calling them
> errors.
> >
> > Keep The Faith, Jim K9TF/WA9YSD
> >
> 73, Jerry, K0CQ
>
> _______________________________________________
> TenTec mailing list
> TenTec@contesting.com
> http://lists.contesting.com/mailman/listinfo/tentec
_______________________________________________
TenTec mailing list
TenTec@contesting.com
http://lists.contesting.com/mailman/listinfo/tentec
|