Owen Duffy wrote in
:

"Peter" wrote in
:
....

My current 5/8 wave ground plan project is simply to get something on
air, however I plans to construct an improved version with the tapped
coil approach.

That lets you shorten it for a bit more gain, and a good match.

If you cut the vertical for 0.6 wavelengths, you should think of
starting with an inductor with reactance towards 1000 ohms.


I meant to elaborate on this a bit more. (Did I hear someone groan?)

If for example, the feedpoint Z of a 0.6 wave vertical over four quarter
wave radials was 150-j500, your tapped coil matching network can be
designed using bulk standard circuit theory to transform 150-j500 to 50
+j0, and nowhere do you use the missing 54° in those calcs. That might
suggest that the "missing degrees" are some kind of explanatory crutch
(or ham speak) that is not directly related to solving the problem.

Owen

On Sun, 14 Nov 2010 02:25:50 GMT, Owen Duffy wrote:

I meant to elaborate on this a bit more. (Did I hear someone groan?)

If for example, the feedpoint Z of a 0.6 wave vertical over four quarter
wave radials...

I'll bite (or groan as the expectation demands) - why "quarter wave"
radials? A rule of thumb?

* * * Rhetorical questions follow * * * *

Quarter wave in physical length?

Quarter wave in electrical length?

Elevated Quarter wave radials?

If elevated, Quarter wave drooped radials?

* * * Philosophical ponderings follow * * * *

The discussion of radials usually attends ground mounted studies in
the 100M band in the 1930s. Those studies sought to reduce loss while
mimicking a conductive ground of infinite extent. Radiators taller
than Quarter wave were treated to feedpoint loading (such as found in
the current topic, albeit with the possibility of it being elevated
and thus muddying the philosophy here). All such historical (and
current AM band engineering) feedpoint loading presumed, basically, a
non-resonant ground system. As Quarter wave long radials imply
resonance (at least in the first read), this would suggest that,
perhaps, this "tuning" should be further examined in light of
feedpoint loading. The conclusion, to my mind, would be that
significant reduction in feedpoint loading could be accomplished by
tailoring radial length (much less drooping that is already part of
the lore).

At first blush, it would seem that the radials would be shorter than
Quarter wave (forgive me for not first confirming this astonishing
leap of faith).

Of course, there is every chance some reactance will remain to be
"tuned" away (returning us once again to loading) - if the mismatch is
deemed significant.

If such is the case, and returning to the original design, what
problem is the Quarter wave length radial rule of thumb responding to?

* * * * Alternative analysis * * * *

Or to put it into the light of other antenna topological discussions,
and in this regard the off-center fed dipole. Here we have an
off-center feed (we rarely go on to describe all such installations as
"vertical dipoles"). We can fully expect that, as such, we are
transforming the expected 70±j0 Ohms into some other value. Quite
frequently in an OCF design, it is much higher - and variable by the
degree of offset. However, for a fixed frequency, this is better
understood and can be anticipated. The proximity to ground and the
geometry (the radials certainly disturb the shape of an OCF dipole,
even if vertical) further change things, but conceptually the monopole
with resonant radials still constitutes an OCF design that is "on
center fed" for the vertical element when it, too, is a Quarter wave
in length.

For many prospective feed points along the length of the OCF dipole,
the only consideration needed is for a ratio transformation, not
tuning. This is usually resolved in a BalUn. Hence "loading" is
removed from the picture through careful consideration of the whole
antenna, the degree of offset, and not through arbitrary assignment of
Quarter wave length radials to all vertical designs.

* * * * Conclusion * * * * *

The concept of a loading coil where its length of wire "replaces" the
missing length of radiator wire is a commonplace for technologists. It
serves the discussion quite well at that level.

The value of this length of wire's inductance is going to vary by
significant value for the many coil form variables available to the
technician. Hence the exactness of this "replacement" is questionable
on the face of it at the engineering level of discussion. This
equivalence "replacement" is forced further into unresolved exactness
if we move the same coil up into the radiator (without changing the
radiator's length).

The same could be said with the treatment of Quarter wave length
radials, which, after all, are a special and not general solution.

73's
Richard Clark, KB7QHC

Richard Clark wrote in
:

On Sun, 14 Nov 2010 02:25:50 GMT, Owen Duffy wrote:

I meant to elaborate on this a bit more. (Did I hear someone groan?)

If for example, the feedpoint Z of a 0.6 wave vertical over four
quarter wave radials...

I'll bite (or groan as the expectation demands) - why "quarter wave"
radials? A rule of thumb?

It is just what I modelled, so I was declaring the context. The thread
started on 2m, my discussion was in that context, and the usual
application would be elevated radials, I modelled free space.

I used a quarter of the free space wave length. It is not that important
because as you note, matching the feedpoint impedance deals with the
length issue. The reason I didn't specifiy any slope is that they were
horizontal.

Other configurations are possible, but the numbers will vary.

I suggest that as the vertical length approaches a half wave, a set of
shorter radials, and perhaps three might well provide adequate
decoupling... but Z will differ again.

I was not trying to publish a working design, rather to give some info
on the way these things behave.

Owen

....

On Nov 13, 8:25*pm, Owen Duffy wrote:
If for example, the feedpoint Z of a 0.6 wave vertical over four quarter
wave radials was 150-j500, your tapped coil matching network can be
designed using bulk standard circuit theory to transform 150-j500 to 50
+j0, and nowhere do you use the missing 54° in those calcs.

That's because the lumped-circuit model assumes that all signals
travel instantly at faster than light speeds through the coil. At
instant, faster than light speeds, the coil cannot possibly occupy any
degrees of the antenna. When the real-world speed of light limit is
taken into account by using the distributed network model, the degrees
occupied by the coil falls out as part of those real-world
calculations. The lumped-circuit model is simply flawed for the
purpose of trying to determine the degrees occupied by the coil. I am
finishing up an article on this subject. At 3.5 MHz, the velocity
factor of the 100 turn, 10 inch long coil is 0.04, which makes the
coil occupy 26.4 degrees when used for a 3.5 MHz mobile antenna.

The "Axial Propagation Factor" from the Hamwaves Inductance Calculator
at:

http://hamwaves.com/antennas/inductance.html

can be used to determine the number of degrees occupied by a loading
coil. For the above coil at 3.5 MHz, the axial propagation factor is
1.8118 radians/meter. Multiplying by 1.4554 converts it to degrees/
inch. The coil is 10 inches long so: 1.8118(1.4554)(10) = 26.4 degrees
occupied by that loading coil at 3.5 MHz.

We can model a transmission line as lossless, but none exists in
reality.

We can model a loading coil that occupies zero degrees of the antenna,
but none exists in reality.
--
73, Cecil, w5dxp.com