On 2/7/18 7:57 AM, Roger Parsons via TowerTalk wrote:
I had been expecting a discussion here on this recent QST article, but there
has been very little. So I thought I would jump in. Answering my own question,
I do not feel that the article does present the whole truth.
It seems to me that there are two self evident cases where an object placed
close to an antenna does not cause loss:
(1) Where the object is perfectly conducting, it may change the radiation
pattern, but as it has no resistance there can be no losses;
(2) Where the object is perfectly insulating, it may affect the characteristics
of the antenna (by changing the dielectric) but as it can pass no current,
there can be no losses.
In all other cases a loss may occur, and I have no reason to doubt the general
methodology described in the article.
However. The NEC based analysis is based on an antenna and a broadly resonant
tree in free space. A tree in free space is considerably less likely than an
antenna being there! (Actually, as there is currently an expensive motor car in
orbit perhaps I am wrong...) The analytical simulation considers an infinitely
long tree next to an antenna, again in free space.
Perhaps a right circular cylinder is an accurate representation of some
particular tree, but it doesn't seem to fit the generalised case. Trees are
ground mounted and have a ground system which probably has higher conductivity
than their trunks and foliage - and which actively seeks out water. They also
have top loading of almost infinite variety. The cedar tree that I can see from
my window has very complex and spread out branches and foliage, whereas a palm
tree (which I can unfortunately not see) appears to be quite close to a
monopole with a some top loading.
That would be a logical next step for modeling.
I think that the model of "single cylindrical lossy tree" next to
"vertical" is a good place to start - for a lot of these problems, you
want to bound the range of values before making it more complex.
From first principles, the effect of a lossy cylinder perpendicular to
a dipole is going to be small, as long as diameter of cylinder is <<
length of dipole. There is also the "lossy cylinder next to the
feedpoint" issue, which could be answered with a series of model runs
moving the tree closer and farther from the antenna.
One would look at the currents induced in the tree next to the vertical,
if they're sufficiently small (< 20dB?) then the resonance or not of the
tree (topload, earth load, etc.) probably isn't a big deal.
Since the tree diameter is << wavelength, treating it as a uniform
cylinder is probably valid, rather than modeling the bark, cambium, and
heartwood separately (which would be quite the ordeal in NEC.. rings of
wire meshes with carefully chosen properties) I'd go with a FDTD type
model for that kind of thing.
Because a tree is lossy it will have a very broad resonance, but it seems to be
stretching credibiity to suggest that a 5m high tree would significantly
influence a 1.8MHz vertical. Or that a 50m high tree would have significant
coupling to a 28MHz vertical. In each case the tree is likely to be very far
from resonance.
I'd venture that the tree isn't "resonant" in that the losses are
probably high enough that the current in a particular segment of the
trunk is driven mostly by the coupled fields from the adjacent segment
of vertical. And, if it's lossy (i.e. the currents are small), it's not
going to radiate much, so the far field pattern isn't going to be
affected a whole lot.
Easy to test, though, run the model with different lengths of tree
and/or antenna.
I could go on, but my feeling is that although the conclusions reached in the
article are reasonable for the model adopted, they are likely to greatly
overstate losses in the real world.
73 Roger
VE3ZI
ps Perhaps there has been discussion on another reflector?
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|