On Thu, 12 Oct 2006 23:30:17 -0400,
(Fred McKenzie)
wrote:
So far I've learned that the majority of loss is in the matching network.
Actual gain of a shortened dipole is only about a half dB down, but
radiation resistance is so low that I-squared-R losses in the tuner are
significant.
You are right that the gain of a short practical dipole is good (and
that requires that the losses are low).
The transmission line is a potential source of high loss when feeding
a short dipole.
There was an article on QST some time back on using a 66' dipole on
all bands. I have written a review which contains some graphs that
cast light on where the losses are, the article is at
http://www.vk1od.net/LOLL/index.htm .
It turns out that for most practical configurations, it is difficult
to achieve good overall efficiency when the dipole length is below
about 35% of a wavelength. You will probably need parallel wire
transmission line rather coax for a shortened dipole.
Someone mentioned that you could reduce losses in the tuner by adding
external loading coils. It seems to me that you would just be moving the
loss out of the tuner into the loading coils.
The question of whether an order of magnitude of improvement can be
attained appears to be answered. If you define that as ten dB of
improvement, then probably not. The total penalty is most likely less
than that.
Taking an "order of magnitude" to mean a tenfold increase or decrease
(10dB), it will be challenging to improve an antenna tenfold by just
adding length unless it was very inefficient in the first place (too
short, lossy feedline + ATU).
If you coax centre fed a 40' dipole on 3.6MHz with 30m of RG58, you
would find opportunity for tenfold improvement by increasing length
(to a particular value).... but only because it was so hopelessly
inefficient as configured.
By my 35% suggestion, a typical parallel line centre fed 40' dipole
will peform ok down to about 8.6MHz. If it is coax fed, it will only
be reasonably efficient on its series resonances (~12MHz, ~36MHz).
Owen
--