Topband: Open Wire Lines - Fact and Fiction

[Tom Rauch]

I'm not sure what post Eric was referencing, because he didn't really quote
which post he was referring to.
My original post was 22 October 1998. I did make an error in a second post
on Aug 8, 2003 where I said:

"I had a 1500-foot open wire line of number 8 solid copper spaced just over
2
inches, and I impedance matched the line at both ends through broadband
transformers. I measured a power loss of 2% on 4 MHz, including the
transformers."

That should have been about 4% loss on 2 MHz, not 2% on 4MHz. We were
talking about 160, not 80 meters. Also since we are looking at this closely,
my line was 1400 feet long. Not 1500.

> In conclusion, I do not think an open wire (2) line with a loss of
> .005
> db/100' is possible in the HF ham bands. I would say that a long line like
> this would exhibit at least 1 db. total loss, from one 50ohm port to the
> other.

Not even close! My data was based on what I observed through measurements of
a real transmission line, and that is the final proof of any estimate.

I'm also puzzled by this:

<<Second, the impedance of the line calculates to 412 ohms, which requires
the
noted transformers to get it down to 50 ohms. Broadband transformers, as
stated in the data, would have to be constructed using ferrite cores. It is
not
possible to build such transformers, especially at transmitting power
levels,
which do not have losses that range from small but noticeable, to
overwhelming.
If
you doubt that, just spend about 10 minutes talking to anyone who has ever
designed a broadband PA. The idea of building these matching transformers
with
power losses in the 1% range is just not thinkable. It would be possible to
build transformers using low-loss coax, which would not be fully broadband,
but
would have losses that could be kept down to a few tenths of a db. over a
single band.>>

That is a disappointing comparison. Transformers for matching two ohms or
less to 50 ohms are considerably more difficult to construct with very low
loss than a 50 to 450 ohm transformer. Even so, I can do that in most HF PA
cases when I am not working with space and cost restraints!!

I can and have built 300kHz to 100MHz transformers for 450/50 ohm
applications with under .1dB loss over that entire range, and with an SWR
under 1.1:1 over that range. Transformer performance at higher Z from and
into non-reactive sources and terminations is nothing like the performance
in very low impedance applications.  This is especially true if efficiency
is intentionally sacrificed for cost and space, and the source impedance is
very low and very reactive (as it is in most SS designs)!!!!

While Eric greatly overestimated feedline radiation, even Gary's more
optimistic estimate failed to consider line layout:

"The field intensity from a loop is proportional to the area enclosed
by the loop -- in dB, that's 20 log(Area1/Area2).  Applied to the above
analogy, radiation from a 2-inch by 1500 foot loop is -67 dB relative to a
750 by 750 foot loop. -67 dB corresponds to a loss of 0.0000009 dB
(e.g. 0.2 mW out of 1 kW). That's not much radiation."

If conductors of the feedline are transposed at small fractions of a
wavelength along the line (perhaps by "twisting" the line), radiation from
each section cancels at distances large compared to the "twist distance".

My line was twisted every 20-40 feet, since it was primarily going to be a
160 and 80 meter line.  Even the -67dB estimate is too high. On ten meters I
could hear "signals" listening to the dummy load terminated line. But then
the "twists" are a wavelength apart, and certainly no longer cancel farfield
signal response of the line! In that case I'm sure Gary's calculations are
correct.

The conductor resistance, according to my calculations, accounts for ALL of
the loss I measured. This makes sense to me, since there was no dielectric
to speak of. It also is a very well known effect that conductor resistance,
below upper UHF, accounts for the bulk of transmission line loss. At HF we
can pretty much ignore dielectric losses.

My measurements, made with good measurement protocol, were within reasonable
error of what theory predicts. 1400 foot of #8 copper in a 450 ohm line has
about the same loss as 100 feet of good quality 50 ohm RG-8 size coaxial
line on 2 MHz. It is far under 1dB, that is certain. If properly constructed
it also has negligible radiation. BE SURE to twist or transpose long
balanced lines at small fractional wavelength intervals!! This would include
long two-wire reversible Beverage antennas, especially with wide wire
spacing.

73 Tom