[TowerTalk] 75 ohm - v - 50 ohm coax

[Jim Brown]

On 2/18/2013 5:47 AM, Stan Stockton wrote:
> Cut the coax to a common multiple of 1/2 wave length for as many bands as you can.  You can easily get 20-10 with the length you have without losing too much coax and maybe 40-10.  It won't help the loss but will make your readings look the same on both ends without any further matching.

Yes. I use a lot of low loss 75 ohm coax in my station.  Much of it is 
carefully cut lengths of 1/2-inch CATV hard line that are some multiple 
of a half wavelength -- 3 wavelengths feeding a 15M monobander, 2 
wavelengths on a 20M monobander, three wavelengths on a 10M monobander.  
I'm also using Belden 8213 to feed high fan dipoles (up about 110 ft) 
for 80 and 40M.

In an earlier post, I noted that if you make an NEC model of a simple 
half-wave dipole, you can export the impedance curve to SimSmith and add 
various lengths of feedline to it and study the result.  And as I noted 
in that earlier post, these resonant lengths of line can (and usually 
do) have the effect of broadening the SWR bandwidth of the load 
presented to the transmitter.

Also, as Stan has observed, you can feed a tri-bander with a length of 
75 ohm coax that is 1 or two wavelengths on 20M.  The 1-wavelength line 
will be 1 1/2 wave on 15M and 2 wavelengths on 10M.  The 2-wavelength 
line will be 3 wavelengths on 15M and 4 wavelengths on 10M. If your 
feedline needs to be longer, simply add 50 ohm coax to it on either 
end.  If the resonant lengths of 75 ohm coax are a bit long, coil up the 
excess.

I've taken this a major step further by using a Vector Network Analyzer 
to measure the complex antenna impedance over the band, exporting that 
data to SimSmith, and designing matching stubs to improve the matching.  
Adding a suitable stub at the antenna end of the cable can match the 
load presented to the 75 ohm line, reducing the loss in that line, AND 
presenting a matched load to the transmitter.  It's usually not 
practical to use this technique for antennas that cover more than one 
band, but it's both easy and effective for mono-band antennas.

SimSmith is free and very easy to use, but you must understand 
transmission line concepts to be able to use it.  Google to find it. The 
author has written some excellent tutorials, and there's a rather 
different tutorial on my website that goes through what I've described 
here in a lot more detail.  In addition to allowing us to design 
matching networks, SimSmith also computes and displays all the losses 
between the transmitter and the antenna, so we can study the usefulness 
of our designs.

http://audiosystemsgroup.com/publish.htm

Another point with respect to loss in that coax.  Because its diameter 
is quite large, loss is MUCH lower than the RG8-size coax we usually 
use.  It is also lower because the loss in ANY coax at HF is I-squared R 
loss in the copper, and the current in 75 ohm coax is proportionally 
less for the same power level.  Further, the ADDED loss due to this 
relatively small mismatch is quite small, even for rather long cables.  
So, since the matched loss is much lower to begin with, the total loss 
with the mismatch is usually less than with even the best RG8 and 
RG213-size coax.

And yet another observation. The loss in coax depends on the impedance 
on the ANTENNA end of the coax, NOT the match at the transmitter.  High 
dipoles are a closer match to 75 ohms than to 50 ohms, while low dipoles 
are a closer match to 50 ohms. The reason is the mutual impedance into 
the antenna of the reflection from the earth.  You can see families of 
curves showing this for various antenna heights in ON4UN's book, and in 
the ARRL Antenna Book. I suspect that there rae similar graphs in an 
RSGB publication. That's why I'm feeding my high 80/40 fan dipoles with 
75 ohm coax -- I get a closer match between the antenna and the feedline 
over a greater bandwidth (this matters on 80/75M more in the US, because 
the band is 500 kHz wide).

73, Jim K9YC