[TowerTalk] 1/8 wave spaced 80m verticals

[Jim Lux]

At 08:43 AM 9/10/2003 -0700, Richard Karlquist wrote:
> > Can anyone help me with the construction of a lewallen phasing
> > network for 2
> > 1/8 wave spaced 80m verticals or sort out my christman feed. I
> > have used 2 x
> > mfj1792 top loaded verticals. I have 40 radials 1/8 wave long on each
> > antenna. I have tried the christman feed to these antennas but i
> > suspect the
> > phasing is not quite right as front to back is only about 15db. I used the
> >
>
>You see a lot of books talking about characterizing the
>antennas and then designing a network for them.  I have
>found that is is much easier to just use an adjustable
>phasing network and tune it until the currents are correct.


I heartily agree.... there are so many "unknowns" in a real system that you 
can go batty trying to model them all and account for them.  I figure you 
use the model to bound the problem (about what range of values will you 
need), and then go and do the experiment for real.  ("Analysis paralysis" 
is a very real danger with phased arrays.. why go out and stand in the hot 
sun trying to figure things out when you can sit in the cool 
airconditioning in front of your computer...)


>The adjustable phasing network is simply an adjustable
>inductor and adjustable capacitor in an "L" network.
>If you don't have these components, you can use an
>ensemble of known fixed inductors and capacitors by cut and try.
>
>First, check the feed coax to make sure it is exactly an
>electrical quarter wave.  In that case, you don't have to
>measure the antenna current, you only have to measure the
>voltage at the other end of the coax, as it will be proportional
>to current.

Excellent point... saves having to fool with current probes. High Z scope 
probe beats the current probe any day.


>You want the feed voltages equal in magnitude and 135 degrees
>out of phase.  You can measure each voltage vs ground with
>an RF voltmeter and adjust them to be equal magnitude.  Then connect
>the RF voltmeter *across* the feed voltages and adjust for
>a voltage 1.85 times as much as the individual voltages.
>Iterate back and forth between magnitude and phase until
>both are correct.  This "3 voltmeter method" is described
>in more detail at http://www.n6rk.com/driving_the_7_hex.pdf
>pages 4 and 5.  I made my own RF voltmeter with a
>Schottky diode, bypass capacitor and pocket voltmeter.
>I put about 100 mW into the antenna for these tests.

The three voltmeter method can even be extended to measure mutual Z in a 
multiport network, much like a network analyzer (it becomes the 6 or 
3*Nports voltmeter technique....), but the math gets pretty hairy, and the 
calibration even more so. (For those technically inclined, look up "6-port 
network measurement"... there's a bunch of stuff from NIST on the web... 
essentially, with a suitable set of scalar measurements (i.e. RF voltage) 
and appropriate cal standards, you can measure the complete characteristics 
of an arbitrary 2 port network)

Jim W6RMK