[TowerTalk] 4 Square phase line length question

[Jim Lux]

----- Original Message -----
From: "daniel hearn" <dhearn at ix.netcom.com>
To: "Michael Tope" <W4EF at dellroy.com>; "Tom Rauch" <w8ji at contesting.com>;
"Jim Lux" <jimlux at earthlink.net>; "Dan Levin" <dan at andlev.com>;
<towertalk at contesting.com>
Sent: Sunday, May 09, 2004 3:31 PM
Subject: Re: [TowerTalk] 4 Square phase line length question


> I use my Iwatsu SS5711 4 ch. 100mhz scope for tuning the 4 sqr antennas I
> have built. I have 4 equal length  coaxes carrying the signal from current
> xfmrs at the tower bases to the scope amp inputs. All 4 current xfmrs are
> first placed on the same tower base to assure that the 4 channels are
> identical in phase and amplitude at the CRT screen. I found that they were
> extremely closely matched.
>   My first 4 sqr used 4 Hytowers with the stubs for 20 meters and up
> removed. I used a homemade hybrid coupler for the 90 deg shifted signal
and
> 1/2 wave of coax for the 180 deg. on 80m. Two of the towers were relay
> selected for broadside operation on 40m in one direction and another pair
> for the other direction.
>   The current phase and amplitude values on 80 were less than the desired
> values and really suffered when going from SSB to CW. After several years
of
> use, the 40m option was disabled and the stingers were strapped to the top
> of the tower. The stingers were lengthened by extending the 2 inch portion
> with sleeved tubing. This greatly improved the bandwidth of the system and
> it was used that way for a while even though the currents were not as
close
> as desired in amplitude and phase.
>   Ultimately I abandoned the hybrid coupler and built a W7EL type LC 90
deg
> phase shifter. This was easily tunable to the correct phase shift but the
> amplitude was not right. A tapped auto xfmr allowed the amplitude to be
> tweaked to the desired value. This put all of the amplitudes and phases at
> the desired value. Unfortunately, this solution is quite frequency
> sensitive. I tuned it for the CW band and had plans to relay select a
> similar circuit for SSB but left Dallas to move to Spokane before that
> happened. The towers were sold at the time of the move That system worked
39
> zones from Dallas and the other (z23) was picked up from Spokane.  73,
Dan,


I think that this points up what you really need to make this work...
something with enough degrees of freedom to adjust both phase and
amplitude - which the W7EL network + autotransformer will do.  I think that
any kind of quad hybrid scheme is going to be doomed in the long run,
because virtually all quad hybrids (particularly ones at HF with lumped
components) are very sensitive to the termination impedances.. you can
either fiddle with networks to get the impedances write (using a fixed
hybrid) or use a network which is adjustable.

For the 4 square, 90 degree excitation case, a baseline configuration could
be a pair of pi or T networks, which would allow you to set the phase shift
and power division ratio.  A 3 element network can independently adjust the
impedances at each port and the phase shift: The phase shift is either 0 or
90 (or two values differing by this amount, which is easier), one port is
simply the impedance at the coax phasing line (i.e. the transformed antenna
feed point impedance), and you set the other port impedance to get the right
power division ratio and impedance seen by the transmitter.   There are lots
of circuit topologies possible, and some may be more "convenient" than
others.  (http://home.earthlink.net/~w6rmk/antenna/phased/design.htm and
http://home.earthlink.net/~w6rmk/antenna/phased/networks.htm discusses some
network synthesis approaches)

The broadcast industry has been doing this for decades, and there are
network synthesis/adjustment procedures in several antenna handbooks
(although I haven't seen any on the web recently).  There's probably some
moderately systematic process that one could go through using a scope to
fiddle with the knobs on a standard 3 element tuner.  A bit of
sophistication should be able to do it with a computer controlling suitable
autotuners.

I think, without having an exhaustive proof, that one can do it with 4
adjustable components (at least for the single band case).

One could use something like inexpensive MFJ C.L.C tuners and a piece of
paper to record the settings, after one has dialed in the phases with a
scope.  A version of this approach was used for a 2 element array, in one of
the ARRL antenna compendia, including the use of a phase reversing switch to
help dial in the phase (adjust for null on the desired signal, flip the
phase of one element).

Given that most people don't put up their antennas on a nice homogeneous
ground plane with perfectly identical antennas, etc., some sort of
measurement capability is going to be needed to properly adjust the array.
Unfortunately, this isn't like a multielement beam antenna on top of a big
tower, where mechanical tolerances can be held tight enough that you can
create a "cookbook" design that works out of the box without measurement.