[TowerTalk] Reverse Fed Towers

[n0tt1 at juno.com]

On Tue, 15 Dec 2015 17:45:39 +0000 wosborne44 at gmail.com writes:
> I have a tower that has a base that is in concrete and grounded.  I 
> would like to make it a vertical without installing insulators.   
> Has 
> anyone used elevated radials with reverse feeding, i.e., connecting 
> the 
> center conductor to the radials and the shield to the grounded 
> tower?  

Hi William,

Yes, I did that this Summer/Fall, for a 7mhz antenna.

Details, details.....
I have a 70-ft BX tower and I installed
12 radials (taking Rudy, N6LF's advice) that are
insulated from the tower in groups of
4 aluminum wires.  

BTW, it's interesting to note that our favorite 
frequency standard, WWV, has 9 radials
on their vertical dipole antennas.

The radials attach to a "buss" that
is insulated from the tower legs with a short, 
ceramic "strain" insulator.  The insulator 
is clamped to the tower via a small bracket 
and two SS hose clamps.  Each "buss" is
connected to a center insulator via 3 wires 
stripped from RG8 type coax and wrapped 
in tape to prolong their life.  Each radial terminates
in SS hardware on the "buss" to provide a mechanical 
connection.  Each radial electrically connects to the "buss" 
via a homebrew wire clamp and a short copper jumper 
with AL/CU terminals.  

Connection to the tower is at 3 places via a 3/16" thick
aluminum plate, triangular in shape, with 3 "arms" that are
made adjustable in their length.  At the ends of the arms
I have a block of aluminum with slots cut into them that 
clamps onto the heavy sheet metal legs of the tower with
a couple of SS screws.  The top of the plate holds the 
center "beehive" type of insulator and the bottom 
holds a PVC electrical box.  Inside the box I have a 
1:1 balun/choke wound on a FT240-61 (gasp!)  toroid.

The output of the choke connects to the beehive insulator
and the supporting aluminum plate.  The "leads" are reversed
so that the radials are at DC ground via the shield of the
coax matching section that feeds the input of the balun.  
The shield is grounded at the base of the tower.  The
matching section is the usual 1/4-wave length of two 75-ohm
RG-11 coax in parallel.

Up on the tower, with a antenna analyzer connected to
the input of the balun, I found, as expected,
no noticable change in input impedance when I shorted the
coax shield to a tower leg.  (Remember, the shield connects
to the radials.)  So, there's no issue with running the coax
feed over to a tower leg (inside the tower) and then down to
the base.

About radial and tower lengths....I did start to model the
antenna using EZNEC 6.0+.  It was, for me, a waste of time
because of all the angles of the tower, the 3 jumpers that
feed the radial from the center of the tower, etc etc.  A complicated
mess!  So, I started with a vertical tower length of 234/f and a little
longer 
for the radials...something like 34 feet each.  I expected 
that the vertical radiator length would be shortened 
by about 6-10%.  It turned out that it was much more than that.  
I ended up about 7 ft shorter than the 234/f length.  I ended 
up with the radials being about 33ft each.  

I measured the current in each wire on the 
"output" of the choke/balun and found that they were 
virtually equal at resonance.  Per Rudy's white paper,
(Google it) it's important to balance the currents in this antenna.

The instrument I used was a homebrew current meter.  
I powered up the antenna with my MFJ-269 antenna analyzer placed
right at the input of the choke/balun and the current meter 
probe was clamped onto each output lead of the choke/balun....
obviously not at the same time.

I attached some photos for you.

Well, that's about it except to say, that I'm pretty sure that
the radials could be attached directly to the tower legs and then
feed the vertical section fed with a gamma match.  That's not
been tried here, but it's on my list of things to do for another
vertical antenna.

73,
Charlie, N0TT