[TenTec] Bazooka antenna.. More than you wanted to know!

[Dr. Gerald N. Johnson]

On Mon, 2008-06-02 at 09:16 +0100, Steve Hunt wrote:
> Folks,
> 
> At the risk of stirring up a hornets' nest ......
> 
> .... I continue to see claims that the Double Bazooka exhibits a wider 
> VSWR bandwidth than the equivalent thickness half-wave dipole. But I 
> thought that W2DU had showed conclusively that the "reactance 
> cancellation" mechanism claimed for the DB was a fallacy,  and that AI1H 
> (ex-W1DTY) had showed that the real explanation for any bandwidth 
> increase was losses in the coaxial elements. In other words, the same 
> effect as putting a resistor across the feed point!

I too have made that analysis, but I figure the slightly increased
bandwidth is mostly due to the use of twinlead for the end sections
giving a fatter conductor and so greater bandwidth. The two stubs that
are wired in series have a slight reactance compensation that can give a
very small bandwidth improvement.  That computation and associated
experiments are why I call the antenna sham.

> 
> Am I missing something? Is there subsequent work that has proved W2DU 
> and AI1H wrong? Or do folks just like using unnecessarily-complex, lossy 
> antennas :)

P. T. Barnum said it well.
> 
> I have an interest in the topic because I just did some related 
> experimental and modeling work on using coaxial elements to "shrink" 
> the size of a HexBeam. Yes, you get the expected "velocity factor" size 
> reduction, but you also get unacceptable losses introduce by the coaxial 
> stubs. A HexBeam driver constructed of RG58 would exhibit coax losses of 
> about 13dB. If you're interested you can read about it at:
> 
> http://www.karinya.net/g3txq/coax_antennas/
> 
> I know my test configuration was different from the DB, and that DB 
> losses will not be of the same order, but the message is the same: 
> "quarter-wave inductive coaxial stubs are lossy, low-Q, elements."
> 
> Now I think I'll turn off the computer for a few days until the hostile 
> reaction dies down :)
> 
> 73,
> Steve G3TXQ
> 
> 
Most of us look at a dipole's feed impedance in series reactance 
(R + jX) form. And there both R and X change as we sweep through the
dipole's resonance. That change is slower when the outer end of the
antenna is fat, like a conical dipole. But if we look at the feed point
conductance as parallel (G + jB) and plot the changes in conductance its
apparent that the conductance component is virtually constant over a 
20% bandwidth, but the parallel reactive component only varies. And it
varies in the opposite direct from a parallel resonant tuned circuit. So
a parallel resonant tuned circuit in parallel with the feed point will
compensate for the reactance change giving a nearly flat SWR over the
chosen bandwidth.

I have found and built (more than one) antenna for 80 meters with such a
lumped (mostly) parallel tuned circuit. My design takes a dipole
resonant at 3750 or 3800 KHz and puts a tune circuit at the feed point.
I use a 3 meter length of RG-58 (RG-8 would have less loss and the
current is high) shorted at the end away from the feed point resonated
by a 2000 pf capacitor. That capacitor needs a high current rating. The
one that WB0BQV has used for decades is a 5" diameter by 5" long
transmitting mica capacitor. That's required for a KW power level. His
antenna has a SWR of about 1.3 from 3.5 to 4.1 MHz (used to be active in
Army MARS).

A purely coaxial stub also works if the characteristic impedance is low
enough at the antenna. I've found a 25 ohm stub (measured the effects at
VHF) significantly widens the impedance match of a dipole (like 20%
bandwidth for a dipole made of 20 gauge buss wire) wired in parallel
with the feed. In that one, I mounted the stubs alongside the feed line,
no balun. You can also achieve the same result on my 80 meter antenna by
connecting two stubs, one 3 meters long (shorted away from the feed
point) accompanied by an open stub to make the two stubs a quarter wave
(in coax) long. They can be coiled up. Just in my VHF experiment I was
using a foot of RG-8 for each stub and that doesn't coil well.

After "inventing" all these things I found a two volume tome in the
library called, "Very High Frequency Techniques" that reported on
research accomplished at the Radio Research Laboratory of Harvard
University during WW2. They used the paralleled stubs and series stubs
to match a dipole for aircraft communications over the full 100 to 150
MHz range. I invented nothing new. Those volumes was published about the
time I learned to read in 1947 and didn't know about antennas.

So the double bazooka has little benefit and its accompanied by a
structurally weak antenna that has a lot of surface area in the coax to
and the twinlead to collect wind and ice with little metal inside to
provide tensile strength. It will fail with broken conductors unless
kept INSIDE. It will have a short life.

And then its a single band antenna (as is my 80meter broad band wire)
where you can't force it to any other bands with a tuner.

However, if the two stubs are rewired at the feed point so they are in
parallel with each other and the coax feed rather in series you achieve
my experiment with the effective 25 ohm stub that DOES have
significantly increased bandwidth. Its still wimpy in wind and ice but
until it breaks it will have that increased SWR bandwidth.

The original double bazooka has a 1/2 wave (in coax) length shorted at
the ends. At the feed point, the braid is cut and the two braids are the
feed point. This is the bad design. If both the braid and center
conductors are cut, and the three braids are connected together (two
stubs plus the feed line) and the three center conductors are connected
together, then this gives the 25 ohm characteristic stub impedance.

If I wanted to work ONLY 80 meters (other than ten meters all the HF
bands are narrow enough that a plain dipole has adequate bandwidth)
without a tuner, I would build a dipole from 120 feet of #12 copper wire
in the conventional manner and on my coaxial feed, I could use a couple
quarter wave stubs tied to the coax which would give me the mechanical
strength of the wire (and I might use #10 copperweld for the top,
immensely stronger than #12 copper, left over from the 80 meter double
extended zepp I used to have up that was made of #10 copperweld, some
326 feet long) and the broadbandedness of the stub. Or I might just go
with the 3 meter stub and the 2000 pf transmitting mica capacitor.

73, Jerry, K0CQ