M3KXZ'S TRI-BAND VERTICAL
MOXON

Editor's Note: Fresh off
his success with his beach mounted vertical Moxon, Pete Millis,
M3KXZ, started thinking about multi-banding. The MoxonProject
website has several multiband Moxon designs each of which is unique
and approaches the subject from differing perspectives. Pete's
design follows in this tradition. Shown below are a series
of emails and MessageBoard postings sent between Pete, Allen Baker
KG4JJH, and your humble webmaster.
4/28/2006:
Hi Steve.
Just a quick update for you.
I have now come up with a design for a tri-band
"Moxon" that doesn't suffer from the effects of interaction
between elements. Cebik shows a design for the "simplest 17-12
Moxon" (<http://www.cebik.com/moxon/1712m.html>), but
this does not have the close end coupling between driven element
and 12m reflector. He also discusses the problem of making a Moxon
work on consecutive bands and suggest skipping a band to minimise
interaction between elements.
The design I have come up with uses a "non-resonant
on the ham bands" driven element (an EDZ). I want to operate
on 20, 17 and 15 on one antenna. The EDZ is therefore only about
67% the length required for the full size 20m driven element. This
has enabled me to shorten the entire vertical dimension as I have
reduced the folded back section on the driven element and lengthened
the folded section on the 20m reflector.
I have incorporated reflectors for 17m and
15m inside the perimeter of 20m antenna. But instead of doing what
Cebik and others have done (i.e have the straight reflector section
and then ends folded forward - which in the case of Cebik's design
mean that the ends of this reflector are no longer close couple
to the ends of the driven element), I have done something different.
Based on the idea that close end coupling is very important, I have
put the ends close together, but then set the element up as a "V".
This sets a good distance between centre of driven element and centre
of reflector.
It's hard to explain in words, so I've attached
the .EZ file for you to look at. Have a look at the patterns if
you get a chance. The modeled F/B ratio and gain is superb. (DOWNLOAD
EZNEC FILE HERE)
  
Another enhancement has been to move the
feedpoint closer to the ground to make setting up easier. I will
be connecting my 4:1 balun (the W1CG design) and tuner here. Shifting
the feedpoint like this has very little effect on performance
of the antenna, but makes the practicalities more straightforward.
I know in some ways this is a more complicated
set up (i.e. it requires a tuner) but the performance should be
far batter than a fully nested Moxon.
Interested to know what you think. I am
working on getting this built this weekend and tested during the
week. My modeling has been checked by another ham and all seems
to be good.
Kind regards
Pete Millis, M3KXZ
4/29/2006:
Hi Steve and Allen.
High SWR's don't so much affect the radiating efficiency of an antenna, but
they do cause increased transmission line losses when fed with coax. For
this application, I am using a highly efficient 4:1 balun (the design by
Charles Greene W1CG) at, or very close to, the feedpoint, and using my
Elecraft T1 atu to tune close to 1:1 - then coax to the rig. This way, the
losses are minimised.
Martin Gillen (VA3SIE) is also working on a vertical "Moxon" for 15 and 10.
His initial design had a straight reflector for 10m, but we exchanged a
couple of "you show me yours and I'll show you mine" emails, and he has now
incorporated the "V" reflector as it greatly improves the F/B ratio.
Martin has been doing some very interesting stuff using a feedline
optimisation program, and has calculated twinline feeder lengths that bring
the SWR right down to below 2:1. This is something that I will probably
adopt in the final build.
Forgot to mention - ignore the way the currents face in the model - these
only show the magnitude of the current, and not the direction the antenna
is firing in. This works in the usual way with driven element and
reflectors.
5/4/2006 (via MoxonMessageBoard):
I have made a modification to my vertical
beach Moxons - the feedpoint has now been moved down to just
over a metre above the bottom corner of the driven element.
This has been done purely from a practical point of view, and was
necessary in helping me with development of my tri-bander (20/17/15).
A requirement for getting good performance
from the tri-bander was to reduce the amount of wire within
the perimeter of the Moxon. On L.B. Cebiks site there is a
description of a simple dual band Moxon that has 2 nested
reflectors and one "non-resonant on the ham bands" driven
element. My tri-bander uses this sort of driven element but
has better coupling between driven element and reflectors.
The reflector elements are all tuned for
resonance on each of the ham bands. The driven element requires
matching to the rig with a balun and atu. Due to the high
SWR levels on such driven elements, it is important to keep
the feed line between antenna and tuner as short as possible.
By moving the feedpoint so it's not far above the ground, I can
now mount the balun and the atu right there, thereby minimising
transmissions line (TL) losses. The SWR on the TL between
rig and tuner is close to 1:1 so even the use of very lightweight
coax becomes acceptable for portable operation.
The EZNEC modeling of the antenna shows
that very little difference in performance can be attributed
to feedpoint location. Moving the feedpoint to a low point
on vertical Moxon's is a useful proposition for anyone who
wants an easier way to route the feedline away from the antenna.
For a Moxon mounted above the ground, you could have the feedpoint
at the bottom corner and drop the feedline down vertically, thereby
getting rid of the problem of routing the feedline perpendicularly
away from the centre of the antenna.
5/5/2006:
Hi Pete,
I took a look at your 20m Beach Moxon on
the website. Very interesting and a great idea! I wish that I lived
near the ocean to try your new Moxon application. I tried increasing
the antenna height to see the effect. It appears that you have optimized
the pattern and gain for a very low height above ground (salt water).
I have been very busy at work lately and
haven't had much time to look at your tri-band version. I would
be interested to learn how the testing is progressing. As Steve
points out, the SWR does appear to be high, especially on 20m and
less so on 15m. You are correct in that a good antenna tuner should
be able to make this work. (I took a first look at the W1CG balun
and made a note to try one.) However, I would be surprised if an
autotuner will tune the 20m band. Most are limited to 10:1 SWR.
OK, I just reread your first email about
using optimized twin-line lengths to get the SWR down to 2:1. Very
good, sounds like you have covered all the bases.
Keep us posted.
73,Allen, KG4JJH
===
Hi Allen.
Thanks for your email.
I built 2 of the tri-band Moxons over the
past couple of days. The first built from quite heavy wire (1.3mm
core + 0.85mm insulation) doesn't pack quite as small as I wanted
for portable. So the second I built yesterday is made form computer
ribbon cable wire (.32mm core and .45mm insulation). I remodeled
the new wire in EZNEC and found it necessary to very slightly shorten
the tuning tails on 20m reflector and the 15m reflector.
As noted, the SWR is high on 20 and 15m,
but the Elecraft T1 had no problem tuning to below 1.5:1 via the
W1CG balun. I still haven't gone through the process of optimising
feedline lengths as I felt that, provided the balun and tuner could
handle it, the shorter the feedline between antenna and tuner the
better. As such, I have just 50cm of feedline from antenna feedpoint
to balun, and 20cm of coax from balun to tuner. This way, even though
the SWR is high on the coax, the calculated losses really are minimal.
For construction, I used small offcuts of
plastic damp proof coarse (the stuff used over here to stop damp
creeping up walls when houses are built). This is really strong
plastic sheet, which is easy to cut and holes can be made in it
with a soldering iron. These are used at the points where the "<"
shape wires connect to the "[" shape wire, and do a good
job of holding things where they should be. Also, being plastic
sheet it is also very light and easy to stash away.
The Moxon was just suspended at 4 corners
using lengths of shock cord, and the fishing poles were angled away
from each other on their ground stakes to enable sufficient tension
to be applied across the top. There was no need for guying which
is handy.
Putting the feedpoint low down has really
helped with the practical aspects of it all, and it really has next
to no impact on performance or radiation pattern.
In the early part of next week, I'll be
setting up on the beach down here (lovely weather at the moment,
which I hope continues to Monday!) and will be sure to sort out
some pictures.
As soon as I have some more pictures and
test results, you and Steve will be first to know.
BTW, Mr Cebik has showed an interest as
well, so it may even make an appearance on his site at some point
(never know!).
Cheers for now, and keep busy
Pete Millis
|