[TowerTalk] Hex Beam

[Eric Gustafson]

Recently, there was a post regarding the Hex Beam reviewed in the
latest issue of CQ magazine.  I read the article and visited the
Hex Beam web site to gather as much information as I could on
this device.  I was intrigued enough to build an EZNEC model to
see how the claims stacked up against what might actually be
possible.  Please do _not_ read this the wrong way.  I did _not_
model the actual beam that is for sale because there is not
sufficient information about its construction detalis to generate
an accurate model of the device.  What I _did_ do was the
following:

1.  Started with a regular hexagon scaled so that a "W" shaped
    reflector element (see the article or web site for a
    description of the antenna) would be able to fit properly a
    few inches to one side of center.

2.  Built a single band (20M) wire model in EZNEC based on the
    original hexagon conforming as much as possible to the
    information given in the description of the antenna in the
    CQ article.

3.  Spent some time tweaking various aspects of the model to
    optimize its operation in the bottom end of the twenty meter
    band.  I was attempting to achieve the claimed F/B ratio,
    forward gain, and 3% bandwidth.


This has turned out to be an interesting exercise.  I have
learned quite a bit and was surprised at one finding.  Here is
what I'm getting from the model so far:

1.  The Hex Beam configuration is indeed a useful way to fit a
    relatively well performing 2 element beam into a smaller than
    full size space.

2.  F/B ratio in excess of 20 dB can in fact be achieved with a
    two element parasitic (reflector) array in this
    configuration.

3.  Forward gain of about 6.5 dBi (4.35 dBd) can be achieved.
    But not simultaneously with a high F/B ratio.  F/B ratio at
    maximum gain is in the 6 dB region.

4.  The feedpoint does not match 50 ohms directly.  The resistive
    component of the feedpoint impedance when the beam is
    optimized for F/B is in the mid 20 ohm range.  I used a
    hairpin to match the antenna up to 50 ohms.

5.  The three percent bandwidth claim requires some
    interpretation.  It depends on what you are using for the
    definition of tha antenna's bandwidth.  Three percent is not
    achieved for either gain or F/B.  Surprisingly, even though
    this antenna is less than full size, three percent (and a bit
    more) can be achieved with respect to VSWR.

6.  The claim on the web site that the pattern goes bad very
    rapidly as the frequency is lowered (much more so than as the
    frequency is raised) is fully confirmed.

7.  I did not find any indication that EZNEC was having any
    trouble at all modeling this antenna (NEC-2 Version).


Here is a table of the performance I could get with the model I
made (based on AWG #12 wire).  These numbers were generated for
an antenna mounted at a height of 0.5 wavelength above EZNEC
"good" ground (high accuracy ground model) and taking the azimuth
pattern cut at the maximum gain takeoff angle (27 degrees).  The
modeled antenna was optimized for maximum gain consistent with
F/B ratio in excess of 22 dB at a frequency of 14.025 MHz.

1.  Pattern Bandwidth = 66 KHz (0.47%)

    I define this for beams as the bandwidth over which the beam
    remains "pointy".  "Pointy" for me means that the F/B ratio
    remains greater than 20 dB.

    Frequency	    Gain       VSWR       F/B Ratio
      (MHz)         (dBd)      (x:1)        (dB)
    ---------       -----      ----       ---------

     13.984         3.54       1.232        20

     14.025	    3.29       1.033	    24.7

     14.050	    3.13       1.121	    20


    So this is a fairly compact beam with more than reasonable
    performance over a bandwidth of about 0.47 percent.


2.  VSWR Bandwidth = 510 KHz (3.6%)

    This is the bandwidth where the VSWR remains below 2:1.  More
    than 3% is pretty good for a compact antenna constructed from
    thin wire rather than thick tubing.  While gain varied a bit
    over the range, it did remain positive.  And the F/B ratio
    pointed the same direction throughout the frequency range
    despite quite a wide variation in value.

    Frequency	    Gain       VSWR       F/B Ratio
      (MHz)         (dBd)      (x:1)        (dB)
    ---------       -----      ----       ---------

     13.920         3.76       2.0          10

     14.025	    3.29       1.033	    24.7

     14.430	    1.39       2.0	     5.3

I don't claim that this is absolutely _all_ the performance that
is available from this configuration.  But it is all I could get
with a reasonable amount of "fiddling" time spent optimizing the
thing.  If anyone is interested in the model file as a starting
point for further optimization, I'd be happy to share it.  It is
small enough that if there is enough interest, I could even post
it.

73, Eric  N7CL


--
FAQ on WWW:               http://www.contesting.com/towertalkfaq.html
Submissions:              towertalk@contesting.com
Administrative requests:  towertalk-REQUEST@contesting.com
Problems:                 owner-towertalk@contesting.com
Search:                   http://www.contesting.com/km9p/search.htm