[TowerTalk] supergain

rick@area.com rick@area.com
30 Jan 2000 18:50:57 -0000


A quick note here on what the theoreticians say about "supergain"
which they call "superdirectivity".   This discussion assumes lossless
elements and matching networks (for now).  The most gain you can get
additively is a power gain of N (in dB 10 log N) for N elements.  This
occurs for spacings of around 1/2 wl.  When you go to closer spacing,
for an optimum driven array, the maximum achievable gain goes up, due
to subtractive gain.  Unlike the Yagi principle that gain is proportional
or whatever to boom lenghth, the driven array optimized for maximum gain
just gets more and more gain as the spacing is decreased.  This is no
theoretical limit to this gain.

Now when we leave the theoretical world and get back to real world antennas,
we find that you can actually only get 1 or 2 dB of supergain before
the bandwidth, ohmic losses, and tolerance sensitivities eat you alive
as the theory predicts that it will.  Conventional Yagi designs can also
get this kind of supergain, so the decision to use an all driven array 
would have to be justified on  some other basis that gain.  Another problem
is that if you maximize gain, you don't get low sidelobes.

It is very easy to illustrate these principles with EZNEC.  Just build a
four element array with say .1 wl spacing and drive it so that the 
elements almost but not quite cancel each other out.  The gain will be
surprisingly high for a .4 wl beam.  The bandwidth will be very narrow
and the currents will be very high.

Rick Karlquist N6RK
www.karlquist.com

--
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