[Antennaware] Dual fed K9AY

Pete Smith n4zr at contesting.com
Tue May 26 13:16:13 PDT 2009

Maybe taking it farther afield, but I have done modeling experiments 
(plus one winter's Europe-only experiments) using two K9AY loops 
end-fire, spaced 135 feet, with 180-degree phasing obtained by simply 
swapping the leads to one input transformer and feeding the two loops 
from a coax tee with equal-length feedlines.  The model suggests an 
RDF of 11+, which is a whopping improvement over a single loop and 
gets into Beverage territory.  The phasing produces two deep nulls 
perpendicular to the axis of the main lobe, while still maintaining 
useful F/B ratio and reducing the peak takeoff angle to about 23 
degrees.  F/B is actually better in the model using less phase shift, 
but I settled for 180 degrees because according to ON4UN at that 
angle no phase errors are introduced by any mismatch.

The test pair of antennas worked pretty well, but I lacked the 
experience and instrumentation to experimentally verify the 
pattern.  I have been noodling over the idea of a 4-loop array, using 
single loops at each end and a classic 4-way loop at the intersection 
in an L pattern of 135 feet per leg.  If I can figure out some 
reasonable switching, that should give me NW, NE, SW and SE with 3-db 
bandwidths of about 80 degrees.

Anyone want the model?

73, Pete N4ZR

At 12:05 PM 5/26/2009, you wrote:
>Excellent thread here guys.  So much so that I have switched out of
>digest mode and into real time mode.
>A few notes on the K9AY - some perhaps more relevant than others.
>1.  The far field pattern is ALL vertically polarized - consistent with
>the antenna behaving as two verticals with a phasing line connecting
>2.  The connection to ground is 100% common mode to the two verticals.
>All my K9AY modeling includes what I call Rg, which is a resistance in
>the ground leg.  Values around 100 ohms appear to give good correlation
>to field measurements I have made of the pattern here in Georgia, where
>I have decent ground properties.  Front-to-back is very sensitive and to
>some lesser degree gain is sensitive to Rg, exactly as you would expect.
>I can't figure out how one would model the antenna as two separate
>verticals, as I can't see how the common mode resistance could be
>included in such a model.  Maybe someone has a bright idea for this?
>Meanwhile, I don't really see any point in modeling the antenna as two
>verticals since the model of the antenna as a loop gives good results.
>3.  Radials cannot be placed directly under the loop wires without
>severe pattern distortion.  They can be placed at right angles to the
>loop.  Many of us run orthogonal K9AY loops - the best answer for that
>configuration appears to be radials at 45 degrees to the loops.  I use
>no radials - just a six foot ground rod, in part because I'm not
>interested as much in F/B as I am in RDF.  Meanwhile I have excellent
>F/B with just a ground rod.
>4.  While the F/B is very dependent on the termination resistance, the
>actual noise performance of the antenna, as defined by RDF, is a
>different story.  RDF changes only about plus or minus 1% for
>termination resistances from 340 to 600 ohms, with Rg=100 ohms.  I need
>to pull out my model runs from a few years ago and check RDF as a
>function of Rg.
>It may take a day or two for me to get back to this, but I am fascinated
>by the possibilities.  I'll stick my neck out and take a guess: I don't
>expect to see nearly any improvement in gain, but with variable delay in
>a dual feed loop you should be able to adjust the tilt angle of the null
>in the back which will make the antenna APPEAR to have more F/B to the
>user, while in fact it probably doesn't.  I'm also guessing RDF will
>remain nearly unchanged by going to dual feed.  With a phasing box like
>the DXE or MFJ you should be able to adjust both the tilt and the depth
>of the rear null.  If cross-fire feeding is used, the settings for delay
>and amplitude of the phasing should hold reasonably constant over a wide
>frequency range.
>I hope I'm wrong!  Just being able to control the rear null tilt angle
>however is worth running the second coax for me.
>I have a movie and presentation on the K9AY loop at www.sedxc.org    As
>the movie shows, a properly built K9AY, without radials, can produce
>phenomenal F/B under some conditions.  Isolating grounds and ferrite
>beads to prevent further pattern deterioration from common mode current
>on the feedline are the key to getting these things working well.  And,
>of course, keeping them uncoupled from everything else in the yard
>that's vertically polarized - like transmit verticals.
>Hal N4GG
>Antennaware mailing list
>Antennaware at contesting.com

More information about the Antennaware mailing list