Excellent discussion John and Terry.
Interestingly enough if you take the 8 elements in a 200 foot circle you
can optimize the peak RDF to nearly 15dB. M.T. Ma in his book outlines his
math to optimize feeds for the specific purpose of maximizing directivity.
Using his mathematics produces 15 dB of peak RDF in this circle with
horrible side lobes and a narrow beam. If you set the 8 elements for
"crossfire phasing" you can get an RDF of 14 dB or so. However using either
of these phasing and amplitude schemes requires a circuitry system accuracy
that is presently unobtainable for backyard installations. Relaxing the RDF
to 13.45 and playing with the phasings allows an array to be realized using
the technology at hand. Lowering the available RDF of these 8 elements to
13.45 causes the close in side lobes to decrease and the 90 degree side
lobes to increase. Interestingly enough these 90 degree side lobes are at a
very low elevation angle. Further lowering of the RDF allows a very clean
pattern as John and Terry have pointed out. At my location the side lobes
generated by keeping the RDF at 13.45 cause none or little harm to the real
SNR at my location Just as Terry has pointed out.
My goal for designing the all active 8 element array was the very same as
Joel has mentioned and that was to build the best performing array that I
could for myself in my environment. After all the tests I have made at this
location comparing receiving antennas, I wait with great anticipation the
tests Joel is making. It will be most interesting to me if one can tell the
difference between systems with an RDF within a dB or so on very weak
signals. I have successfully measured 1 dB difference in S+N to N ratios
between antennas using Spectran. There are things to be learned here.
From: Topband [mailto:firstname.lastname@example.org] On Behalf Of Terry
Sent: Wednesday, December 17, 2014 7:57 AM
To: 'John Kaufmann'; email@example.com
Subject: Re: Topband: 8 circle: DXE vs Hi-Z
Bravo, John! Well stated on all points. No disparaging comments noted. I
especially appreciate your discussion of RDF and what it actually means in
As you commented, in practice, RDF is calculated (EZNEC, etc.) assuming
equal amplitude distribution of noise over all 3D free space. On 160m, in
North Florida, I am sure that the assumed equal distribution of noise is
never the real case. It is actually possible to calculate RDF for any
arbitrary spatial distribution of noise, but to do so you must
mathematically characterize the noise distribution in 3D - obviously a
formidable task at 1.8 MHz. Usually everyone just takes the simplifying
leap and assumes that all the bad noise is coming from off the back and the
sides of the antenna, in some well-behaved average way. Noise sourced from
the front of the antenna pattern is largely ignored.
Antenna patterns are often "optimized?" to reduced side lobes to near zero
levels. The main lobe is generally broadened as a result of such side lobe
optimizations. Now consider the case of non-uniform noise distribution,
with a high noise level broadly sourced at the front of the antenna pattern
and lower noise levels sourced on the sides and back of the antenna pattern.
By minimizing noise reception in the side lobes, the main lobe is now
broadened and thus is exposed to a greater solid angle of high noise source.
Furthermore, the increased exposure to high noise takes place in the main
lobe, which has the highest pattern field gain. The actual antenna RDF would
be substantially degraded as a result of additional received noise power.
For this example, optimizing the antenna pattern for minimum side lobes
would actually degrade the antenna's environmental SNR.
RDF is a very useful metric for comparing receiving antennas. But, we must
use the concept in its entirety - we cannot ignore the system aspects that
are hard to measure, calculate, or characterize. Perhaps W7EL will
incorporate an arbitrary 3D noise model in his next EZNEC update?
John wrote in part:
...RDF as a receiving metric:
RDF is indeed a very useful metric for comparing receiving antennas.
However, we need to be aware that it assumes the ambient background
(atmospheric) noise is uniformly distributed in 3-dimensional space, which
is not always true in specific instances. For this reason, RDF may not
exactly predict the differences between two arrays in any given situation.
It is possible for a system with a lower RDF to equal or even outperform
another system with higher RDF under certain noise conditions. If the noise
were always uniformly distributed, then RDF would perfectly predict relative
receiving performance (actually SNR).
The next point about RDF is that it is calculated for a specific signal
arrival direction in three dimensional space. In terms of azimuth, it is
the peak direction of the forward lobe. In elevation, it is common practice
to use 20 degrees, which can be considered appropriate for DX reception. If
the signal arrives from a different azimuth or elevation angle, the SNR
advantage predicted by RDF may not actually be realized. I have seen a
simple low dipole with a lousy RDF occasionally outperform my 8-circle
system by a large amount when the elevation angle of arriving signals is
very high and the RDF advantage of the array cannot be realized.
As RDF gets higher, the beamwidth of the antenna system generally gets
narrower. You can see this if you look at chart #2 in K7TJR's Dayton
presentation (http://www.kkn.net/dayton2014/HiZ_DAYTON_2014_7n2.pdf). This
brings up another point. By making the RDF very high, you are necessarily
restricting the angular sector over which the antenna delivers its best
performance. This is fine as long as the angular sector coincides with a
direction that is important to you. The flip side is you give up some of
that performance outside that sector. For switched arrays with a finite
number of selectable directions, that could be a disadvantage when a
direction of interest falls halfway between contiguous switching directions.
Looking at the pattern of the array will tell you what you give up in the
"in between" directions.
These comments with respect to RDF are not intended to be disparaging. On
the contrary I do believe RDF is an excellent tool for comparing receiving
antennas. You just have to aware of what it actually means in practice...
73, John W1FV
Topband Reflector Archives - http://www.contesting.com/_topband
Topband Reflector Archives - http://www.contesting.com/_topband