A couple of words about RDF (Receiving Directivity Factor) and DMF
(Directivity Merit Figure). Both of these may be used as a kind of
figure of merit for receiving antenna performance, but neither of these
fully characterize everything that is important in determining
signal-to-noise ratio for a receiving antenna. A degree of judgment
needs to be applied in addition to looking at the number values.
RDF was originated by W8JI and it is calculated by measuring the forward
gain in the "desired" direction minus (in dBi)the total average gain of
the antenna. This is easy to determine because EZNEC will calculate
both of these. It's generally a good parameter, but there are a few
problems.
The first problem is, what is the "desired" direction. Most experienced
DX'ers will have a good idea as to what is needed, but the inexperienced
ham won't, and the desired direction may be different for different
applications. But maybe that's OK on an individual basis because
conditions may be different for different people or for different
applications. For example one person may choose 25 degrees elevation,
and another person may choose 15 degrees elevation as the desired angle
and the direction to be in-line with the lobe. I saw a write-up on one
guy's website where he used the desired direction as whatever the
elevation angle was for peak gain. That made a low dipole look really
good. The point is that numbers generated by different people are going
to be different for the same antenna.
The second problem is that the parameter makes no distinction for lobe
shape. In other words, you can squish the forward lobe into any shape
you desire as long as you don't change the "gain in the desired
direction" and overall average gain (just like a balloon) and the RDF
number won't change. The shape however will change the S/N ratio.
The third problem is that a lot of the response in the forward lobe can
be moved to the rear (just like the balloon analogy again). If the
"gain in the desired direction" and average gain are not changed, the
RDF number won't change. Front-to-back ratio MAY be more important than
a wider front lobe to some people.
The point is that the signal-to-noise ratio of the antenna may change
significantly and the RDF number remain the same.
ON4UN recognized some of these short-comings in the RDF parameter, and
developed his own parameter, DMF. DMF is the gain in the desired
direction minus (in dBi) the average gain in the rear half hemisphere.
This is a little different from RDF, but it has some of the same
problems. We still have this "desired" direction, so it has the same
problems as listed above which may make the number different for
different applications. The DMF does give a lot of weight to
front-to-back ratio, and so precludes the case of "squishing" some of
the forward lobe into the reverse direction without the number changing.
It doesn't help the problem of squishing the front lobe into various
shapes however. In other words, the front lobe could be horizontally
very wide and vertically very narrow, or you could make it very narrow
horizontally and very wide vertically. Both would have the same DMF
(and RDF). Clearly the lobe with narrow vertical response would be much
more desirable.
When I first looked at these two parameters and saw the problems, I
started developing my own method of measurement, however I never
finished it. With a little expertise in recognizing what is important in
terms of the shapes of the response curves, you should be able to
determine which is best. This however doesn't give you a number, and so
comparison between a large number of antennas becomes difficult. I wish
I had finished my method of measurement now, because it would be nice to
have something that provided better characterization.
Another item of importance but not necessarily a large player in
determining the S/N is the location of the nulls. These may be important
for rejection of QRM from certain directions. That may give more weight
to obtaining a good front-to-back or a null at 90 degrees, depending on
where you are located and what QRM you would like to reject.
Jerry, K4SAV
Terry Conboy wrote:
> I've followed with interest the discussion about phased arrays of
> K9AYs vs. 4 squares, such as the DX Engineering receiving array. I
> decided to model a few variations, and they back up K4SAV's & NI1N's
> comments pretty well. As the basis for performance comparison, I used
> the RDF (receiving directivity factor) employed by W8JI. Here are the
> results:
>
> Antenna RDF, dB
> ========================
> Whip 4.7
> K9AY 7.3
> EWE 7.4
>
> 4Sq Whips 90* 10.1
> 4Sq K9AY 90* 10.2
> 4Sq EWE 90* 10.2
>
> 4Sq Whips 120* 11.6
> 4Sq K9AY 120* 11.4
> 4Sq EWE 120* 11.4
>
> 1 wl Echelon-Bev 11.5
> 1.5 wl Beverage 11.6
>
> [The 4 squares all use 1/4 wl spacing. The 120 degree phasing system
> is very close to what results from using the crossfire phasing scheme
> recommended by W8JI and DX Engineering at that spacing. The crossfire
> system is broadband and can be used over several bands, and the
> current phases will vary accordingly.]
>
> The bottom line is that a single K9AY or EWE (which are essentially
> equivalent if properly installed) is a better receiving antenna than a
> single whip vertical, but using them in arrays may be much more
> complex than required for good low noise receiving performance. The
> only notable improvement for the arrays of EWEs or K9AYs is a
> reduction in the off-axis back lobes, which could be useful for
> attenuation of some low angle QRM. On the other hand, as K4SAV
> describes, the array of verticals has a null at zenith, which could
> help attenuate lightning static and NVIS QRM.
>
> It's also interesting that all of the 4 Square arrays are very similar
> in performance to a 1.5 wavelength Beverage or an echelon of 1
> wavelength Beverages (staggered by 0.186 wl) with 135 degree phasing
> (although the back- and side-lobe structure is different.)
>
> Incidentally, models of the DX Engineering array actually show
> slightly better RDF (a few tenths of a dB) at closer spacings than 1/4
> wl (with appropriate phasing). I suspect the arrays of EWEs and K9AYs
> would be similar.
>
> 73, Terry N6RY
>
>
>
>
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|