KØHA Minimizing Receive Noise


The following is an update on a study which I previously made in trying to compare antennas and their ability to reject a noise (e.g. static) coming from all azimuths and elevations. I thought that this could be achieved by comparing the maximum gain with the average gain of the antenna (root-sum-square value). Originally I went through the time consuming process of calculating these numbers by hand. I suggested to W7EL that he add this type of calculation to his software. Just after my suggestion he added "Average Gain" to his EZNEC software when calculating a 3D radiation pattern. My "Relative Receive Noise" figure is the difference between the maximum gain and the average gain.

Comparing my past hand calculated numbers vs. those obtained via EZNEC shows that the two numbers closely agree in improvement difference when comparing real antennas. There is a bit of a difference when comparing to an isotropic antenna due to my compromising on the number of high elevation angles used in my manual calculations. The numbers below are those obtained via EZNEC software. The figures obtained this way agree very closely with what I have observed with actual antenna use comparisons.

The numbers below were trying to show three things:
   The noise reduction due to using different Beverage antenna lengths.
   The noise reduction due to using a different number of phased Beverages.
   The noise reduction improvement by using a loaded uni-directional Beverage vs. and unloaded
      bi-directional one.

First the obvious: Noise (static) rarely if ever comes equally from all azimuths and elevations. If the noise is coming from something close to a point source then designing an antenna with a null in the appropriate direction will be a much better solution. But, in places like this one the noise comes from a wide azimuth range with a changing heading. Finding an antenna with the greatest average relative noise reduction provides a better overall benefit here in the long term as compared to reducing the noise from only one direction.

Comparing some common 160M antennas and ones which I have personally built and used produces the following results:


  Isotropic ant   0 dB
  2 ele vert     -8.12   Classic 90ºspace/90ºphase array for comparison
  My xmit ant   -10.61
  537.5 Bev     -10.57   280 ohm load
  537.5 Bev      -7.41   Not loaded
  580' Bev      -10.60   360 ohm Load
  580' Bev       -7.50   Not loaded
  881' Bev      -12.20   390 ohm load
  780' Bev      -11.85   330 ohm load
  2 x 780' Bev  -11.96   6' spacing, no stagger
  2 x 780' Bev  -13.56   6' spacing, 1/4 wave stagger feed/phase
  2 x 780' Bev  -13.84   350' spacing, no stagger
  3 x 881' Bev  -15.36

The single Beverage loads above were chosen for the best relative noise, not front to back. A compromise load value choice is often the best. The 3 phased Beverage figure is based on an example which I have actually built and used where minimum relative noise was not the only factor used in antenna design. My personal tests and use experience confirm the relative validity of these numbers. The calculations and personal use has shown that the high ground conductivity here decreases Beverage dBi signal levels a significant amount. The relative receive noise figures go down just a bit with the higher conductivity as well. And, I have tried unloaded Beverages and have always found them to be of almost no value (compared to other antennas available). Besides the above figures, on days when the noise (or QRM) is towards the back of the antenna, the relative noise improvement of a loaded Beverage as compared to a loaded one is more like 20 dB. A reversable two wire Beverage is a better choice if both directions are desired.

The figures below I thought might be a handy reference for some. They were calculated some time ago (before EZNEC) when I was designing a phased almost-parallel Beverage array. They were adjusted to match the above EZNEC numbers. This was the first time that I was using a relative noise figure to help design such an array. While heavy on the numbers, I think that observing the progression of changing one parameter (the phase of one Beverage) is useful in that it shows a variety of things happening with my Beverage array:

    0   -14.08   40   -8.16  54.3  24.97  24  41.6  Prior listed #
   25   -14.21   46   -8.47  52.8  26.86  22  40.5  
   50   -14.35   49   -9.07  50.7  28.25  21  39.5
   75   -14.51   52   -9.99  48.2  29.60  20  38.2
   90   -14.60   53  -10.59  46.8  28.35  19  37.3  Chosen design
  100   -14.62   55  -11.26  45.5  28.70  18  36.0
  125   -14.45   58  -13.03  43.1  26.06  17  34.7
  150   -13.43   60  -15.29  44.9  22.21  16  32.9

  AZ = Center of azimuth lobe
  dBi = dBi value of above lobe
  BW = -3 dB bandwidth of azimuth lobe
  F/B = Front to back of main lobe (NOT to "back" of Beverage wire)
  EL = Center of elevation lobe at peak azimuth
  ELBW = -3 dB bandwidth of elevation lobe

Note that while the Beverages all stay with a 40 degree azimuth, the main lobe shifts clockwise by increasing the phase delay to one Beverage. Being able to adjust the phase of that Beverage (via a MFJ-1026) can be very useful! Looking at the pattern, as my receive noise figure improved, I wasn't expecting a "less clean" looking pattern. However, the decrease in the main lobe width obviously made up for that. Also apparent was the lack of validity in only being concerned with the main lobe width figure. Using this and later phased Beverage arrays has shown that it takes some time to familiarize yourself with the action caused by MFJ-1026 control adjustment. However, once that's done, I've been able to note propagation path changes when there's a different control setting required for best reception.

For more information on this "relative receive noise" figure, see W8JI's site covering Receiving Basics and Receiving Antenna Design: W8JI RDF
He named the noise figure which I pioneered the Receiving Directivity Factor (RDF). He has a variety of other excellent information available at his home page: W8JI Home

If your are interested in the origination of the "relative receive noise" parameter/RDF, please review the following:
W8JI had nothing to do with its first use. First, see Tom's comment below from the following message: http://lists.contesting.com/archives//html/Topband/2000-10/msg00108.html
From it: "The person would could figure out how to use the pattern tables that are exported by Eznec in a spreadsheet to determine S/N ratio would be a top-band hero, because the data would express the effectiveness of an antenna for receiving. We need directivity, not gain."

I previously addressed this over a year earlier, initially: http://lists.contesting.com/archives//html/Topband/1999-09/msg00123.html
and: http://lists.contesting.com/archives//html/Topband/1999-09/msg00149.html

I responded to W8JI's email quoting part of that second message: http://lists.contesting.com/archives//html/Topband/2000-10/msg00117.html

The K6SE reaction to my information is at: http://lists.contesting.com/archives//html/Topband/2000-10/msg00134.html

I had been quietly using my receive antenna figure of merit in the analysis of my three almost parallel Beverage array and my three element parasitic vertical array for over a year while confirming its validity before I mentioned it on the Topband reflector.

I commented about my prior email to W7EL suggesting that he included my figure in his next revision of EZNEC in the following: http://lists.contesting.com/archives//html/Topband/2000-10/msg00137.html
From that message: "This would be the root-sum-square of all azimuths (and elevations) of radiation, including the main lobe (the rms value of the antenna). [2013 Comment: The next revision of EZNEC after my email to W7EL included the "Average Gain" figure for the first time]. I tried those computations and found that they weren't representative of the noise reduction effects of my receive antennas as compared to my figures which leaves the main lobe out of the calculation. However, The average gain figure and my [preferred] number both will generally point to the same antenna having the best directivity. Due to that, I've been using the average gain figure as an easier method than my prior manual calculation. However, I think that the average gain figure underestimates the difference as compared to what I've witnessed with my own antennas." It's nice not to be required to manually calculate the root-sum-square of over 1,000 azimuth/elevation points.

The above shows that I documented the RDF figure using both true root-sum-square "average" gain and the average gain outside of the antenna's main lobe way before W8JI started experimenting with it. He might be able to claim coming up with a nice name for my parameter.


160M Page    NDB   
6M Page


Send KØHA e-mail (Remove all X's from the address)