Some Optimized Antenna Designs

I have been dabbling lately with a simple Optimizer code to work in conjunction with my xnec2c application, a translation to C of the well-known FORTRAN program NEC2. xnec2c has an option that enables the printing of Frequency dependent data (VSWR, Impedance, Gain etc) to a CSV file for use by other programs. Also, xnec2c monitors its .nec input file for changes and re-runs the frequency stepping loop which recalculates new data and prints to the .csv file. It is therefore possible to arrange the optimizer program to read the .csv data file, recalculate antenna parameters and save them to the .nec input file. xnec2c will then recalculate and save new frequency-dependent data to the .csv file. If the optimizer program is arranged to monitor changes to the .csv file, then a continuous loop can be created in which new antenna parameters are calculated and saved to the .nec file, new frequency dependent data are calculated and saved to the .csv file and the loop repeated until the desired results (optimization) are obtained. So I have used my optimizer program to try and refine the design of a few antennas and I have also come up with what I think could be new antenna designs. Here is one that should be of interest to users of G5RV or Windom antennas:

The 5B4AZ 6-40m Multi Band Dipole Antenna

Both the G5RV and Windom wire antennas have been very popular with radio amateurs looking for a relatively simple wire antenna that can cover most of the available HF amateur radio bands. However, both types of multi-band wire antenna cannot cover some of the frequencies now available in the HF spectrum.

I therefore had the idea of trying to combine the design principles of these two antennas, hoping to come up with a new design that will provide coverage of all HF bands in its range. After some preliminary work in "xnec2c", my own graphical version of the NEC2 engine, I came up with a design that looked promising. It was difficult to optimize it by hand, so I then wrote a rather simple but successful optimizer application which I used to automatically optimize the original design.

The result is a design that covers all bands from 7 MHZ to 50MHz with a VSWR of less than 3:1, which makes it very easy to tune with either an outdoor auto tuner at the feed point or, if the coax run is not too long, with an indoor tuner. The input impedance of this design is generally 200 Ohm so a 4:1 balun is needed. And it is also possible to use the antenna on 80m, but the input impedance at resonance is about 12 Ohm so it would require a 1:1 balun to be switched in instead.

The drawing below shows the design and the dimensions of this wire antenna. The shorter half of the dipole is 12.87m / 42' 3" long and the longer half 16.86m / 55' 8" long. The balanced feeder is 450 Ohm and for an open wire feeder it would be 11.86 m / 38' 11" long. For a window line with a velocity factor of 0.91, the feeder would be 10.79m / 35' 5" long. Contrary to the G5RV and Windom, the input impedance of this antenna is ~200 Ohm so a 4:1 balun is needed. It is possible to use this design on 80M with a 1:1 balun.

5b4az-dipole

The screen capture below shows the VSWR plot for this antenna, produced by xnec2c, over a frequency range from 7 to 51 MHz. The VSWR at all the amateur bands between 40m and 6m is 3:1 or less, making this antenna easy to tune on all bands.

vswr plot

Unfortunately due to age related problems and living in a small island with few, if any, materials available for amateur radio home building, even antennas, I have not been able to construct an example of this design. However I fully trust the NEC2 engine to correctly evaluate this type of antenna design. I therefore decided to publish this design, hoping that it may be adopted by radio hams with a better opportunity of building and using it. It should even be attractive for commercial producers and vendors of wire antennas!

Here is the design file used in xnec2c to evaluate the design. The ZO "card" is an xnec2c extension that is not compatible with NEC2 or other derivatives so it should be removed if the file is to be used with software other than xnec2c.

Classic Windom Antenna optimized for 6-40m

This is an optimized design of a Classic Windom (fed with an open wire feeder) for the 7-52 MHz range. Unfortunately the 30m and 17m bands could not be covered by this design, however input impedances are not extreme so they can be matched by a tuner at the base of the antenna. On the rest of the bands, the input impedance is generally 200 Ohm so a 4:1 auto transformer is needed. Here is the design file for this antenna. The ZO card is an xnec2c extension so it should be removed if used with NEC2 or other derivatives.

The drawing below shows the design and the dimensions of this wire antenna. The shorter half of the dipole is 6.52m / 21' 5" long and the longer half 14.76m / 48' 5" long. The single-wire feeder is 12.22m / 40' 1" tall.

6-40m Windom

The screen capture below is the VSWR curve for the antenna. Although the 30m and 17m bands are not covered by this design, the input impedances are not extreme so they can be tuned by an antenna tuner at the base.

6-40m Windom vswr plot

Classic Windom Antenna optimized for 10-80m

This is an optimized design of a Classic Windom (fed with an open wire feeder) for the 3.5-29 MHz range. All bands in this range are covered with a VSWR of less than 3:1. The input impedance is generally 200 Ohm so a 4:1 auto transformer is needed.

The drawing below shows the design and the dimensions of this wire antenna. The shorter half of the dipole is 14.70m / 42' 3" long and the longer half 27.45m / 90' 1" long. The open wire feeder is 11.95m / 39' 3" tall. Here is the design file for this antenna. The ZO card is an xnec2c extension so it should be removed if used with NEC2 or other derivatives.

10-80m classic Windom

The screen capture below is the VSWR curve for the antenna.

10-80m Windom vswr plot

The Extended Double Zeppelin Yagi antenna

This Yagi-type antenna uses a driven element that is 2x 5/8 lambda in length, thus gaining about 3dB gain over the usual 1/2 lambda dipole. The reflector is a little longer than 1.5 lambda and the director a more conventional 0.45 lambda in length. The gain of this antenna is about 9.2 dB, substantially higher than the gain of a classic 3 element Yagi. The input impedance is about 75 Ohm with a reactance of about -300 Ohm so a tuning/matching network is needed for a good match to 50 Ohm. Here is the design file for this antenna.

The drawing below shows the dimensions of the antenna. The driven (middle) element is 2.53m / 8' 4" long. The reflector is 3.03m / 9' 11" long and spaced 0.45m / 1' 6". The director is 0.87m / 2' 10" long and spaced at 0.41m / 1' 4".

2m extended Yagi

The screen capture below shows the gain curve of this antenna and its input impedance values

gain

This design can be used in a circularly-polarized Yagi antenna, if a second structure is placed 1/4 lambda ahead of the first and at 90 degrees to it. It is then only necessary to feed both antennas in-phase, since the necessary physical and electrical phasing is taken care of by the structural design of the antenna. Here is a design file for such an antenna.

Neoklis 5B4AZ
http://www.5b4az.org