This article describes the design of an antenna for "local" contacts on 7MHz, including a simple and efficient matching system that presents a 50Ω load to the transmitter.
The design objectives are:
The antenna is intended to serve mainly local VK contacts. The requirement can be simply met by an antenna with approximately omni-directional characteristics. Even though the location is a semi rural one, a horizontal antenna is chosen for best noise performance.
VK1OD is located in Bowral (abouty 100km SW of Sydney), roughly in the centre of the south east corner of Australia where well over half of Australia's population reside, as illustrated in the map to the right.
The table shows the nearby state capital cities (2002) and the path parameters for 7MHz communications at 0500 UTC with SSN=100.
Note that radiation angles from 31 to 82 degrees suit these cities.
A dipole mounted low to the ground with its legs sloped downwards from the centre was chosen as a reasonably omni-directional horizontal antenna. A mast to support the centre of the dipole at a height of 11m is available, and the legs can be conveniently sloped downwards at about 45° to the horizontal.
The impedance of a nominally half wave dipole is quite dependent on
frequency, height above ground, the nature of ground, and nearby
conductors. The series resistance component changes slowly with
about resonance, and reactance changes quite quickly with frequency.
Fig 1 shows NEC modelled R and X for the antenna which is resonant at
As the frequency increases, the change in feed point impedance can be expressed as the VSWR it would cause on a given feedline. Fig 2 shows the VSWR in a 75Ω line. Conversely, as the length of the antenna is increased beyond resonance at a given frequency, VSWR(75) will increase.
The objective of the matching scheme used is to cause a VSWR of 1.5 on the 75Ω feedline because the effect of that VSWR is that there will be points along the feedline where the the ratio of V to I, the impedance, is 50Ω purely resistive. In the strictest sense, we want VSWR=1.5 at the transmitter end of the 75Ω line, VSWR will be a little higher at the feed point end due to line losses.
First step is to erect the dipole with 1:1 current balun at the feed point and tune it to obtain VSWR=1.5 in the 75Ω feed line at the desired frequency by making it a little longer than a resonant dipole. This is best done by measuring VSWR with a 75Ω SWR meter at the transmitter end of a a feed line length that is a little longer than anticipated for the installation (see below).
Fig 3 shows the feed point impedance that would be transformed by a
length of Belden 1189A (RG6) at 7.1MHz.
The NEC model give above gives specific values for feed point
impedance for the modelled antennas, but the installed antenna will
depart from the model to some extent because the model is a
simple one that does not capture the entire antenna environment, and
some parameters (eg soil type) are guesses and have a certain amount of
error. Nevertheless, we can use the model to guide us to a solution
that is tuned for the actual installation.
We expect that a dipole like that described above, when tuned for VSWR=1.5 in a 75Ω feed line will have a feed point resistance component typically somewhere in the range 60Ω to 80Ω. From Fig 3, select a length of line that is convenient and will suit that resistance range with positive X (since the dipole is resonant below the desired frequency). Fig 3 indicates a length of around 13m will suit, lets choose 15m as a starting point to cover that range.
Now, connect a 50Ω line from the transmitter to a 50Ω SWR meter and then to the 15m of 75Ω line. Measure VSWR(50) at the desired operating frequency, and at adjacent frequencies. It is likely that the VSWR is greater than one, and is less at lower frequencies. Cut a little off the 75Ω line, and repeat the measurements. Repeat the process until the VSWR(50) minimum is at the desired frequency, and it should be very low, less than 1.1.
Fig 4 shows the measured VSWR, Rx and Xs from 6MHz to 8MHz looking into the tuned length of RG6.
Apologies for the graphs, the TAPR VNA software is poor. The red VSWR
scale is from 1:1 to 3:1, so the second grid line from the bottom is
1.4:1. The blue line is R and scale is 0 to 100, and orange line is X
and scale is -50 to +50. Note that the R and X lines in Fig 4 are not
directly comparable with Fig 3 (Fig 3 shows the feed point impedance required
for input impedance of 50+j0Ω vs length, whereas Fig 4 shows the input impedance
with the optimised tuned length of feed line).
Whilst it is possible to adjust the dipole length and feedline line
length for a perfect match, it is impractical as factors like
variations in soil moisture content after rain, seasonal variation in
vegetation etc will cause small variation in the input impedance.
The dipole could be tuned so that resonance is above the desired frequency and VSWR(75)=1.5 at the desired frequency by shortening it from resonant length. This means that feed point impedance would typically be in the range 60Ω to 80Ω and reactance is negative. Looking at Fig 3, you might choose a line length of 23m as the starting point and then adjust the line length until VSWR(50) looking into the line is very low (<1.1).
Fig 5 shows the feed point detail. The RG6 feed line is inside a length of 13mm irrigation pipe as protection from birds, and a W2DU style balun formed of 12 suppression sleeves on the RG6 fits inside the irrigation pipe.
Fig 6 shows an adjustable capacity hat for fine tuning the antenna. It is just a piece of HDC with a dog leg in the middle, and secured to the dipole wire with a split bolt line tap about 900mm from the end.
Fig 7 shows the insulator termination detail. The copper is terminated using crimp sleeves particular to the purpose. These were widely used on open wire telephone lines, but probably not obtainable now.
Fig 8 shows the method of tensioning the dipole. The star picket is a temporary measure and will be replaced with a length of galvanised steel pipe. The winch fitting is a rural fence accessory.
Fig 8 shows the earthing of the supporting mast to a 2.4m long 16mm copper clad ground rod driven into wet clay. Two 25mm2 conductors bond the mast to the earth rod. The ground rod consistently measures around 12Ω resistance indicating a soil resistivity of around 20Ωm (which is quite low). When connected to the mast in its foundation, the combined resistance consistently measures 9Ω to 10Ω.
The temporary end supports were replaced with 50mm diameter steel posts. This raised the ends of the dipole to about 7m in height (included angle 135°) and required a small adjustment in tuning.
Fig 10 shows the end support posts. They are made from 40mm NB (~50mm OD) galvanised steel pipe, with cap added and a chain link welded on for rigging attachment. The winch pictured is a permanent fence wire strainer which cost about $7 at the local rural hardware shop. The other pics show the lateral guys (4mm galvanised FSWR) on the mast at 8m height, and their attachment to an eye in the roof frame and a star picket driven into the ground. The mast was trued up laterally using a theodolite and then the rigging screws and shackle pins wired to prevent loosening. The egg insulators are to prevent unwanted interaction with the dipole.
Fig 11 above shows a scan with a AIMuhf antenna analyser look into the existing 13m of RG6 coax. The capacity loading stingers were adjusted in position to obtain VSWR(75)=1.5 at about 7.070MHz. The VSWR lot above is referenced to 75Ω, and the cyan cursor shows VSWR(75)=1.5 at 7.059MHz. This is quite close enough, antenna tuning will change a little with ground moisture level etc.
In this case, the stingers needed to be move about 100mm to compensate for the new end height which could be expected to affect feed point reactance much more than resistance.
Having used this stinger design for many years, it was time to improve the design. Fig 12 shows the improved design, the 2mm HDC stinger wire is formed around a 6mm round rod, then through the split bolt line tap, and the ends bent out at 90°. The improved stinger seats better against the main line, won't fall out easily, and can be removed without fully undoing the line tap.
Next step is to determine any adjustment needed in the tuned 75Ω coax section.
Fig 13 above shows the same scan referenced to 50Ω. Minimum VSWR(50) occurs at 7.060MHz, and as expected, it is quite low (1.04). It is not altogether surprising that the existing length of line remains optimal as the retuning with the stingers was mostly to correct a shift in reactance caused by the new mounting height.
Note that in Fig 13, minimum VSWR is extremely low (<1.05), and reactance does not pass through zero around minimum VSWR. This is an example that shows the nonsense of the widely held belief that system resonance is synonomous with minimum VSWR, and that reactance always passes through zero coincident exactly with minimum VSWR.
Fig 14 above shows the new cable entry panel. A brick was removed from the wall and an aluminium panel folded up to cover the space and accommodate four N type bulkhead connectors. This is the garage wall, and the inside is unlined which gives ready access to the other side of the connectors.
Fig 15 shows a pair of ceramic feed through insulators with metal hardware fabricated from 5mm stainless threaded rod, nuts and washers. The feed through insulators are to facilitate entry of an open wire feeder.
The panel is grounded via a 6mm^2 conductor behind the wall.
The 13m of RG6 feedline has a loss under the mismatched conditions of 0.22dB and cost A$5. Another 2m of RG58 fly lead to make the distance to the transmitter has a loss of 0.07dB at a cost of less than A$2, giving a total feed line system with loss of 0.3 dB (efficiency 93%) for less than A$7 and delivering a very good match for the transmitter. The balun is about A$10, so all up, the antenna cost is about A$17 plus appropriate copper wire.
An NEC model of the dipole indicates copper loss of 1.8%.
Common mode feedline current at the feed point adjacent to the balun was measured at 50mA with 100W delivered to the antenna system. The choke impedance is 560+j780Ω, so power dissipated in the balun at 100W is I^2R=0.05^2*560=1.4W, 1.4% of transmitter output.
Fig 16 shows where the transmitter power goes.
|1.02||22/05/2011||New higher end supports and revised tuning.|
VK1OD on the 'net. I appreciate feedback, click on the ... in webm...@vk1od.net for my email address.
© Copyright: Owen Duffy 1995, 2012. All rights reserved. Disclaimer.