Home

Old Site

Natural History

Radio Valves

Amateur Radio

Computer stuff

Me

A Portable 20 Metre DX Antenna

A Vertical 2-element Moxon Rectangle

Radiation pattern
Elevation radiation pattern - scale is dBd.

Introduction

I built this because I needed a compact and portable aerial, to take on holiday, that would give me a reasonable chance of working some DX with a small tranceiver operating from a car battery.

Towers, poles and rotatable yagis were obviously not an option. At my holiday destination "throwing some wire into a tree" wouldn't work either.

This is an easily constructed portable antenna that can be carried and erected by one person. It uses readily available components to give a gain and radiation pattern comparable to a 3 element beam at 20 metres AGL.

History of the design

Although the short two element beam, with the ends of the reflector and driven elements folded toward each other, with a small gap, to provide loading, is often credited to L.A. Moxon, G6XN and known as a 'Moxon Rectangle', he credits the design to VK2ABQ. (1) ; Moxon was perhaps too modest as VK2ABQ used a square antenna and Moxon's Rectangle was a significant improvement on the basic idea.

The design is most often used as a horizontal beam with diagonal spreaders and is built as a multiband (10, 15 and 20) array. If used horizontally it needs to be at least ½λ high (more is better) to get low angle radiation, as with any horizontal antenna.

However, if the antenna is mounted vertically it is very insensitive to height, as long as the lower ends are 1 to 1 ½ metres AGL, and it produces a very respectable forward gain, a good front to back ratio and a very wide beam , so that, for instance a west facing beam will provide useful gain to the whole of the new world.

This is not a new idea and there are several designs on the web. I downloaded Pete Millis, M3KXZ's EZNEC file (2) for 20 metres as a basis for experiment in EZNEC. His file has poor choice of the number of wire segments and resegmenting is strongly recommended before use. Poor segmentation can cause significant errors in predicted fields.

I am using different materials and have slightly different dimensions. The dimensions are quite sensitive to the wire diameter and to the insulating material. (I give EZNEC files for my design below). Pete's own website (3) was offline at the time I looked (it is back now) so I decided to do my own implementation.

Pete gives some cautions about EZNEC and average gain. Please note these if you are using the demo version of the software. My design does not seem to show these problems using many segments and EZNEC 3.0 or EXNEC+ 5.0.

Design and simulation

The design was simulated and tweaked with the EZNEC+® v.5.0.14 (4) antenna modelling software. The initial EZNEC file was based on M3KXZ's. Some scaling was done to move the resonant frequency to 14.2 MHz.

Diagram of antenna
Wires with radiation pattern and poles.

Wire lengths

The wire lengths used in the models, numbered in the diagram, are given below. Two sizes are given, both for ETFE insulated, 19/02 silver-plated wire. The SSB lengths give a centre frequency of 14.2 MHz and an SWR below 1.5:1 over the range from 14.1 - 14.35 MHz. The CW model is centred on 14.1 MHz and gives an SWR below 1.7:1 over 14.0 - 14.35 MHz.

Wire Length (for SSB) Length (for CW)
1 1,171 mm. 1,177 mm.
2 7,394 mm. 7,431 mm.
3 1,171 mm. 1,177 mm.
4 1,446 mm. 1,453 mm.
5 7,394 mm. 7,431 mm.
6 1,446 mm. 1,453 mm.
Corner from Feedpoint or centre 3,697 mm. 3,716 mm.
Gap between Driven and Reflector 112 mm. 113 mm.
Half of Driven 4,868 mm. 4,893 mm.
Reflector Total 10,287 mm. 10,338 mm.

A recalculation using 1.5 mm. uninsulated copper wire showed that the wire lengths need to be scaled by 1.5% - i.e. multiply lengths by 1.015.

It seems that the ETFE insulation reduces the velocity factor of the wire by 1.5%. I have modelled 1 mm. copper wire with a ¾ mm. PVC jacket and ETFE wire lengths must be scaled by -2% (multiply by 0.98), that is a velocity factor reduction of about -4% compared to plain copper.

Polythene insulation will be very similar to ETFE. I haven't modelled enamelled or transformer copper wires or aluminium conductors. The conductor material should not affect the length needed and the thin polyurethane or whatever on enamelled wire will also have a negligible effect.

Settings used

These are the settings I used. You will, of course, modify them as you explore the design but these are good starting points

Frequency
14.2 MHz.
Wires
6 wires with dimensions as above were positioned approximately 1.5 m. AGL. An EZNEC wire file is available for download (5) for those who need it for modelling with other programs.
Sources
1 source at centre of wire 2. 70.7 volts (100w in 50Ω). Wire 2 must have an odd number of segments, EZNEC positions the source at the centre of a segment.
Ground Type
Real/High Accuracy but see (6) below.
Ground Description
1 medium. This could be adjusted for local conditions. See (6) .
Wire Loss
Copper (1.74E-08). Silver is 1.59E-08 (7)
Units
Metres
Plot Type
3D. User choice.
Step Size
Ref Level
2.2 dBi (0 dBd).
Desc Options
In the "Desc Options" tab select the "Plot" tab then the "2D Misc" tab. Select a user selected outer ring of 10 dBref. Otherwise EZNEC will plot the far field with a zero dB level of the maximum gain. This will not enable you easily to judge the absolute gain of the antenna.

When generating SWR curves I used a scan from 14.1 to 14.35 MHz, with a step of 0.005 or 0.01 MHz. Bigger steps result in faster calculation.

Computed results

SWR plot
SWR in 50Ω

Notice that the maximum SWR is less than 1.5:1. This chart was not output by EZNEC directly but was constructed from an output data file using spreadsheet software. Plotting the SWR from 10 to 50 MHz showed the only other near resonance to be at about 46 MHz. This is not a harmonic of the operating frequency so no untoward radiation will occur. Strong Band I television signals may be received, but most HF rigs will be well filtered against them.

Azimuth plot
Azimuth plot. Outer ring is 10 dBd.
Elevation  plot
Elevation plot. Outer ring is 10 dBd.

This design produces a forward gain of 7.9 dBd at 6.4° elevation (vertical beamwidth 25.4°) and a 19.5 dB front to back ratio. The horizontal beamwidth is a remarkable 133° at -3dB (4.6dBd). The SWR is less than 1.5:1 between 14.100 and 14.350 MHz. A small increase in size could move this LF for CW work, but the SWR at the ends of the band would rise to 1.7:1. This is still usable with most modern rigs. The scaling factor required is 1.005 or add 5mm per metre to the suggested dimensions.

EZNEC files used

20M Moxon EZNEC file

Materials and costing

2 x 10m fibreglass roach poles.
Mine were premium display poles from Sky Blue Leisure (8), but there are many other suppliers. Some, expensive ones, available from fishing tackle suppliers are carbon fibre and are not suitable (they conduct).
About 25m of 19/02 silverplated, ETFE (9) insulated wire.
W.H. Westlake, (10) but now out of stock. Other wire may be substituted with recalculation if a different insulator or diameter is used. Wire loss is insignificant for copper, aluminium, and silver conductors and will have no noticeable effect. The insulation material has a significant effect on wire length (see above).
2m length of 20x20x4 mild steel angle.
B&Q (11) or engineering supplier.
~75m of 2mm kernmantle cord or substitute for guying.
Mammut from Taunton Leisure (12) , but you could use any other light weight line. Stout nylon monofilament may work but is springy and may cause tangles during erection.
4 x 60mm Jubilee clips
Stainless is best.
4 x 200mm nail type pegs.
Mine came from Marcruss (13) in Bristol, but any outdoor or camping shop will have suitable pegs.
Feedpoint Balun
Toroidal core, 1m RG178 coax, scrap Perspex or Lexan (125x25x4 or similar), panel mount female BNC connector, polystyreene cement.
Sundry
Various cable ties, small pieces of scrap 2 or 3mm thick aluminium. A short length of stout (30lb plus) nylon monofilament. Ask a fisherman or a tackle shop.

Construction

Cutting the wires

Three wires need to be cut, one for the reflector and an identical pair for the dipole. Use the lengths given above or derived from your own calculations. Allow a margin for the corner(s)(50mm.) and the end knot(s)(200mm.) on all three and the feedpoint attachment on each leg (100mm.) of the dipole. Cut long and work from the centre of each element, constructing the feedpoints, then the corners. Final adjustment of the four horizontal legs is then made at the knots.

The dipole centre

The dipole centre or feedpoint is constructed on an off-cut of 4 mm Perspex. A coaxial balun on a dust iron toroid is used to force the antenna into balance and prevent the feeder from radiating.

The dipole centre
Balun

The dipole centre
Feedpoint construction

Wind the balun onto the core using 12 or so turns. Cross over the winding at the half-way point. Secure the winding with polystyrene cement.

The Perspex drilling
Perspex drilling pattern

Using two pieces of Perspex about 120 x 60mm, drill each in the pattern shown. None of the dimensions is at all critical. The six holes that secure the wire should be large enough to pass a loop of wire through. 7 mm. was used in my case. It is very difficult to drill the holes which cross the edge after cutting off the wings. Use water as a cutting lubricant and a fairly low speed to avoid melting the plastic and causing the drill to 'catch'. Use a countersink to chamfer the edges of these holes.

The large centre hole should be clearance on the bulkhead BNC connector. The four small holes on the centre line should be large enough to pass the cable ties which secure the toroid. Positioning depends on core size. The fifth small hole should be large enough to pass the rg178 through and positioned conveniently.

Now cut off the 'wings' - the two outer strips shown in plain grey. Use a fine toothed hacksaw and lubricate with water. Now smooth and round off all the edges with a fine file and/or abrasive paper. Brasso can be used if a 'glass clear' finish is desired.

Attach the BNC connector and the core to the perspex and make off the input end of the RG178 to the connector. Pass the other end throught the small hole. Strip the end and attach the two legs of the dipole. Pass a loop of each wire through the centre hole at its end and loop it over the end of the centre piece. Tighten the loop, leaving some slack at the inboard end to prevent strain. I then coated all the exposed wires with Liquid Tape (14) .

Making the corners

Corner of the loop
Corner fixing and attachment

It is very difficult to tie a secure knot onto the ETFE wire, due to its slippery nature.

My solution is to form a small loop in the wire and fix it with a small (2.5x60 mm) cable tie. This has proved to be effective and reliable in use. It has kept the corners correctly positioned over an extended period.

This method should work equally well for any other flexible wire.

Joining the ends of the dipole and reflector

A bowline is tied in the end of each wire element. ETFE insulated wire is very slippery so twists and other knots may not hold.

Joining the wires
Joining the wires

The two wires are then connected with a length of stout nylon monofilament (50 lb in this case), tied with the common 4 ½ turn knot used to attach hooks etc. Loop the nylon through the wire, take the free end around the standing nylon 3 to 4 ½ times then thread the free end back between the two nylon lines close to the wire. Lubricate (with spit) then tighten the knot quickly by pulling the standing line and the wire sharply.

Here the ends of the wires have been sealed with Plasti-Dip International Performix "Liquid Tape" (14) ; to prevent water ingress. I'm not sure how well this will seal to the ETFE, more later.

The top pieces

The top piece

Two top pieces are needed. Each is made from an off-cut of three millimetre aluminium sheet. There are three holes drilled. The centre hole should be just large enough to fit over the base (largest end) of the top section of the pole so that when lowered onto the pole the top piece rests on the top of the second to top section. This size is variable depending on the source of the poles. The other two holes allow guys and the halyard to be attached. All the holes should be chamfered with a countersink or a much larger drill. The edges of the top piece must also be chamfered. Try to achieve a smooth finish all over to reduce chafing of the lines.

The pole supports

The angle iron is cut into two equal 1m pieces to act as pole base supports.

Angle iron supports

Guys and halyard

Cut and make off four guys about 13 m. long. This length will allow the guys to be at 45° to the pole. Cut a halyard at least 20 and better 25 m. long.

Erection

Tools used

Lump hammer, spirit level, compass, screwdriver, side cutters.

Extend the poles

Work from the top down, pulling each joint tight with a slight twisting motion. Tape the joints with PVC insulating or self amalgamating tape for added security.

Drive in the base supports

Using the level
Setting vertical
Lined up
Lining up

These supports are the main means of keeping the poles upright, so they need to be firm. They also need to be vertical. Hold the support loosely at the top and allow it to act as a plumb bob before driving. Check with the spirit level after the first couple of inches and correct if needed.

Drive one of the base support angle irons to 300 or 400 mm. You can vary this depth to suit the ground conditions.

Use the compass to align the second support. Sight the first support from a couple of metres. The choice of alignment is not at all critical, in view of the wide beamwidth. I used 315° to cover the Carribean and North America, with a chance of JA too. Place the compass on the ground, then go to the second support and line up the compass and the first support.

The distance between the supports should be 3 - 3.5m.

Lay out and attach the aerial wire and guys

Lay out the antenna wire alongside the poles.

Wires and poles laid out.
Wires and poles laid out.
Fixed Top Piece
Top Piece
Feeder attached
Feeder attached

Attach the top corner of the driven element at the top of the pole using one of the aluminium top pieces. Use a short piece of the guying line. This is the pole which will be mounted on the support in the direction of radiation. Remember that the anttena fires towards the feedpoint.

Next attach the feeder to the feedpoint balun. You may wish to seal the connector with self-amalgamating tape.

Now attach the feeder to the pole pulling the upper leg of the dipole fairly taut to guage position. I used two releasable cable ties to fix the coax.

Now fix the bottom corner. I used a short length of line and another cable tie around the pole to position it. It should be about 1.5m from the pole base.

Results

Erected Antenna
The erected antenna

During a short, fairly relaxed, session at 15:30 - 17:00 on 14/SEP/2009 I worked PJ4, YB, and several US stations as far west as Las Vegas. I was pleased that I got heard in pile-ups with relatively little difficulty despite a maximum power of less than 100W. A report of "peaking 59+15" was recieved from one US station.

For a more permanent installation, use 2 inch fibreglass poles from Engineered Composites (15) ; and guy well. Each pole should be quite rigid before the antenna is erected. The additional cost of four 6m poles will be well justified by better wind resistance. Scaled rectangles for 15 and ten or WARC bands could be added.

get the support lines from the poles equal if possible.

Notes, links and references

(1) L.A. Moxon, G6XN "HF Antennas for all locations" p 167. RSGB, 1984. ISBN 0900612576

(2) Pete Millis's EZNEC file

(3) Pete, M3KXZ's website http://www.outsideshack.com/

(4) by Roy W. Lewallen, W7EL. A free version, only limited by number of segments, and therefore complexity of antenna, is available for download from http://www.eznec.com/

(5) 20M Moxon wires file (.txt) The wires are, of course embedded in the downloadable EZNEC model files, so this file will not be needed if you choose one of those. If you plan to use other software then the wires file may be useful. It can be simply edited (delete the comment lines) into a csv file.

(6) See this discussion and values of parameters for ground modelling in EZNEC. Soil Dielectric Constant values .

(7) Dr. D.W. Knight, G3YNH, has extensive data on the properties of conductors at http://www.g3ynh.info/zdocs/comps/part_1.html . I strongly recommend his excellent and very thorough work " From Transmitter to Antenna ", on impedance, impedance matching and high-frequency power transmission as essential reading for anyone who wishes to understand the subject.

(8) http://www.skyblueleisure.co.uk

(9) ETFE Ethylene tetra-fluoro ethylene is a slightly stronger and more resistant alternative to PTFE, used to jacket insulated wire and cable. It has excellent electrical properties and has good enough thermal resistance to be more or less immune to soldering irons. It is the plastic used for the Eden Project domes.

(10) W.H. Westlake Electronics
'West Park', Clawton, HOLSWORTHY, Devon, EX22 6QN.
Tel: +44 (0)1409-253758 Mobile: 07774 635077 Fax: +44 (0)1409-253458

(11) B&Q

(12) Taunton Leisure

(13) Marcruss

(14) available from SOTA Beams.

(15) Engineered Composites Ltd .,
Unit B4, Borders 2 Industrial Park, River Lane, Saltney, Chester, Cheshire. CH4 8RJ
PHONE: 01244 676 000, FAX: 01244 677 267
E-mail: info@engineered-composites.co.uk

© Andy Cowley, 2005-2012     Contact     Home

Valid XHTML 1.1! Valid CSS! I heart Validator Wikipedia Affiliate Button

This page last modified - Tuesday, 16-Mar-2010 23:02:24 GMT, rendered - Wednesday, 22-Feb-2012 22:42:27 GMT

for (none) 38.107.179.237 by www.andycowley.net