• [Introduction]
• [Design]
• [Construction & Installation]
• [Measurements]
• [Feedline and tuner/coupler]
• [Rotor/rotator]
• [Modeling]
• [References]

©2018-2021 F. Dörenberg, unless stated otherwise. All rights reserved worldwide. No part of this publication may be used without permission from the author.

Latest page update: June 2021 (included length of feedline to antenna tuner, added photo 12B & text)

Previous updates: January 2020 (inserted Fig. 17 and Fig. 21 + text)

INTRODUCTION

To be covered:

• Background, my local conditions (previous dipole + height; max span on terrace; < 70% of full size ( about 2x3.5 m) --> noticeably less efficient; ability to cover 30-80 or 30-160), DK2DB, DF7XH. Note: at my current QTH (the south of France), French/European law applies, including "right to an antenna". I live in an apartment building, with a "home owners association" (HOA). In accordance with this law, I do not need permission from the HOA to put up an antenna - independent of the HOA rules & regulations! The law takes precedence over HOA R&Rs. However, for a permanent antenna ( = installed for more than 3 months), I am obliged to notify the HOA.
• 4NEC2 modeling, analysis
• model (downloadable); tapering
• 20-10m or 20-6
• horizontal/vertical radiation pattern for each band and for height 2/4/6m, poor/average/good ground
• 3D radiation pattern, same
• SWR sweep, impedance sweep, impedance at select freq in each band
• Construction
• parts list (tubing, self-tapping screws, feedline connection, DIN3015-1 clamps, center insulator, mast stub, mast clamps)
• tapering decision
• tooling, drilling tube ID, sanding down OD
• Rotor, incl. adapter stub
• Feeding, antenna tuner/coupler/matching
• losses
• Measurements:
• miniVNA sweeps at height 2/4/6m
•  impedance at select freq in each band
• radiation pattern with Web-SDR or Web-KiWi
• Suitability for 6m (50 Mz): not tested at my QTH (in France. depends on dept)

DESIGN

My local constraints: temporary installation, hand-crank mast; --> light weight, manageable for 1 person. I cover 40/80 with my mag loop and am not interested in the 30m band.

Dipole antenna systems become noticeably less efficient when the span is less than about 70-75% of a standard half wavelength dipole. In other words: for a tip-to-tip span less than ≈0.35 λ. Less than that, and losses in the feed line and in the impedance matching unit ("tuner", coupler) become significant.

At 14 MHz (20 m band), this is equivalent to a span of 7.5 m (≈25 ft). So I want my dipole to have close to the maximum span that I can install. Given the size of my terrace and the placement of my antenna mast, I am limited to a dipole with a maximum leg length of 3.4 m (≈11 ft). I.e., a tip-to-tip span of 6.8 m (≈22.3 ft) or 65% of ½ λ at 14 MHz. It will (have to) do!

This dipole will have some directivity  in the 20 m band, and (much) more so in the 17, 15, 12, 10, and 6 m bands.

Minimize wind load (incl. stress on rotor/rotator)

Examples: DK2DB commercial (no longer in production) incl. remote symmetrical tuner near the feed point, seen at Ham Radio many years ago, also advertized. DF7XH clone with improved tuner (continuously adjustable vs 5-band switch of DK).

 

Fig. 1: Evaluation of a dozen tapered/telescopic aluminum tubing configurations with 4 or 5 elements per dipole leg

(Excel spreadsheet: see ref. 3)

I converged on config nr. 3 (which happens to be the same as DF7XH). It has a 2x3.4m=6.8m (22.3 ft) span. Each dipole leg comprises 5 aluminum tube sections, with decreasing diameter: 20-16-12-10-8 mm. All 12 configurations have roughly the same total cost for the aluminum tubing.

CONSTRUCTION AND INSTALLATION

The legs of my dipole are made of aluminum tubing. I used a tubing made of the commonly available aluminium-magnesium-silicon alloy (Al-Mg-Si) type 6060 T-?. This is quite similar to alloy type 6063. Instead of using a single tube section for each leg, The legs are tapered towards the tip, by using five sections of tubing with decreasing diameter. The section lengths and diameters are as derived in the "design" section above. From feedpoint to tip:

• 50 cm of tubing with an outer diameter (OD) of 20 mm and an inner diameter (ID) of 16 mm, so the wall thickness is 2 mm; i.e., 20x2 mm tubing.
• 90 cm of tubing with an OD of 16 mm and an ID of 12 mm; i.e., 16x2 mm tubing.
• 90 cm of tubing with an OD of 12 mm and an ID of 10 mm; i.e., 12x1 mm tubing.
• 90 cm of tubing with an OD of 20 mm and an ID of 8 mm; i.e., 10x1 mm tubing.
• 60 cm of tubing with an OD of 8 mm.

Fig. 2: The aluminum tubing - lengths per configuration nr. 3 in the table above

Fig. 3: Staggered diameters of the telescopic aluminum tubing sections

Each section (other than the 20 mm OD section) is inserted in the next-size-up section: the 8 mm OD section into the 10 mm OD section ( = 8 mm ID), 10 mm OD into 12 mm OD ( = 10 mm ID), and 12 mm OD into 16 mm OD ( = 12 mm ID), and 16mm OD into 20 mm OD ( = 16 mm ID). Insertion depth is about 10 cm (4 inch).

Note: unlike a pipe, a tube is defined by its outside diameter. The ID of a tube is not manufactured as accurately as its OD. Typically, the tube ID is a bit smaller than the OD minus twice the nominal wall thickness. If you get your tubing from an amateur radio supply store, the ID's and OD's should be made accurately enough such that successive diameters are a "sliding fit". I got my tubing from a general tube & pipe supplier, and was not so lucky...

Once the tube sections have been inserted, we have to secure the overlaps in place. For the standard options - see Figure 6:

Fig. 6: Options for fixing overlapping tube sections

(source: ref. 3)

I decided to use "blind" pop rivets with the following dimensions:

• 2.4x6 mm for 1.5-3.5 mm grip range ( = total metal thickness = sum of the wall thicknesses of the overlapping tube sections),
• 2.4x8 mm for 3.5-5.5 mm grip range,
• 3x6 mm for 2.5-3.5 grip range,
• 3x8 mm for 3-6 mm grip range.

Of course, you will also need riveting pliers, and 2.4 and/or 3 mm metal drill.

Options considered for mounting the dipole onto the mounting plate:

• 2 (hydraulic) line clamps per dipole leg
• 1 line clamp per dipole leg + a solid insulating rod, inserted into each leg end at the feedpoint.

I decided to use two line clamps on each side: a total of four DIN3015-1 hydraulic line clamps (a.k.a. "Stauff clamps"), standard model, size 3, ribbed, for 20 mm OD tubing.

Options considered for connecting the 450 ohm "window" feedline to the end of the 20 mm OD tubes:

• stainless steel hose clamps (UK: jubilee clamps). They are asily adjustable/removable.
• machine screws + washers + lug + lock nut.

I decided to use hose clamps. Right before attaching the feedline wires, I filed the ends of the aluminum tubing to remove all oxide, and ensure good contact. I gleaned this connection method from my  friend Rolf (DF7XH). His rotary dipole has alraedy been up for several years, without any contact issues.

Additional installation material:

• 25x25x1 cm poly mounting plate (kitchen cutting board),
• 2x V-bolt (or U-bolt) mast clamp.

Fig. 7: Parts of my dipole mounting plate

Fig. 8: Front and back of my dipole mounting plate

The complete mounting plate as shown above weighs 900 gr (2 lbs). The aluminum tubing weighs about 1100 gr (per Fig. 1).

Fig. 9: Side view of the mounting plate and feedline connection at the dipole's feedpoint

I used a standard 150 cm (5 ft)  section of galvanized TV mast with a 40 mm OD and two TV mast chimney-mount brackets, with  galvanized steel straps around the chimney. All from the DIY-store. I used "double ratchet" brackets, as it is much easier to tighten the strap, if there is a ratchet on both ends of that strap.

Fig. 11: The chimney mounts and strap

I am not using coax feedline between the antenna and my automatic antenna tuner (ATU). Instead, I use 450 ohm "window" line - this is standard for a multiband antenna with an impedance matching "tuner" that is not installed directly at the feedpoint of the antenna. The windows significantly reduce windload, so the cable will not flop around as much. I have placed my automatic antenna tuner in the attic, close to where the cable passes between the roof tiles. Total length of the window line is 3.95 m (≈13 ft). This is about 1/10 λ on 40 m and 0.4 λ on 10 m. Of course, the cable is part of the complete antenna system, and it introduces a frequency-dependent impedance transformation between dipole and antenna tuner. B.t.w., this antenna system also turns out to generate a very decent signal on 40 m. Note: the ability of a tuner to match the impedance at its feedline input (i.e., the antenna side), says absolutely nothing about the losses inside the tuner - which may be very high!

Unlike coax, this type of feedline should not be attached directly to a metal mast. I could not find suitable spacers in the local do-it-yourself stores. So, I decided to improvise. I bought a section of dense foam tubing that is intended for insulating hot water / heating pipes. It has an outside diameter of about 6 cm (≈2.5 inch). Swimming pool "noodles" are another option. I cut the tube into a number of 5 cm (2 inch) thick slices. I used tiewraps to attached them as pairs. See photo below. I used longer, wider tiewraps to attach each pair to the mast. I then attached the 450 Ω window line to the outer foam rings, again with a tiewrap. See photo above. Note: always use black tiewraps: they will hold up in sunlight, whereas white tiewraps will not! Time will tell how well my spacers hold up in sun and wind...

Fig. 13: Spacers for attaching the 450 Ω window line to the mast

(I used rivet hole punch pliers for making attachment holes in the windows line - you can also use a simple awl)

Fig. 12A: Close-ups of the installation with the feedline spacers

Within a couple of months after installation, the antenna system was already exposed to several storms with 100 km/h gusts - no damage to any of the parts! Of course, the foam tubing is not intended for outdoor use. The photos below what the spacers look like after two years exposure to lots of sunshine (QTH in the south of France) and many (!) stormy days. Luckily, the foam tubing is quite inexpensive.

Fig. 12B: Condition of the foam spacers after two years exposure

In July of 2022, after another year of frequent strong winds and lots of sunshine, the foam had deteriorated further and I decided to replace all the foam parts.

There are three cables between the "antenna + tuner" system to my transceiver:

• Aircell-7 coax cable from the remote automatic tuner,
• 4-wire control cable to/from that tuner ,
• 3-wire control cable to the antenna rotor.

The original cables are all about 12 meters long (≈40 ft). Once I installed the antenna on the roof and the tuner in the attic, these cables were just long enough to exit the attic below the rain gutter on the terrace side of my apartment. To avoid rain water running down the cables into connectors, there are drip loops shortly after they exit the soffit:

Fig. 14: The rotary dipole, and the three cables exiting the attic via the soffit

(the coiled-up cables are the three extension cables)

To reach the transceiver in my shack, I made three extension cables. They have connectors on both sides, so I can disconnect these cables when not in use. So I don't have to worry (as much) about lightning protection. Note: my standard home-owner's insurance covers liabilty for (legal) hobby activities, including amateur radio!

The far end of the extension cables re-enters the apartment via a door to the right of the one shown in the photo above:

Fig. 15: The three cables re-enter the apartment - one cable passes through the cat trap door

The coax cable connects to a BNC jack on the wooden frame of the screen door. I use that jack also for other antennas, such as my Small Transmitting Loop antennas (a.k.a. "Magnetic Loops"). From the BNC jack, a coax cable passes via the PVC door jamb to the transceiver. The tuner control cable connects to an adapter cable that includes a section of ribbon cable. My automatic tuner used to be installed on the outside wall on the other side of the terrace.So I already had this adapter cable for years. The ribbon cable easily passes underneath the door. I did not have a rotor and rotor control cable before. That cable simply enters via the cat trap door.

My antenne is placed about 4 m (13 ft) from the four TV antennas of the apartment building: two satellite dishes and a mast with two UHF antennas. Vertically, my dipole height is half way between the latter two. Given the length of the dipole legs, the tips of my dipole pass within 80 cm (22 inches) of the nearest ( = lower) UHF antenna. There do not appear to be standard guidelines for calculating if this will cause problems. Obviously, I do not want to destroy all television sets and distribution amplifiers of my 27 neighbors in the building! I verified compatibility on 40-10m by transmitting a carrier signal of 10 watt on all bands, while simultaneously watching a satellite and a UHF TV channel. I then repeated this with and 80-100 watts. There were no signs of interference or damage. This test was done on a dry day. I may repeat the test on a rainy day, just in case...

Fig. 16: Placement of my rotary dipole and the four TV antennas

Fig. 17: My rotary dipole from a distance

ANTENNA ROTOR / ROTATOR

Obviously a "rotary dipole" must be rotated. So I needed a rotor (Europe + UK; "rotator" in the USA).

My antenna, including mounting hardware, weighs about 3 kg (≈ 6.6 lbs). This includes a 40 cm long POM/Delrin/acetal rod with 35 mm diameter as "top mast" stub. The rotor documentation only states a limit for the length of the top mast: 3 ft (≈ 90 cm), not for the antenna weight.

My requirements --> light "TV antenna" rotor will do. Decided to get a Channel Master 9521HD, which is an improved model 9521A. It sells for $145 in the USA (2018 pricing). It settings are easily changed between 50 Hz and 60 Hz AC supply.  Ref. datasheet etc. This rotor uses a 117 volt synchronous AC motor. If you rotate back & forth a lot, eventually the angular position indicated on the control box becomes inaccurate. A "synchronization" cycle must than be initiated on the control box or via the remote control. This only takes a minute or two. The unit can also be configured for "automatic sync" after every 35 movements.

As I have access to the entire attic, I routed the rotor control cable (20 m / 60 ft) below the roof tiles.

Fig. 18: Channel Master CM9521HD TV antenna rotor/rotator system

The lower clamps accommodate a mast diameter of up to 2 inch (5 cm), but the top clamps will only accommodate a 1.75 inch (4.5 cm) diameter mast stub (3 ft (90 cm) max).

The dipole has a symmetrical "figure 8" radiation pattern. This means that you only have to rotate the antenna over 180° to cover all directions. This is important, because the feedline  (450 Ω window line, in my case) wraps around the mast and rotor, when the antenna is turning. Limiting rotation to 180° reduces the feedline length above the rotor.

Fig. 19: Azimuthal map, centered on my QTH - to quickly determine antenna bearing direction

Map tool? e.g., https://ns6t.net/azimuth/azimuth.html

FEEDLINE AND ANTENNA TUNER / COUPLER / MATCHING UNIT

Options, needs (incl. expected impedance range to deal with), my choice(s), losses.

Multi-band operation --> not coax to ATU/TRX, but symmetrical feedline (window line, twin lead, or ladder line).

Abt 5 m / 15 ft of 450 Ω window line between the dipole feedpoint and the ATU. ATU installed on one of the rafters below the antenna. Then 18 m / 60 ft coax (I used Airflex 7) and ATU control cable to the shack. The cables leave the attic at the far end of the roof, and enter my shack via the terrace. Connectors, so I can disconnect these cables (and teh rotor control cable) when not in use. So I don't have to worry (as much) about lightning protection. Note: my standard home-owner's insurance covers liabilty for (legal) hobby activities, including amateur radio.

Feedline converts the feedpoint impedance! Ref. calculator.

The resonance frequency of this antenna is around xx MHz, so it is operated as a non-resonant antenna. Contrary to popular myth: a non-resonant antenna radiates just as well as a resonant antenna. The only (yes, the only!!!) advantage of a resonant antenna is that around its resonance frequencies, it typically has a feedpoint impedance that is easy to match to a coax cable. That's all!

MEASUREMENTS

First impression: noticeably quieter (in the sense of better signal-to-noise ratio, not deafness, hihi) than the twice-as-big (2x7m) non-rotary wire dipole that I used to have on my terrace, installed at a height of only 2.5m (8 ft) above the terrace floor.

The first radiation pattern that I determined was for the 10m band. It is a "receive" pattern, as I used a beacon transmitter somewhere in my area (I do not know the distance, but there was very little QSB on the signal). It transmits a full-time constant carrier on 28.1 MHz.The pattern shows a steep null of -9.3 dB. Note: I only measured for an azimuth of 0 - 180 degrees, as the antenna is symmetrical. In the plot below, the values for 185°-355° are simply copies of the values measured for the 0°-175° range. Step size was 5°, except around the null-direction. I measured the relative strength (dB) of my receiver audio with the excellent "Spectrum Lab" audio spectrum analyzer freeware by Wolfgang Büscher (DL4YHF). The pattern distortion around 70°-90° corresponds to the direction of the two satellite TV and the two UHF TV antennas (see Fig. 15 above).

Fig. 20: Measured relative signal strength of a nearby constant-carrier beacon on 28.1 MHz

(pattern measured on 3-Aug-2019, 18:30 UTC)

The next plot shows relative strength of two carrier pulses sent by me in the 20m band, and received at a Web-SDR at a distance of 1100 km. The second pulse is with the dipole pointing at that receiver, the first pulse in a direction 90° away. The difference is about 10 dB, which is consistent with the polar plot above.

Fig. 21: Measured relative signal strength of carrier transmitted by me on 14068 kHz

(signal measured on 4-Jan-2020, 11:30 UTC)

The next image shows the impedance and SWR of my antenna, measured directly at the feedpoint, with the antenna installed on the chimney and pointing in three directions:

Fig. 22: Sweeps of my miniVNA antenna analyzer with antenna pointing in 3 directions

(the above image is a dynamic gif with 3 frames that change every 3 sec)

miniVNA antenna analyzer sweeps (at feedpoint, after feedline; 20-10 / 20-6, zoom per band, various heights.

Photos of installation, incl. w miniVNA via Bluetooth/WiFi.

Fig. X: Measured feed point impedance sweeps for h = 2.5, 4, and 6 meters above my terrace floor

Fig. X: Measured feed point impedances for h = 2.5, 4, and 6 meters above my terrace floor

Compare to calculated/modeled sweeps and table. Explain significant differences, if any.

Short feedline to auto ATU near base of mast, symmetrical ATU in shack.

Need common mode choke?

Feedpoint impedance range to be accommodated at the various heights + freq bands (simulated/modeled vs measured)

Resulting impedance ranges at the output of the feedline = impedance ranges at the ATU input.

ATU losses for these impedance ranges.

Antenna currents.

Horizontal pattern via WebSDR/KiwiSDR or with local beacon.

Suitability for RX/TX on 6m /50 MHz (if EDX-2 goes upthat high)

SOME MODELING

First, let's have a look at the radiation pattern of a dipole antenna in "free space". That is, far away from all objects and materials that can couple with the dipole or that can otherwise affect the generation and propagation of the antenna's electro-magnetic radiation. This 3D pattern has a doughnut shape, with the the dipole poking through the doughnut hole. All cross-sections that include the entire dipole, look like a figure-8.

The above image was generated with the excellent 4NEC2 freeware simulation package. The model of my tapered dipole is provided as ref. 2. Interesting and beautifully colorful as the above 3D pattern may be (especially if you like doughnuts, hihi), it is not the pattern of an HF dipole antenna in a realistic installation on earth, even if the earth would be homogenous and without any objects anywhere near the antenne. The images below show the 3D radiation pattern of the same horizontal dipole, for frequencies in the 20, 17, 15, 12, 10, and 6 meter bands, and at installation heights of 2.5, 4, 6, and 10 meters (8, 13, 20, and 33 ft). Ground type is assumed to be "moderate" (conductivity and dielectric constant). They were generated with the same 4NEC2 model as above.

I selected these particular installation heights, as they represent the (low!) height of the 2x6.5 m wire general purpose dipole that I had for many years at the corner of my terrace, the 6 m hand-crank mast that I recently acquired, a height in between the latter two, and the 10 m height of the next-size-up telescopic mast.

Fig. 23: 2x3.4m dipole installed at height of 2.5 m (≈8 ft) over moderate ground

(3D radiation patterns for the 20-6m bands)

Fig. 24: 2x3.4m dipole installed at height of 4 m (≈13 ft) over moderate ground

Fig. 25: 2x3.4m dipole installed at height of 6 m (≈20 ft) over moderate ground

Fig. 26: 2x3.4m dipole installed at height of 10 m (≈33 ft) over moderate ground

Fig. 27: 2D patterns for 14.1, 18.1, 21.1, 24.9, 28.5, and 51 MHz - each at a height of 2.5, 4, 6, and 10 m

(non-normalized view - full size zoomable view is here; click here for normalized view; note: different gain scales above 21 MHz)

We can draw some general conclusions from the above patterns (which apply in general to all horizontal dipoles):

• For installation heights below 1/4 λ, most of the radiation is upward. The 3D "blob" is only pink at the very top, and has minimal horizontal directivity in the two directions that are broadside to the dipole. This is a "cloud warmer" antenna, good for local and regional communication. But I am interested in DX....
• At around 1/4 λ, the 3D "blob" begins to flatten and horizontal directivity becomes more pronounced. The oval upward pink area of the 3D "blob" begins to split into two distinct pink areas. Both are broadside to the dipole, and have a relatively high take-off angle.
• At around 0.5 λ, the blob resembles a flattened "free space" blob.
• For height above 0.5 λ, an additional lobe begins to appear in the vertical direction. This lobe becomes more pronounced as height increases.
• Around 0.8 λ, this upward lobe splits into two high-angle broadside lobes. This angle decreases as height increases. There are sharp nulls between these upper lobes, and the lobes below them. Also, small side-lobes appear, in the direction of the two dipole legs.
• Above 1 λ, 1.5 λ, 2 λ, etc., additional upward/broadside lobe-pairs appear, and all the those pairs become flatter as height increases further.
• The number of lobes in each of the two broadside directions ≈ 2h / λ.
• 0.3 λ really is about the minimum installation height for obtaining a reasonably low take-off angle. I have a 6 m (20 ft) hand-crank mast. At full mast extension, the top is at a height of around 0.3 λ for frequencies in the 20 m band, and - of course - more than that at higher frequencies. Mounted on my rooftop, the antenna is also about 6m above the concrete floor in my apartment.

The next figure shows that, for installation heights below 0.5 λ, the "quality" of the ground does not seem to impact the shape of the 3D radiation pattern a lot. However, when soil conditions cause high ground losses, effectively less power is radiated. So, ground type does impact the overall size of the blob.

Fig. 28: Impact of ground "quality" (poor/moderate/good) on the radiation pattern at 14.1 MHz for various installation heights

Fig. 29: Impact of installation height on feed point impedance (h = 2.5, 4, 6, and 10m; average ground) - 4NEC2 simulation

(the pink curves show primary resonance around 21 MHz; gray vertical bands per IARU Region 2 / USA bandplan)

To make it easier to determine what the calculated feed point impedances are (and have to be accommodated via feed line plus tuner/coupler/matching unit at that feed point or downstream), here in tabulated form (incl. 80 and 40m).

Fig. 30: Calculated feed point impedances for h = 4, 6, and 10 meters; average ground

(based on the 4NEC2 simulations shown above)

REFERENCES

• Ref. 1A: "Rotary dipole for the shortwave bands from 10 to 20m", 2-page product sheet from www.id-elektronik.de
• Ref. 1B: "Rotary Dipol für die Kurzwellenbänder 10 bis 20m"
• Ref. 2: 4NEC2 model of the short 20-10m dipole antenna, by Frank Dörenberg (N4SPP), November 2018
• Ref. 3: Figure 11 in "Antenna Materials and Accessories", chapter 20 in "The ARRL Antenna Book", 21^st edition, 2007