all-antennas-2

IV3PRK Pierluigi “Luis” Mansutti

160 Meters: DXing on the Edge

All my 160 m. antennas, through the years - part 2.

SUMMARY: my first Waller Flag - a 2-wire Beverage - new 4-square receiving array - the Hi-Z 4-square array - a Beverage on ground - a 3-wire mini loop -

- the TX3A DHDL - doubling the DHDL - DHDLs in broadside configuration.

2007: upgrading the original rotatable Flag to the Waller Flag.

Why it did NOT work: all the mistakes of my first attempt to be avoided!

After hearing directly how well Jose Carlos N4IS was working DX stations on 160 m. also in the summer season, and knowing the success achieved in a few years by Doug NX4D himself, I was attracted by their receiving antenna. It is a rotatable end-fire close array of two Flags, originated by Doug Waller NX4D and thus named “The Waller Flag” by Jose Carlos N4IS, who built a couple of them with some improvements. All the electrical and mechanical details are on his web page With over 11.5 dB of RDF, it can be compared to a rotatable broadside array of two Beverages. But, as usual, there is no free lunch with antennas and its drawback is a very low signal level, around “ - 55 dB”, which is about 40 dB lower than the Beverages. Hence, not only one, but two good preamplifiers are needed and be very careful against “common noise”. Very weak signals of -55 dB are extremely vulnerable and are easily overwhelmed by this enemy flowing over the coaxial cable shield and entering through gaps in the system.

We will see which were the gaps - i.e. the mistakes made in this first attempt - corrected and not repeated ten years later with my second Waller Flag, which is has been very successful, and still working fine!

Basically, the original W7IUV Flag was a single rectangular loop, that I was going to replace with two smaller loops (same vertical wires length, but much shorter the horizontal ones) fed, with end-fire phasing on the same boom, as shown in the Eznec “Imagine View” below.

I did a huge modeling work on EZNEC, all reported in PDF file, dated August 2007

followed by a second one, in Sepetember, with details on "Tx antenna detuning”

After detuning and made the tower invisible, I got these elevation and horizontal patterns, reported here, side by side, in order to see the difference between the two flag types.

W7IUV Flag Waller Flag

The take-off angle is lower, and the lobe is much sharper in the Waller Flag, favouring the DX signals. In my personal situation, it should be a great benefit in reducing the Russian and eastern European QRM while working Far East and the Pacific on the same direction. Unfortunately, compared to the single Flag, there is a high angle back lobe, caused by the close spacing of the two loops, and almost impossible to reduce.

The azimuth pattern of the Waller Flag is also very good, only 83 degrees of beamwidth, better than a typical four-square array but, as stated previously, there is a drawback which requires attention: the gain is 55 dB “negative”, thus 25 dB lower than on a single Flag, and we need two clean preamplifiers after being very careful with transformers.

It's worth taking a look at the two graphs on the right. In the F/B graph there are no major differences, apart from the peak in the 330° direction, with the tower behind the flag.

Where we find a great marked advantage is in the RDF, with a wide stable difference of over 3.5 dB: awesome! RDF is the most important and direct parameter used in ranking receiving antennas.

On the Tom’s site: we see that a 1.75 wl Beverage has an RDF of 11.16 dB.

The Waller Flag is above that, and can be rotated.

The construction…

In the second half of September 2007, a big storm broke a fiberglass spreader of the existing rotatable Flag, so I had to take it down and hurry up with the new Waller Flag construction. But, with a last check on the Eznec model, an error arose with the wires diameters, which I never changed from the original #12. As I was going to use two of my old 10 meters log-yagi elements (split and insulated from the boom) as vertical wires for the loops, I corrected their diameter to that of the aluminium tubes. My God:. the pattern resulted destroyed... unbelievable!!

So, I went back to the #12 wire (actually better, I changed to #14) and I modeled the two vertical insulated aluminium elements (in two halves as they are) at a distance of 30 cm. from the vertical wire they are supporting. Everything went in order again, with the lobe and the FB as before! A little more complicated in the construction, but quite easy to do.

… and the disappointing results!

The Waller Flag did not work at all! Despite all the emails exchanged with Jose Carlos and Doug, who did their best to assist and encourage me, the answer received from Larry Molitor W7IUV, to whom I asked information on how to handle a gain of -55 dB, put the word "Finish" to my first attempt at WF, and after a month I decided to return to the previous flag.

All buildindg details, results and comments are on this PDF file:

So, I immediately rebuild the W7IUV flag as it was before, with the same dimensions. All technical details and info are on Larry's website and I put mine on another short PDF in november 2007:

But what was wrong with my Waller Flag attempt?

A wide explanation came from Jose Carlos, N4IS, on March 2015, when I was in Ecuador and JC was trying to convince me to install there the Waller Flag. He wrote:

09/03/2015, Hi Luis!

I have a deep admiration and great consideration for you, I understand your passion for 160m.

Let me tell you few observations as constructive comments.

Few years after you try the WF, I went to your site to see the DHDL and I saw few pictures that helped me to understand why your WF attempt failed.

1- The feed line from the WF was running on the air, together with a lot of other cable. Common mode noise transfer from one cable to another because it runs outside the shield. Several cable together end up having the same current outside the shield.

2- The feed line MUST be grounded at the base of the tower, and at the entrance of the house; after that you can run the cable to the second floor.

3- Just few days ago I measured 30 dBm increase in common mode noise removing the ground of my feed line at the base of the tower.

4- I solation of the boom, not sure it was isolated.

5- Most important. You built a signal switch with open frame relays. When you use a high gain preamp 20db or more, any wire without shield 1 cm long can capture several dBm in noise from another wire 50 mm apart. For RX antennas switch, all relays must be RF relay with high isolation from the poles to the coil of the relay.

6- You detuned your tower but not detuned the elevated radials that were resonant as well.

I am saying the points above to motivate you to try a WF again, Just the same frame you have can support 2 loops, the way it is today. Use twisted pair from Ethernet cable, just one pair. Choke the feed like and ground it twice before enter the house.

You will be able to dig signal 6 dB below noise comparing with one single loop you have today.

Just my two cents, 73

JC N4IS

That was all rigth! The boom was not isolated, the mast was not isolated and the tower was not isolated: all metal stuff tied together. And worst of all, the coax cable was not going down the tower to the earth (it was not possible because on a terrace) where should had been installed a common mode choke with a ground rod, but was driven horizontally for twenty meters over the roof, with the rotator control cable, and then entering the house in the TV cables conduit, down two floors to reach the shack.

No common mode chokes for sure… let's imagine how much noise was picked up by horizontal cables ove the roof and what happendd to a -55 dB RF signal!

Well learned the lesson, all these mistakes have been corrected, and ten years later I built successfully a second Waller Flag - probably the last antenna in my life, as it still works fine - and fully documented in two dedicated pages from the home page of this site:

2008: a two wire Beverage.

At the end of 2007 I acquired, in order to build my son's house, a new piece of land - 3.300 sq. meters - bordering with mine; but, my son is still leaving in Ecuador, and so that space was available for further antennas. At first, I thought at the possibility to try again some kind of Beverage, despite it was elongating towards South, not my most needed direction.

Some years ago I was used to have, in the winter season, a 175 m. Beverages stretched outside of my property to the NW direction, and it was for sure better than any other receiving antenna but, along a public road, it was a nuisance. After reading this excellent article his two-wire Beverage,

and going back to books and an old 1997 article by Tom Rauch, W8JI,

I decided to try also such an antenna, to lay in the new available space, far enough from the Tx antenna and its elevated radials. Unfortunately, it could cover only North and South directions, not the best I was aiming at.

Basically, my reversible antenna is made with a pair of twisted wires derived from a telephone cable 129 m. long, stretched like a classic Beverage, two meters high and gradually going down to the termination grounding rods. The impedance of the twisted wires, as a transmission line, should be about 150 ohms and, as a Beverage antenna, around 450 ohms. I used the same number of turns of OH2BEN for all the transformers, and wound them on the binoculars BN73-202. These are the pictures of my matching/switch boxes, and one installed at the feeding point, where a buried coax cable arrives.

As usual, I put my circuit diagram in a PDF file, with all details, and the reports and comments about on air tests. Click here

The antenna was working and switching correctly but, on the air tests, while in the reversed direction - to the North - the performance was very good, even better than the Pennants, in the forward direction - to the South - the desired signals were covered by very high noise.

Again, I tried everything, also walking around with my MFJ 852 - Line Noise Meter - modified to match a ligth portable 3 el. yagi to search offending sources, but all in vain.

Later, following the discussions about my noise with my friend Gary, KD9SV, the culprit was found and the "noise mystery" solved, as reported in this short article:

But that was not all: something else was wrong! A few years later, while designing my Reversible BOG's in Ecuador, I have been aware on a mistake made in the schematics of this 2-wire Beverage. And it was not my fault, but originated by an error in the 3rd edition of Low Band DXing (1999) and amended in the 5th edition (2010). The primary windings of the transformers T1 and T4 must NOT have a common point, as - per mistake - done in the left circuit by connecting togheter the shields of the feedlines (or in my switching circuit). And must be provided a serious Common Mode Choke for each cable.

ON4UN's Low-Band DXing ON4UN's Low-Band DXing

3rd Edition - 1999 5th Edition - 2010

Fig. 7-23 Fig. 7-115

2008: the new 4-square receiving array.

After long studies and modelling with Flags and Pennants phasing, followed by not so satisfying results in the real world, I decided to go back to the phased verticals Rx arrays. In 1994 I had built a very difficult one: the 4-square mini-phased array by K9UWA-KD9SV-W7EL design and that has been my best receiving antenna for almost ten years. But it was very critical, due to the high Q of the vertical dipole elements (inductance loaded, rather than resistance), too much WX dependent, and required a lot of maintenance with inductors rebuilding and toroids substitutions. I think it was the only one in the world still in use, and even one of the authors, KD9SV, suggested me to update to the new concept of verticals phasing developed by Tom, W8JI.

From the original W8JI design, I modified with Eznec+5 the basic element in order to use all the aluminium tubes of my old 4-square array.

I put also the element tips above the top hat radials, reaching a total high of 9.5 meters, to get more gain (sensitivity) and trying to further decrease the inductive part of load for the lowest Q of the antenna. Element diameter is tapering from 35 to 25 and to 20 mm. with the tips being of 12 mm. tube.

Top hat wires are 7.5 meter long and come down at 2.25 m. level. The loads are 73 ohms of R +296 ohms XL, thus a low Q of 4 which allows a low SWR for a stable phasing on the narrow frequency band of my interest.

After a long study with Eznec, I came to the final, very elaborated, design for 8 directions switching. By means of end-fire/broadside feeding (left pattern) we get the lobes at 0 - 90 - 180 - 270 degrees, while using crossfire feeding (rigth pattern) we get le lobes at 45 - 135 -225 - 315 degrees. All EZNEC images, patterns and tables have been put on this PDF document:

In this design, the separation between the 9.5 m. high elements was 21 meters (29.7 m. diagonal) and could just fit on the new lot, 27 meters wide, but in a dangerous situation, however, due to the risk of too many obstacles nearby. On the left: a 2 meters high fence and the 220 V power line. In front: a 1.50 meters high fence and a crossing telephone cable, 5 meters high On the right: the main telephone line (and a few months later will be added the public lighting on that road).

On the Google map below, I added all my 160 m. antennas in the month of March 2008, plus the projected new 4-square array.

After choosing the most suitable coax cable (Micro Tek H-325 AL-PE) 75 ohm with a VF of 81%, at first, as recommended, I measured the velocity factor on a quarter wave line with the AEA CIA analyzer and I found it something different from the published data. Thus, I calculated the phisical lengths of all phasing lines and reached the final real design - with 20 meters of separation between the elements - which gave these meaningful results: Gain Take off angle -3dB BWdh Front to Back RDF

1) End-fire/broadside feeding: -15.48 dB 22° 115° 13 dB 9.89 dB

2) Crossfire feeding -23.51 dB 20° 74° 29 dB 12.02 dB

Awesome the RDF in crossfire, while a wider coverage in the first configuration.

All plots, tables, schematics and construction details on this second PDF file,

from which are taken the following pictures on the preapairing work.

After completing the central switch/combiner box - using 14 Tyco/OEG printed circuit board relays, taking all attentions for grounding and common mode isolation, as well as on separate windings of binocular transformers -

the system was fully tested in the shack with all the phasing lines connected and switchable to the four elements, closed with a 75-ohm resistor to simulate the antenna loads.

Then, at each el. base, in these boxes ===> must be added resistors to reach 75 ohm, and null out the capacitative reactance by means of homebrew variable inductors of abt. 25 µH.

The matching boxes are fixed on the wooden antenna supports, wrapped in the buried section (1 m.) with a soft copper sheet for better grounding. (See upper right picture).

And ground connections were the main issue. After many e-mail exchanges and deep discussions with my helpful and expert friends Lee, K7TJR, and Jerry, K4SAV, I followed Lee’s advice. All provisions have been taken to avoid ground loops and unwanted signal or noise incursion between radials and feedlines, and to get also the best possible lightning protection.

As seen in the next ground system drawing, there are three separate grounds:

1) the verticals have their own ground system and are isolated from the coax lines and from the central box;

2) the feedlines to the elements and the phasing lines are grounded only at the combiner central box through a separate ground rod;

3) the RG213 feedline to the shack is isolated from the combiner and has another separate ground rod outside of the array.

So, the antennas are isolated from the central box, and the central box is isolated from the feedline going to the shack. Jerry says that item 2 or 3 are unnecessary but, with my hostile environment and all the constrains of my lot, I didn’t want to leave anything unattempted! This is the schematics of the ground system:

It was easy to null out reactance and achieve the required R 75 + jX 0 on three verticals, but … strange and difficult on antenna n. 4 (photo below), which is too close to a single telephone cable crossing my lot to reach a neighbor depot.

Anyway, I managed to get rid of it, and I made the first “on air” tests. The next morning, tryiing to listen to USA, I registered on Orion S-meter the following noise readings: S 1/2 on the NW pennant, S 3 on the rotatable flag, and S 9 on the new 4-square array !

Then, I connected each of the single 4-square elements directly to the feedline going to the shack, and the results were amazing: almost 40 dB of noise difference between the best (vert. # 1) and the worst (vert. # 4) confirming that:

there is no problem with common noise on the feedline
the noise seems to be originated only by the telephone line, and not the power line.

Removed the offending telephone cable.

So, the problem had to be solved with the Telephone Company, and ask them to find another route for that cable (Sep. 25, 2008). But being unable to get a reply from them, I found the cooperation of my neighbours, who let me make some tests while disconnecting any modem or fax equipment from their line. Nothing changed… so I decided to climb myself the two telephone poles, unhook the offending cable and find for it another route, down on ground, about 30 meters far from the array. The noise immediately decreased, and then I had to match again the impedance of each elements to restore 75+jX 0 ohm.

All this story, with installation details, images with graphs on S-meter readings and the comments on performance “on air” can be found on this PDF file:

These were the the noise readings on Icom Pro II S-meter on the single elements of 4-square, before and after removing the guilty cable:

Vert. 1 - Vert 2 - Vert 3 - Vert 4

before ……….. S 0 S 5 S 2 S 8

after …………… S 0 S 2 S 0 S 1

and these are the noise readings on whole 4-square:

West - NW - North - South

(side) (diagonal) (side) (side)

before ………. S 6 S 6 S 4 S 7

after ………… S 2 S 1 S 1 S 0

The noise is reduced by three s-units on Vert.2 and by seven s-units on Vert.4 (the closest to the offending telephone wire). But it is not yet perfect due to some noise still captured by Vert.2 : that’s my local environment problem !

2009: trying to improve the 4-square receiving array.

After a couple of months of tests, I realized that the 4-square was working quite well, with a front to back ratio better than on any other antenna but, generally, the Pennants were more “quiet”. Despite all my common mode cares and the feed line tests were not bad, I decided, as a last resource, to substitute all the buried RG213 feedline (150 meters), with the same RG6 style 75 ohm CATV SAT cable used for the array delay lines.

With the combiner box on the workbench, to substitute the output Xfmr, I decided also to add the small T-attenuators in order to improve the F/B ratio as suggested by my friend Lee, K7TJR, and described also on the ON4UN “Low Band DXing” from page 7-26.

In the combiner circuit we have three lines with different phasing, but also with different losses, which may cause a few degrees of variations from what calculated. So, the small T-attenuators are added on the lines with lower losses in order to null those differences to balance outputs.

Though I have not yet been able to find a Vector Voltmeter or some other suitable equipment for accurate measurements of RF magnitude and phases for precise adjustment of the two 1K ohm potentiometers (marked 1 and 2 in this photo), I calculated what should be the shunt resistance and set their value as described in this PDF document:

Now, we get a good F/B raport of 30 dB on all the eight positions of the array, but generally, only on strong signals. The very weak DX stations are still better heard on the Pennants! What can be done next ?

2010: from the homebuilt array to the K7TJR 4-square array with Hi-Z amplifiers.

After one year of many tests:

and “On the Air” DXing, it was confirmed that my new 4-square Rx array is working very well, with a F/B of 30 dB, but it never outperformed the Pennants system, where the S/N is always better on the weak DX signals.

The issue could be in the particular lot where the 4-square is located, surrounded by very close power and telephone lines, or in the array stuff itself. So, in a sunny dry afternoon I repeated once again the noise readings on the SDR-IQ receiver, described in the above documents, and came to this conclusion:

«… we must guess that the combiner, phasing circuit and feeding lines are working correctly: ALL THE NOISE is originated in the antenna elements. At this point I guess the problem could be in the loading wires of the vertical elements and/or the expanded ground system».

So, I put again, at almost one wavelength of distance, the 1.843 test oscillator and began to modify the 4-square elements, removing the loading wires and shorting them, one by one, while registering the noise variations.

With all the elements reduced to simple whips of 6.33 meters, the noise level became the same as in the best Pennant and the 1843 KHz signal a few dB better, despite the greater distance from the test oscillator ( 160 m. vs. 30 m.). Of course the impedance has not yet been matched to the 75 ohms feedline and, hopefully, there is a lot of room for improvement and getting a good pattern in a 4 square phased array.

It has been proved that a well designed and carefully built 4-square Rx array, with the W8JI top loading elements technique , does not work in my environment. A F/B of 30 dB was achieved, but only on strong signals …no way to pull out of the noise the weak DX, better on the Pennants!

The loading wires and the elaborate ground system did a good job for stable impedance matching and correct phasing, but produced a huge NOISE PICK-UP (may be due to the surrounding power and telephone lines).

Now back to EZNEC to find the impedance of the simple vertical after removing the loading wires this is the result:

R = 0,55 – jX 1214 ohms.

With such an increase, the inductance required for a stable matching is more difficult to deal by means of the old classic stuff. So, the best solution could be to leave the traditional array with passive elements and give a thought at the active antennas system with high-impedance amplifiers.

In over 40 years of DX activity I did never buy a commercial antenna, but working with small electronic circuits is beyond my technical capabilities, so I decided to test the 4-sq. array by K7TJR with his Hi-Z amplifiers. Lee is a good friend with a long experience in low-band receiving techniques.

Lately his production grew up a lot, from 4-sq to 8-circle, and on his great website

you can find all the stuff needed for Rx phased arrays at the cheapest prices of the market!

Most important, Lee is always helpful and his technical documentation is very useful to understand how the system works and to make the right decisions, on the personal needs.

So the first decision is to choose between an array optimized for maximum RDF (Receiving Directivity Factor) mode “A” or the best Front to Back ratio mode “B”. In the mode “A” you have a sharper lobe with a RDF of 12 dB, but two side lobes, while on “B” the lobe is broader, but with a better back null.

But these plots are for an array of at least 80 feet on a side in a clean open space free of fences, power lines or nearby metallic poles…. which is NOT my situation (more on this later). The operation mode depends only by the length of the two delay lines. Full instructions are given on the calculating procedure. In my case, with a 20 meter side dimensions (67 feet), the longest delay line should be about 48 degrees, for max RDF, and about 72 degrees for best F/B.

In the mean time my municipality came ahead also with the public lighting and so a new metallic pole has been added at a 20 ft. distance from a 4-sq. element : for sure it will not help !

Mod. A

Mod. B

Though it did not make the miracles I hoped, the K7TJR array works very well from MW, where an impressive F/B of 36 dB is shown on a BC station, to 80 and 40 meters. On 160 m. the F/B is “only” 20 dB, but my array side dimension are less than optimum and the noise still the same being too close to a high fence and power line on the left side and to the telephone line on the other, with an added new lighting metallic pole at about 20 feet from one ant. element. On the left picture you see the K7TJR small controller on top of my previous one. It needs only a 3-wire control cable (out of the 16-wire cable I had before) and it’s ready for use: no loading wires, no radial system, and no need to tune or adjust anything! Click here for all details and test reports:

September 21, 2010

2009: trying the first BOG, really a Beverage on ground.

In the month of February 2009 I used all my possible weapons in order to listen for a Pacific DXpedition over the North Pole path. After installing a two-wire Beverage, I decided to try also for the first time a Beverage on Ground, as long as possible.

(And that was a mistake - as I will learn later- the BOG antenna must be shorter than a Beverage!)

From the far southern ground stake of the 2-wire Beverage I laid down on the ground a single wire 230 m. long straight in the northern direction, where I drove another ground rod. With a load resistance of 330 ohms (lower than in a regular Beverage) I got a good uniform SWR at the input of the RG213 feedline, through a KB-1 Beverage transformer (9:1 ratio). As seen in the short paper “2-wire Beverage Follow-up” this antenna is very quiet, but also the signals are much lower, without any real improvement in the S/N ratio, and thus useless so far.

Anyway my experience confirms that “You never have enough antennas” and, once in a while, every antenna or piece of wire can be the right one!

Unfortunately the B.O.G. is a typical temporary antenna (a nuisance for footing) and cannot be kept for a long time.

2011: a 3-wires mini-loop by Dr. Dallas Lankford design.

From 2010 to 2011, while going on with greater projects, I tried also to realize some ideas of dr. Dallas Lankford, who published a lot of files about BC and lowband antennas on this website:

After given up with the Quad Delta Flag Array project, due to my lot constrains, see

here: “How a great Rx antenna array can be destroyed in a bad environment”,

I turned my attention on a new small receiving loop and, in June 2011, decided to build it. I belong to the old generation of electronin tubes and relays, without experience with transistors and pc boards, but this construction has been quite easy with the “Ugly bug” technique.

From the original design I duplicated the Hi-Z amplifier (bought everything by Mouser) with a couple of minor modifications: I found a very cheap shielded CAT5 cable and wound 20 turns around a big FT-240-31 for more than 3.000 ohm of impedance as a common noise choke on the antenna side… I ordered the Norton amplifier to Clifton Laboratories, but not yet mounted it.

As we will see the Hi-Z J310-J271 amplifier is working well with correct impedance matching and there is enough signal for the tests, which are reported in this PDF file, with all details:

This is the complete diagram of the loop and Hi-Z amplifier:

The photos show, from left to right: the Hi-z amplifier in aluminium box, with banana plugs; then, mounted in a sealed electrical plastic box, on one side of the supporting fiberglass arm; and, on the opposite side, a temporary container with the 10 turns loading potentiometer.

The first tests were on some known AM BC stations: NO BACK NULL was found while adjusting the potentiometer, but a very DEEP SIDE NULL was found at the minimum resistance value. Thus the antenna does not work as a Flag, BUT WORKS VERY WELL AS A LOOP - to cancel offending signals - on Broadcasting AM band, where a station can be reduced by 30 to 40 dB. With such a sharp side null, I could install it on a rotator, but on 160m. it’s way down and, in any case, my problem is the general noise level and not a strong QRM coming from a definite direction…. so not worth for sure!

2010: the “TX3A” DHDL - Double Half Delta Loop - receiving antenna, and doubling it.

The DHDL - Double Half Double Loop - receiving antenna belongs to the family of Pennants and Flags, and its design originated in Florida, as the Waller Flag, by Jose Carlos, N4IS, and developed by George AA7JV, who used it, for the first time, in his TX3A DXpedition on Chesterfield is. in November 2009, hence the name given.

The basic design is essentially two half delta loops interconnected to create a single antenna, with one transformer and one load resistor, as well documented by AA7JV on a page of his beatiful TX3A website:

The innovative idea of connecting those loops at low level, without a long top wire, allows an easier and lighter construction. I have been always fighting with noise and interactions, and tried every kind of RX antennas, within my constraints and limited space. I found that the best ones, i.e. more quiet, are still the Pennants, their best characteristic being the “ground independence”, i.e. no ground connection.

So all my past experiences encouraged me to enter EZNEC with the original tx3a.ez file they sent me, and investigate on the best possible design which could fit on my property. The RDF achievable with two of these arrays is above 12 dB, really impressive, and the gain is “only” – 27dB, which is double (or better to say, half the loss) of the -55 dB of the Waller Flag. That means less problems with common noise and preamplifier requirements. With 74 meters of total length, I could get the same results of a long Beverage into the North direction where my toughest and most needed Pacific DX area is located.

This is the “Antenna View” imagine with the horizonthal pattern of a single TX3A DHDL, whose dimensions are:

vertical wires 1 and 6: from 2 to 9 m. high (7 meters); horiz. wires 4 and 3 (at 2 m. high): 11 + 11 meters, with these results: gain - 28 dB; F/B ratio: 33 dB; RDF: 10 dB.

Two of these DHDL's, in an end-fed configuration, with 30 meters of separation (and thus 22 + 30 + 22 m.) can give, with the same - 28 dB gain, a F/B of 35 dB, and a beatiful RDF of 12.28 dB.

All details, with pictures and tables can be found on this PDF document:

After having examined with Eznec all possible variations and configurations, it has been confirmed that this antenna wants to work whatsoever, so next I took into consideration the available supports and the best placement among my fruit trees. The physical beauty of this array is that everything is at about 2 m. height: loads and feeding boxes with the low connecting wire stretched between them without disturbing anybody. Furthermore the entire receiving loop is extending above the metallic fences. I bought four fiberglass fishing rods 7 m. long, thin enough to support a very light 1 mm. aluminium wire (a cheap Japanese type sold in garden markets for bonsai use) and tied them on 2.5 m. wooden poles (used for trees support).

The transformers used are wound on binocular BN73-202. For 75 ohm cable, I calculated 2 turns on the low side and 7 turns on the high side, but after installed I got a perfect match by removing one turn, Thus the final Xmfrs are both 2 turns on primary and 6 turns on secondary.

The two DHDL are in end-fed configuration, and these are the azimuth and elevation beautiful plots (primary black traces) for correct phase reversing. I added the recalled bleu traces for normal phasing, that is, after inverting one of the feedlines connection, and we see that a broader pattern at higher angle results, which could be useful in some occasions. So, I put a couple of relays in the closest box containing the binocular transformer to provide an easy way to swap the feeding phase.

On the air tests of these first days are well promising despite propagation or activity from Central Pacific is not as good as it was some weeks ago. Europeans and Russians high angle signals are always better with the normal broad lobe position; the same with Japan, which is at 45 degrees. The only one I heard better on the reversed phasing – best weapon position – has been Merv, KH7C, as it should be, straight through the North Pole at low angle. But this path is exceptionally possible only at low solar minimum, without Aurora or any geomagnetic activity. Generally the signals are skewed East or West of the Polar Cap, so I’m afraid that such an array, in the North direction, will be useful only for a limited time at low solar cycle…

In any case it has been a positive new experiment with receiving antennas and the next one will be with a broadside configuration. This will give an even better pattern than the end-fire, and an RDF reaching 13 dB with a separation of 100 meters between the loops. Unfortunately, while the elevation angle is higher with a broader lobe, the horizontal one is very sharp and in my particular situation it does not cover any needed direction. It could be placed only into 90 or 270 degrees, not enough for my most wanted area of the United States.

So, I haven't tried the physical construction yet, but I have published, as usual, all the tables with the Eznec model results, including the broadside configuration. Rys, SP5EWY, took this cue and successfully built such an array at his second DX station. On January 27, 2010, he published the following comment on Topband Reflector:

Hi All,
Beeing inspired by an article written by Luis IV3PRK http://www.iv3prk.it/user/image/..-rxant.prk_tx3a.pdf (Thanks Luis!) I'v decided to build Double Delta Loop used by George on Chesterfield in a broadside configuration. Because of TX3D Australs DXpedition I'v made 250m Beverage in their direction and and just a week ago I'v built also this described Delta system.
I was trying both antennas in my second QTH 45 km from my house (its an orchard) yesterday but Deltas was not working and on Beverage TX3D was too weak. When I checked I discovered that one of wire in a delta was broken , so practically today was the first day I was able to compare antennas.
When I started to listening on Deltas in the first minutes I'v thought they are broken again because there was almost not any noise in a RX...but checking the band I'v found a lot of strong NA stations with very clean signals. I tested Deltas with the Beverage many times and always a difference was huge. Even a strength of stations was about 2S better on the Beverage the difference in the level of noise was not comparable...it was like day and night.
Using the Beverage I haven't heard even a trace of TX3D....on Deltas I'v heard them about 20 minutes ....but with deep QSB.

73 Rys SP5EWY

Rys sent me his building details:

«The distance between Deltas is about 95m. Resistors are 1250 ohm, tranformers 12:1 bifilar windings, 2 pieces of coax 75 ohm each one 50m connected together to transformer 2:1 and to 75 ohm coax. I used similar fishing bats... 8m high but I removed the last thinest part , so they are 7,1 m high attached to 2m high steel elements. The wire is 0,7mm. TX3D was my #301 so I'm really VERY glad!»

and the picture of his SP5EWY second low-band DX location showing one of the DHDL elements and far, on the back, his ¼ wave vertical.

The environment is quite different from mine… but now I am sure that the broadside design works OK, maybe I will give it a try in the future, if needed in my only possible location. For the moment, in that north-south direction, it can be only left in end-fire configuration.

January 28, 2010 Luis IV3PRK

But just one month later, in February 2010, two Pacific DXpeditions were expected: TX4T from 320° and H40FN from 40°. So, I decided to dismantle that useless 360 degrees DHDL array and to move the single deltas on the northern side of my lot, near the existing Pennants, and point them to the most desired directions. That area is less influenced by noise of the utility lines, and interactions with the TX antenna, which remains on the back. It’s a pity to waive the better RDF of the end-fire configuration but I must be content within my space limits.

100 ohm

In the mean time I received a good advice from Carlos, N4IS, on a new feeding system and common mode noise blocking. His idea was quite simple: the use of twisted pair wires - which are balanced - instead of coax cable - which is unbalanced! Easy to do: I stripped a 15 m. length of CAT 5 data cable and extracted the 4 pairs of twisted wires; the characteristic impedance of the pair is 100 ohms, so I just removed one turn of the DHDL binocular transformer to get the correct impedance.

The 15 m. length is just arbitrary, what I had. At the end of each twisted line I put a binocular isolating 2:1 transformer to match the 50 ohm coaxial cable going to switching box, where another CMC, with its a ground rod, is installed.

More details can be found here:

Unfortunately, I did all this work in vain, as I didn't manage to hear those DXpedions on Tahiti and Temotu (which I still need), while I had already worked in October 2009 TX5SPA, a polish DXpedition on Austral is.

In the next three years I went ahead with further Eznec modeling, and mounting, testing and dismounting DHDL antennas at the search of the best performance in my most convenient location.

2013: the broadside array of two DHDL's.

In September 2013 we were waiting for a new DXpedition from Lord Howe is. by VK2CCC as VK9LL.

Tomas loves 160 m. and promised a big effort, as usual, on that band. Propagation in this period is not yet as good as it was in November 2009, when I worked him as YJ0CCC with his 100 watts only, but I decided to try all my possible in order to work my last needed DXCC country in that area. In the nearby field, the grass mowing had been completed. It’s laying north-south, useless for Beverages, but some kind of broadside array can be fit to point to the target direction, which is at 65 degrees.

This is a recent Google Earth image of my location - between two public roads - with superimposed all my antennas left on September 2013:

1) the Tx shunt-fed tower with still the 4 elevated radials;

2) two groups of three pennants each, switchable also in broadside feeding;

3) the W7IUV rotatable flag;

4) the new double DHDL broadside array (red lines) with the connecting feedlines (thin red lines) and the switching boxes between them (white small squares) and its directions 65/245 degrees.

This is the EZNEC printout of the DHDL's broadside model, whose tables and patterns can be found here:

The following is the final azimuth slice of the chosen broadside array at 23° elevation angle (red trace) superimposed to a single DHDL (blue).

As I had already a CAT5 cable coming there from the shack I provided switching directions also for both single DHDL’s. I did it in order to test how the same loop can perform different if placed 90 m. apart in my noisy environment.

For a normal use some of this switching is in excess!

This is the complete DHDL broadside drawing; alll the details and measures are in the same PDF document:

Remember that in broadside feeding the lines must be always equal, no matter how long.

These are the four antenna (A) two switching (B) and central (C) boxes completed (left photo).

All the system has been checked for correct switching and impedances outside the shack (right).

Below: one antenna box mounted on the wooden pole at the base of the fiberglass fishing rod; to the rigtht, the central box with a common mode choke on the feedline to the shack: 12 turns of RG6 cable on FT200-31 core.

And finally, this is the southern DHDL in better free space, and its central crossing point.

The inpedance readings on the AEA CIA analyzer were perfect 75 ohms on the four single DHDL's positions and very good also the two broadside directions. ====>>

I had not yet enough time to make many tests on the air and comparisons with other receiving antennas, but two cases are sufficient for me to say that the DHDL broadside array is working ok:

1. after carefully listening to EA3JE who is only 1.000 km. far, but at 245° bearing, exactly on my back array direction and with a strong and stable signal for true s-meter readings: above 20 dB of difference by switching directions.

2. I managed to hear and work VK9LL – a very difficult one in just a short opening – which was the target and the object of all this work, thanks Tomas for giving my # 306 DXCC country on 160 m.!

Tomas wrote a very interesting report with all his set-up details, still on the web:

and sent me the 160 m. log of his DXpedition, which I analyzed on DX Atlas

and wrote this paper:

Those data have been used by Carl Luetzelschwab, K9LA, for a deep study on ducting propagation, and gave origin to this page on my site “The 160 m. propagation mystery”:

Finally, it had been worth it! This time, all the work I'd done hadn't been in vain, even though I knew it was only temporary. In fact, I was already in the process of moving to Ecuador, and all the antennas would be dismantled within a few months.

September 30, 2013 Luis IV3PRK