Phased Pairs of Antennas for HF Reception

Matt Roberts - matt-at-kk5jy-dot-net

Published: 2023-02-10


Antenna Pattern
One question I frequently receive is how to build a phased-array from a pair of receiving antennas.  I have talked about such antenna pairs in various articles, including the loop-on-ground and phased vertical SRL articles.  Using a pair of identical receiving antenna elements, with a carefully selected phase delay between them, allows for creation of interesting antenna patterns, such as the cardioid.  Such patterns enable the antenna to attenuate stations or noise from undesired arrival directions.

I will admit that I have limited experience with phased pairs.  My location is such that DX arrives from all directions, and sometimes from multiple directions at once, so unidirectional antennas aren't as helpful as I would like.  Nonetheless, I hear from many people who would like to try these antennas, so here are my notes on how to pursue building HF phased pair receiving arrays.


Try a Single Element First

If you want to build a pair of receiving antennas — especially if your goal is to get rid of some local noise source — try a single antenna first.  E.g., if you are thinking of building an SRL or a LoG pair, use a single SRL or LoG at your location for a few weeks to see if that antenna type is compatible with your gear and your location.  Any antenna type that will help your reception in a phased pair will also give you at least some improvement as a single element antenna, and for a tiny fraction of the price (and space).

For many people, that smaller improvement may be enough by itself.  Most of the time, it has been for me.

Some elements, such as the SRL, have some really nice patterns already.  The SRL has two nulls that can be easily turned to point at a nearby noise source.  That alone might cancel out your QRM problem, and get you on the air for just a few dollars.

If you choose an antenna element that doesn't like your location, e.g., because it couples to some nearby noisy device more than your transmit antenna, building two of an ineffective antenna will likely perform no better.  Adding pricey phasing hardware will leave you even more frustrated.

Some well-meaning pocket-protector types will tell you that "pattern is everything." Pattern is important, but in the real world, you will find that some basic receiving antenna types just work better than others in your local environment.  The antennas that work best for me are entirely different from the ones that work for my HF contesting friend across town.  The location and what's around it make a lot of difference.

This is probably the best advice of this entire article — if you are considering a receiving antenna pair, try a simple, single element first, to see how it plays.


Initial References

If you want to experiment with phased arrays, even just a pair of antennas, you're going to need to do some reading on the subject.  I suggest starting with the ARRL Antenna Book.  The most recent version I have is the 23rd Edition from 2015.

In that edition, Chapter 6 contains information about how to generate predictable phase delays in both receiving and transmitting antennas.  Specifically, Section 6.2 deals with Driven Arrays, which is what I talk about in my receiving antenna articles.

The material on building driven pairs of antennas applies equally well to transmitting and receiving antennas.  The one benefit of a receive-only array is that the components do not have to stand up to large RF power levels, so you can use less expensive reactive elements.


Design First, Then Build

Another critical tool for building phased pairs is the computer model.  There are at least three different — and free — modeling packages available for amateur radio use, all based on the NEC engine.  These include EZNEC, MMANA-GAL, and 4nec2.  I have used them all, but most of my models are done in EZNEC.

Modeling an array will allow you to experiment with different delay values and antenna spacings, without having to cut any wire or cable.  It will also let you experiment with feedlines, to see the effect of changing type or length of lines.  When modeling both antennas and feedlines together, you get the benefit of seeing the effect of reflections and mismatches on your antenna pattern.


Phased Pairs — The Easy Way

There are two different approaches to building phased pairs that I have actually tried, but if you want the easiest solution that is as close to turn-key as it gets, you will want a commercially built phase controller, such as the DXE NCC-1 or NCC-2.

In its simplest configuration, you build two identical antenna elements, space them apart in your yard, then connect them to the phase controller using identical lengths of feedline, and that's it.  Everything else you need to get the right phase angle between the antennas is a matter of flipping switches and turning knobs on the controller box, until you get the effect you want from the antenna.

You can even download the PDF manual for either model from DXE, and it will show you recommended configurations of hardware to build a phased pair of receive antennas.  DXE likes to use CB whips as the elements, but I prefer the SRL, because the nulls to the side give sharper patterns for end-fire configurations.  But this is a matter of what your goals are.

Other elements could be LoGs, DoGs, or any other simple antenna element.  The goal is to make the controller box generate the delay you need to get the effect you want.  Since the phase is continuously variable, you can literally just turn the big knob until you get the best sound from the other station.  The actual pattern generated may not be the one you think, but so what?  This is what makes the controller the best option for most people.

Ideally, you will want to follow DXE's recommendation to put weather-resistant preamplifiers out at the elements, even though this isn't strictly necessary for identical elements with identical feedlines.  The goal of remote preamplifiers is two-fold:  First, you want to capture as much signal as you can, and amplify it where you will get the best SNR possible.  Second, using a remote amplifier provides a buffer between the untuned element and its feedline, keeping feedline reflections at a minimum.  This keeps nulls nice and deep.

People look at the cost of a phase controller and wonder if it is worth the cash.  The short answer is yes, unless you are building an array for only one band and direction, such as a DX antenna to Europe for 160m, or a directional transmit pair.  The controller is simply unbeatable for allowing you to fine-tune the pattern you want quickly.  When you switch bands, you just re-adjust the phase to optimize the desired signals.  If you want to flip the pair to the opposite direction, you just flip switches.  It really is the Cadillac solution, and DXE has helpful people to help get you moving.


The Cheap(er) Way

If you really do just want a receive pair for a single band, you can build such an array with passive elements.  I described this briefly in the SRL pair article, but here is the basic recipe.

Once you have identical elements in the yard with the desired spacing, you can run feedlines of different lengths to a passive combiner, such as the DXE RSC-2, SC-50, or SC-75.  The phase delay is generated by the differing feedline lengths.  It is the difference in line lengths, that generates the delay.  The delay calculation should take into account the feedline type, since the velocity factor differs between cable types.  The delay is also dependent on the frequency used.

The difference in line lengths is the delay angle desired, multiplied by the wavelength of operation, multiplied by the velocity factor of the line.  The delay angle is a multiplier, where 360° is a multiplier of 1.0, 180° is 0.5, 90° is 0.25, and so on.  This is covered in more detail in the ARRL text.

The output of the combiner feeds the receiver directly.  If you want to change the direction of the array, you reverse the feedlines, so that longer and shorter lengths of the delay section feed the other elements, respectively.  If you want to switch bands, you change the cable length to get the delay needed on the new band.

When modeling this type of phasing network, I found that it is of particular importance to use buffer amplifiers on each antenna, and to keep the feedline lengths from the antenna elements to their respective buffer amplifiers equal.  Ideally, the amplifier would be out a the antenna, but if this isn't practical, at least keep the feedline lengths equal.  Between the amplifiers and the combiner, the feedline lengths will differ, of course.  Buffer amplifiers keep reflections on the mismatched feedlines from interfering to corrupt the overall pattern.  If the antenna elements are matched, either by resonance or by resistive loading, this becomes less of an issue.

The buffer amplifiers do not need to have positive gain, but they certainly can.  Whatever the gain of the amplifiers, they should be very similar to each other.  Generally, you want both legs of the antenna system to be as close to identical as you can get, so that when they are combined, you get a clean mix of the two phased signals.

This is another reason to consider going with a dedicated phase controller device.  By the time you buy two good quality amplifiers, a good quality signal combiner, cut all the feedlines for the different bands and patterns you want, then get it all wired and tested, you have spent a good chunk of what it would cost you to get a used NCC-1.  Again, the passive approach is probably best for a single band and direction.


The Antiphase Pair

One special case of the passive phased pair is the antiphase pair.

If you take two vertically-polarized antenna elements, and separate them by an appropriate distance, and then feed them 180° out of phase with each other, you get a sharp, bidirectional, endfire pattern, similar to an elevated horizontal dipole antenna.  When I model this, I get even sharper lobes than the ARRL text describes.

Such a pattern gives modest gain in two diretions, but very deep nulls everywhere else.  This kind of antenna could be used to emphasize one or two DX directions, but cancel out every other direction.

This is probably the easiest phased array to actually build because it doesn't require delay calculations or circuits.  To get the 180° phase difference, simply feed one antenna backwards from the other antenna.  E.g., if using identical vertical loops, simply reverse the antenna terminals of one antenna.

Otherwise, this antenna is very similar to the passive approach described above, but without the delay line.  Everything about the two antenna elements north of the combiner is identical.  It's the spacing and the reversal of one antenna "polarity" that does all the work.

It should even be possible to feed such an arrangement without any buffer amplifiers, although I have not tried this.


Other Options

There are even simpler options for directional receiving antennas besides the phased array.  The Beverage and rhombic antennas are obvious examples.  But there are others, including the terminated loop.  That antenna can be made any size, and a mast-mounted preamplifier would give enough signal for just about any HF band.


Experiment

Be ready to experiment.  If you don't want to tinker with different configurations and see how they work, then don't go the cheap route.  Get the phase controller, because that is the least effort to get to a working antenna pair.  Phased pairs are an advanced topic for most amateurs, so be ready to learn new things.

If you aren't willing to invest some effort, this topic is not for you — get a beam antenna or build a Beverage...

Copyright (C) 2023 by Matt Roberts, KK5JY.
All Rights Reserved.