Saturday, March 19, 2016

Loop Feed and Measurements

In the previous post I showed the loop antenna I'm using at present. Full wave loops will resonate on all harmonics so a 40 meter loop can also be used on 20, 15, and 10 meters, however, you can't get a perfect match on all bands. What's more, if the antenna length and feed impedance are adjusted for a perfect 1:1 match on one band, the other bands will suffer. I prefer to go with a "happy medium" approach, where all bands provide an acceptable match. With this antenna that means an SWR of around 2:1. That happy medium occurs, for this antenna, at about 115 ohms.

Below I'll describe the matching device I built for this antenna and the resulting measurements. The objective for the feed is to match unbalanced 50 ohm coax to a balanced 115 ohm load. The feed shall also choke off or isolate common mode currents so the signal travels inside the coax and not on the shield.

A transmission line transformer configured to transform from one impedance to another is always an unun. It cannot provide isolation. If we also want isolation we must use two separate transmission line transformers in series. In this design, isolation is provided by a transmission line transformer configured as a 50 ohm 1:1 balun followed by a transformer that steps up the impedance. Loss in a well designed transmission line transformer is less than one percent so the combined loss of two in series is still negligible.

We begin with the impedance transforming unun using a T140-61 core and some 16 AWG magnet wire. Type 61 material is more expensive but I chose it for this transformer because it has very low losses from 3 to 30 MHz. This toroid works like an autotransformer so there is some magnetization of the core and low loss is, thus, important.

Ferrite is a very hard material so I like to cover toroids with one layer of Scotch 33+ to provide some padding for the wire. This ferrite is non-conductive and the edges are smoothed so the tape probably isn't necessary but it makes me feel better.

Next, I cut small strips of the fiberglass tape shown above to make the transmission line as shown below. The tape has a silicone adhesive, is rated to over 500 degrees F, and most importantly it doesn't stretch. Taping the wires like this makes it much easier to wind the toroid core while keeping the conductors side-by-side.

Below is the finished toroid with seven trifilar turns. These will be wired in series and used like an autotransformer with a 1.5 to 1 turns ratio, giving a 2.25 to 1 impedance transformation. In other words, this configuration is a 2.25 to 1 unun, transforming 50 ohms to 112.5 ohms.

FT140-61 toroid with seven trifilar turns

Next we use an FT240-43 toroid and a two conductor transmission line made from 14 AWG magnet wire to make the isolator or 1:1 balun. The objective here is to come up with the highest common mode impedance using the least amount of wire. I chose type 43 material which is cheaper, has much higher permeability, and higher loss, for the following reasons:

  • We want to minimize inter-turn capacitance in order to keep the self-resonant frequency of the balun as high as possible. That means we want as few turns as possible. Type 61 material has a permeability of 125. Permeability of type 43 is 850, allowing fewer turns to reach a certain inductance.
  • Type 43 material has higher loss but in this application it doesn't matter -- it's actually a benefit. Core magnetization in this type of 1:1 balun is theoretically zero. If there's no core magnetization then there's no loss. The only core magnetization that might occur results from the small common mode currents that we're trying to get rid of. If those disappear as heat in the core, so much the better. Core heating in this kind of balun only becomes significant at power levels well beyond 1500 watts so it can be ignored.
  • Type 43 material costs much less than 61, which is nice, but didn't drive the choice.

Below is the finished isolator with six turns, connected to the 2.25:1 unun, ready to be installed in an outdoor rated housing. Note the crossover. The crossover has no effect on the performance of the transformer and provides for the input and output wires to come out on opposite sides, which is handy in this case.

Below are the two transmission line transformers installed in a Bud NEMA4 waterproof polycarbonate box. All hardware is 18-8 stainless.

Shown below is the first test setup. A 115 ohm 1% resistor is fixed across the balanced output terminals and the analyzer was swept, taking 500 samples from 3 to 30 MHz. SWR measured nearly flat with 1.01 at 3 MHz, rising to 1.09 at 30 MHz. The photo shows the analyzer at 14.1 MHz reading 51 ohms, exactly what one would expect with 115 ohms on the output terminals.

Below is a 500 sample sweep from 3 to 30 MHz of the final installation. I originally made the antenna using 143 feet of wire, which in my experience is a little long so I can trim it. Because this installation has one leg of the triangle parallel to and 10 feet off the slope of the hill behind my house. The antenna resonated about 3 percent low. My calculation said to shorten it by 5 feet, which brought the antenna to exactly where I want it.

The antenna can be used on all four bands without a tuner.

Look for me at night around 7070 kHz on PSK31, Contestia, Olivia, Thor, PSK63F, and Hellschreiber modes. 73 de KW2P.

Monday, March 7, 2016

Clarksburg 40 Meter Loop Antenna

This QTH is problematic for antennas. The lot is very narrow, just 25 feet. The house is tall, long, and narrow, just 18 feet wide. It's kind of like living on a sailboat. The attic is impractical to access. There is an attic trap door but it's too small for an adult to fit through. There is no back yard, just a steep drop behind the house and the front door opens to the sidewalk. Ground radials are not an option. However, there is a tall tree, part way down the hill just behind the house.

My favorite band is 40 meters so that was the main target. I considered several antennas including an end-fed wire, an off-center fed wire, a vertical dipole, a C-pole, and a loop. Modeling the various antennas with NEC2, the C-pole for 40 meters looked promising. It's shorter than a dipole and needs no radials but it's a single-band antenna.

I decided to go with my favorite antenna, the full wavelength loop. The question was if 143 feet of wire could be made to fit. Loops work well, are easy to tune, have high radiation efficiency, and work on all harmonics. Using my laser measure I found that there was just enough space for a triangle running from the upstairs window, to the far branch of the tree (40 feet), then down to an insulator, and back up to the feedpoint at the window.

Using my line launcher and a one ounce lead sinker, I shot non-conductive nylon monofilament line through the tree. As expected, it overshot and landed across the power lines behind the house. I slowly pulled the line back off the power lines and lowered the sinker to the ground. There I attached paracord and hauled it up, over the tree, and to the house. With the paracord I hauled into place the 18 AWG copperweld wire for the antenna. Shown below is the configuration.

143 Foot (43.5m) Wire Loop
The photo's angle gives a false impression of the shape of the loop. The top segment from the tree to the house is actually almost level. It's impossible to take a straight-on photo of the antenna.
The antenna is invisible unless one looks carefully. Even I have a hard time seeing it.
This loop performs very well. Loops almost always do. Below are some plots showing the nearly omnidirectional pattern. Polarization is about 80 percent vertical and 20 percent horizontal.

Antenna Geometry. Feed point is the top right corner.

Pattern on 40 Meters

Pattern on 20 Meters

3D Surface Depicting Gain on 20 Meters

I'm still refining the feed and matching network for the loop. I'll post an article about that soon.

You'll find me using digital modes, PSK, Contestia, Olivia, Thor, MFSK16, and Hellschreiber at 7070 and 14070 kHz. I hope to see you on the waterfall! 73 de KW2P.

Wednesday, February 24, 2016

On the Air, Finally

After a long hiatus, I'm back on the air. QTH: Clarksburg, West Virginia, Harrison County, locator EM99tg. Primary mode is digital. Antenna is a 40 meter vertically oriented triangular loop, which gives good performance on 40, 20, 15, and 10 meters.

The Rig

I'm also working with my daughter, KC7BNH, in Coeur d'Alene, Idaho, to help get her back on the air as well.

Monday, December 17, 2012

Maritime Mobile Nets / Frequencies / Schedules

The KW2P Ham and Maritime Mobile Net Frequencies list has moved to a new URL here:

If you find any errors or have corrections, please let me know.  An easy comment form is here:

Wednesday, August 18, 2010

KW2P Beverage Antenna Designs

In a perfect world we'd have full-size 160m Beverage antennas fanning out like the spokes of a wheel from a centrally located shack, and the feedpoints would all be located near the shack. Most of us don't have the necessary 80 acres of land so the feedpoints to our Beverages often end up far away and must be fed through long runs of coaxial cable.

For example, let's say I want to install a unidirectional Beverage aimed northeast and the shack is located in the northeast corner of the property. The Beverage wire must extend 800 feet towards the southwest of the shack, the termination resistor must be located at the shack end and the feed is all the way at the southwest end. I have to run the 800 foot Beverage wire plus 800 feet of coax to bring the signal to the shack. You can't do anything to change the geometry of this problem but I'll show here how the coax can serve both as the feedline and the Beverage wire.

Sometimes we build reversible Beverage antennas that require long runs of coax plus distant relay boxes to perform the required switching. The coax is often buried, making it susceptible to physical damage, especially on farmland, and subject to contamination from constant exposure to moisture. Buried coax can be punctured by nearby lightning strikes. Locating the damage and making repairs can mean replacing the entire run of coax.

Some time ago I developed construction and feed methods for Beverage antennas that eliminate or at least reduce some of these problems. Recently I searched all the antenna reference materials I could find, assuming that someone else must have also developed the same or similar designs, but found nothing. So I'm publishing the designs here. Hopefully they'll be of use to others.

There are three embodiments of these ideas discussed below. These designs make use of coaxial cable for both functions: both as the feedline and as the Beverage antenna element. These designs take explicit advantage of the fundamental characteristic of coaxial cable--that RF traveling on the outer skin of the shield has no bearing on or interaction with RF traveling inside the coax. The signal propagation modes that occur on the inside and outside of the coax are also completely different and propagate at different speeds. Due to the skin effect, the the RF currents traveling on the outer shield travel on the very outside surface and do not even appear on the inside surface of the shield.

Each of the designs are shown using 75 ohm coax. The designs can be adjusted for 50 ohm coax simply by changing the turns ratio of the matching transformers. CATV distribution cable such as RG-6-messenger is ideal for these designs because it includes a steel messenger wire running parallel to the coax and molded into a weatherproof jacket. The steel messenger wire allows the cable to be pulled to much higher tension than would be possible with coax alone, allowing longer spans between supports. RG-6 is also a non-contaminating coax so it will last a long time exposed to the weather or buried. It's tempting to use RG-59 because it's cheaper and it weighs about a third of what RG-6 weighs but it has a contaminating jacket.

The designs discussed focus on 160m but the same principles can be applied for any band.


The first diagram shows a typical unidirectional Beverage antenna with the terminating resistor at one end and feedline at the opposite end. The feedline is connected through a transformer that transforms the impedance of the feedline to an impedance closer to that of the antenna to improve coupling and minimize loss.


EMBODIMENT ONE: The basic idea.

The basic design concept is shown in the following diagram. The additional embodiments below use the same technique described here. The coaxial cable is suspended above the ground and the outer skin of the coax shield serves as the "wire" of a classical Beverage antenna. The tiny currents induced in the antenna wire (the outer shield of the coax) are referenced to earth ground and are presented to the 450 ohm primary of matching transformer T-1, exactly as in the classical Beverage shown above. T-1's secondary connects to the shield and center conductor of the coaxial cable feedline, which happens to be the same coax that forms the active element of the antenna. The RF signal injected by T-1 propagates inside the coax to the opposite end (the terminating resistor end) of the antenna. Note that T-1 is an isolation transformer with two independent windings.


At the terminating resistor end of the antenna, we are faced with the problem of extracting the signal we want, which is propagating inside the coax, while preventing the RF currents traveling on the outer surface of the coax shield from flowing beyond this point. This is a common problem in antennas that is solved by means of a balun (L1). However the problem in this case is bigger than we usually face with ham antennas. In the case of a Beverage antenna we are likely working at 1.8 MHz, which means the inductances required are large. We are also working with an impedance that is 10 times higher than what we normally work with so the required inductances are that much higher still. (Remember that the balun is concerned with blocking the outside surface currents at the 500 ohm impedance of the Beverage. The 75 or 50 ohm internal impedance of the coax is irrelevant as far as the balun is concerned.)

The rule of thumb for baluns is to present an inductive reactance that is 10 times the impedance we're working with. For 50 ohm coax you aim at 500 ohms. In this case, the impedance of the Beverage wire is 500 ohms so we'd like to see the balun present 5000 ohms of reactance. At 1.8 MHz, this is a relatively huge amount of inductance--about 450 uH. However, working in our favor is the fact that losses at the terminating resistor end of a Beverage have somewhat less effect on signal output than losses at the feed end of the wire so we can fudge down on the 5000 ohm requirement and call it 2500 ohms. But even so, we are still looking at 225 uH. Suitable baluns are discussed at the end of the article.

EMBODIMENT TWO: Feed it anywhere.

It's probably obvious to some readers that since the coaxial cable (in terms of the signal traveling on the inside) is untuned, its length does not matter. The coax does not have to continue for the full length of the Beverage antenna as shown above and the feedline can be brought off at any point. The advantages of this are clear. Instead of worrying about where the endpoints of the antenna are with respect to the shack, all the antenna has to do is pass nearby the shack and the feed is brought off at the nearest point. Several Beverages covering different directions can be installed and as long as they pass near the shack at some point the feedlines can all be very short.


EMBODIMENT THREE: Reversible KW2P Beverage

This variation may also be obvious to some readers. Note that I have never built and tested this variation but I have no doubt that it would work fine. I'm hoping to find and acquire a piece of land large enough to try this out.

Reversible Beverages invariably have relay boxes at the far ends of the antenna to switch between feedline and terminating resistor in order to reverse the antenna pattern. The concepts shown above in embodiment two demonstrate bringing the feedline off at any point along the antenna's length. The same method can be employed to bring the terminating resistor to certain points along the antenna or all the way to the opposite end. Directional switching can take place in a single box located at either end of the antenna or at certain points along the antenna's length. Switching directions is simply a matter of swapping the feedline for the resistor at L1 and L2.


Now comes a question: Note that in the first two embodiments, the length of the coaxial cable(s) did not matter. In this third embodiment, I assume that the lengths of coax are halfwave multiples (electrical length), taking into account the velocity factor of the coax (inside). The reason for the 1/2 wavelength multiples is to ensure that the resistance of the termination resistor is reflected accurately at the other end of the coax as a pure resistance. However, if the impedances of the Beverage wire / matching transformers / coaxial cables are all matched closely enough that SWR inside the coax is low, the lengths should not matter and it should not be necessary to hold to 1/2 wavelength multiples. This remains to be tested empirically. For now I show 1/2 wavelengths because I know it will work.

Lightning Survivability

One thing to consider when building Beverages is ease of construction and low cost of components like transformers and baluns because these components are frequently destroyed by lightning. A Beverage is a very long wire so lightning strikes hundreds of feet away can still induce plenty of current to vaporize baluns and transformers, puncture insulation, etc. Spark gaps at strategic locations are inexpensive, low-tech, and well worth the effort. Each support should be equipped with a ground rod and spark gap. Nothing will save you from a direct or very close hit but spark gaps will protect against most of the nearby hits. It is also a good idea to frequently inspect the antenna and spark gaps. A spark gap that was vaporized by yesterday's storm won't protect you today.

Matching transformers cannot be made lightning resistant but fortunately they are cheap and easy to make. If you're going to make one, make several at a time. You'll need them.

Suitable Baluns

There are several ways to build a suitable balun for this antenna. The type of balun needed in these designs is actually an unun (unbalanced-unbalanced), also called a "current balun", or a Collins balun. What the balun is doing is it simply uses inductance to choke off or block the antenna currents traveling on the outside of the coax, keeping them "up on the antenna wire" and not traveling down the feedline into the shack.

Construction parameters to consider are:

1) Keep inter-turn capacitance as low as possible to keep the self-resonant frequency of the balun as high as possible so the antenna can be used on 80m and even 40m if desired. If operation on bands other than 160m is not planned then self-resonant frequency is not that important as long as f0 is well above 2.0 MHz.

2) Baluns can be and should be designed with lightning surges in mind. A direct hit near the balun will likely destroy it no matter how it's constructed but most lightning surges will come from induced current from nearby strikes that a properly made balun can withstand. The more surge resistant you make the balun, the less often you will have to replace it. Making the balun more lightning resistant mainly consists of making the balun physically large and distributing the surge voltage across the whole balun so that it doesn't arc over between turns. The best way to achieve this is with a long solenoid-shaped coil that keeps the first and last turns as far away from each other as possible and distributes the voltage evenly. This also minimizes inter-turn capacitance mentioned above.

The Numbers for Some Solenoid Baluns

Here are some numbers for suitable baluns

Form Dia.Coil Len.TurnsuHCoax Len Ft
4.5 in56 in**168246198
12 in11 in33239104
12 in18 in54448169
18 in7.3 in22254104
18 in11 in33461155

RG-8X, RG-59
Len Ft
4.5 in30 in124243146
12 in7 in2924491
12 in11 in46464144
18 in5 in2024794
18 in7 in29451136

Len Ft
4.5 in21 in107251126
4.5 in36 in185456218
12 in5.3 in2724685
12 in8 in41452129
18 in3.71924889
18 in5.327441127

Len Ft
4.5 in6.5 in6525176
4.5 in10.5105446124
12 in2.32324772
12 in3.333451104
18 in1.81826584
18 in2.424444113

** A single layer of RG-6 on a 4.5 inch form is impractical. One way to cut down on coil length and reduce the amount of coax in the balun is to wind a 2-layer coil. A conventional 2-layer winding that runs to one end, then reverses back over the first layer gives a neat-looking result but is a bad idea. It brings the first turn very close to the last turn and defeats the main reason for winding a solenoid instead of a toroidal or Collins balun: high arc-over resistance. However there is a quasi-scramble way to wind these that keeps most of the benefits while cutting the length of the coil and the amount of wire required nearly in half for the same inductance. Wind two or three turns on the coil form, then cross back and wind two or three turns in a second layer over the first three turns. Then put three more turns on the form, cross back and put a second layer on those, and so on. It's not as neat looking but it creates a 2-layer coil that keeps almost all the benefit of a single-layer coil.

Coax losses at 2.0 MHz are shown below. Values are for 150 feet and SWR of 1.5:1. Except for RG-174 miniature coax, losses at 2 MHz are so low they can be ignored.
RG-6A/U750.604 dB
RG-8X500.631 dB
RG-58500.788 dB
RG-59750.678 dB
RG-59B750.734 dB
RG-174501.74 dB

Collins Baluns

A Collins balun is very easy to make and consumes about half the amount of coax as a solenoid. However, unlike a single-layer solenoid it offers no special resistance to arc-over from lightning.

To make a Collins balun, simply wind the coax lightly on a round form, all in one spot, then slide the coax off the form and tape or wire-tie it into a ring. Unless the coax is very limp and pliable, an extra pair of hands from a helper is useful for this step. A slightly tapered form can be handy for sliding the coax off. Small plastic wastebaskets or barrels often have a good shape.

30 turns of RG-6, 12 inches in diameter, yields about 450uH and consumes 94 feet of coax (in contrast to 170 feet for the same inductance on a solenoid)

Toroidal Baluns

Toroidal baluns can also be used with these antennas. These antennas are receive-only so there are no issues of saturation or power handling. The target inductance is 450 uH so we want to use the highest permeability ferrite that will handle the frequencies of interest. Good old Type 43 is probably the best choice. It has a permeability around 850+ and handles 160, 80, and 40 meter frequencies. An FT-240-43 core yields 450 uH with 20 turns. The problem is the window of this core is 1.4 inches in diameter so it will only accommodate 12 turns of RG-6 (0.332" dia.). Stacking two cores reduces the required number of turns to 15, but that's still too many. Since the balun is really just an inductor, and inductors in series add together, we can get even closer to the goal with two 225 uH baluns in series. This gets the number of turns required down to 13 on each core which is still more than the 12 turns that will fit, but close enough.

On the other hand, building the balun with RG-59 (0.242" dia.) coax solves the window area problem because 19 turns fits easily through a 1.4 inch window. And only one core is needed instead of two, which may be a concern because FT-240-43 cores cost about 10 dollars apiece.

I hope these Beverage designs can be of use to other hams, make your installations easier to build, easier to maintain, cheaper, and more reliable.

73 de KW2P

Thursday, February 26, 2009

Google / Feedburner Mess, Please Update Your RSS Readers

As you probably know, Google bought Feedburner back in 2007. Recently Google "offered" to move my Feedburner account and feeds to the new Google / Feedburner. Since eventually everyone will have to move or their RSS feeds will stop working it seemed like a good idea. I pushed the button and all appeared to go smoothly and automatic. But it didn't. It broke all of my blog feeds, and since I maintain nine different blogs, it's a significant problem. Some feeds stopped updating, some began pointing to the wrong things, feeding comments out instead of posts, and other nonsense. It's a mess that I'm still cleaning up.

The Maya Paradise blog is now fixed, I think. The feed buttons on the Maya Paradise News Blog are updated and if you are a subscriber, please re-subscribe to the new feed before Google shuts down the previous feed system entirely on March 16th.