NQ4S - Delbert wrote:
I've run into a problem designing a compact antenna and could really
use some more experienced help.

Due to CCR I'm not allowed to have any permanent antenna structures
outside. Currently there are 2 HF and 2 VHF/UHF antennas crammed in
my attic. The HF antennas work to a point, but they are not overly
efficient. Beyond Europe and South America, most DX stations come
across as just a whisper.

Over the weekend I did some research on a compact antenna for the back
yard of my town home. The goal was to come up with something on 20M
which is compact, light, and easy to put up and take down.

While going through the ARRL antenna book my attention came across a
chart of field strengths for various combinations of phased arrays.
One that really had my interest were the ones at only 1/8 wavelength.
At 20M this is only about 8 feet, which is very doable with some PVC
pipe for a mount on top of a 30 foot telescoping pole.

You have to realize that those graphs assume zero antenna loss. While
that's a good assumption for a full size wire dipole, it's not a good
assumption for a Hamstick. More about this below.

Spending a couple of hours playing with EZNEC I figured out that an 8
foot separation between two dipoles running about 140-degrees out of
phase would give a pretty good gain at about a 30-degree takeoff angle
when mounted 30 feet off the ground. It looks like a really nice
improvement over a single dipole.

Being that a full dipole would run over 30 feet in length (longer than
my entire back yard), my natural idea was to use a couple of
hamsticks. I'm already using two hamstick equivalents for a dipole in
my attic. It works, and is nice and compact. Using two Hamstick
dipoles on the 8 foot boom would give a turning radius of around 8
feet. Very doable.


A way to guess at a Hamstick's loss is to model a wire that's the length
and diameter of the Hamstick. Model it as a vertical over perfect
ground, and look at the feedpoint resistance. Then, assuming that the
input resistance of the Hamstick without matching is about 50 ohms, you
can guess that the loss resistance is roughly the difference between 50
ohms and the resistance of your wire model. For example, if the wire
input R is about 10 ohms, then the Hamstick loss resistance is roughly
40 ohms.

Now make your model array with elements each made of a wire, a source,
and a resistor (load) at the source. Make the resistor value equal to
the loss resistance you got from the experiment above. If you want to
model dipoles, double the wire length and the loss resistor value. Run
EZNEC and note the gain. Save the model. Then delete one of the elements
and run it again, again noting the gain. The difference between the two
is approximately the gain of a two-element Hamstick array over a single
Hamstick element. Still think it's worth the effort?

Close spaced arrays with gain inevitably have relatively low feedpoint
resistance. This makes them more sensitive to loss, so it becomes
impossible to realize the perfect-antenna gain with lossy elements. A
particularly interesting array is the 90-degree spaced, 90-degree fed
cardioid, like in the EZNEC example files. You'll find that the gain of
that array over a single element stays constant regardless of how much
loss you add to the elements (as long as you add the same amount to
each, and always compare to one element with loss included). In that
particular array, one feepoint resistance goes up and the other goes
down, in such a way as to make the sensitivity to loss exactly cancel.
Most arrays don't share that interesting property, though, and
close-spaced arrays certainly don't. I'd be surprised if you got more
than a dB and a half or so from your array. You can, however, get a good
front-back ratio, and that in itself might make the effort worthwhile.
You'll have to be the judge of that.


Now we reach the point where I'm running into trouble. What would be
the best way to feed this antenna? My first thought was to run a
piece of coax up to the first dipole and then a piece from there to
the second dipole to provide the 140 degree delay. I'm running the
dipole in my attic without any sort of matching network, and it works
fine. Will running two dipoles in the same way be close enough to 50
ohms impedance or am I really missing the boat here?

You're missing the boat. You've got to read more of Chapter 8 of the
ARRL Antenna book, where you found the graphs. It probably is possible
to feed them with coax only, but not with a 140 degree difference in
length. See "The Simplest Phased Array Feed System -- That Works" in
Vol. 2 of _The Arrl Antenna Compendium_ if you want to do that. Examples
CardTL and 4SqTL furnished with EZNEC illustrate this method. There's
more information in the Antenna Notes files accompanying them. If you
don't want to use that method, you can use an L network as described in
the Antenna Book.


Another variant I've considered is running coax up to the top of the
pole and into a 1:1 balun to stop any RF coming back down the cable.
The balun would be close attached to the first dipole. From there a
piece of 300-ohm ladder could be used as a phasing line running from
the balun to the 2nd dipole. I'm less confident that this would work
due to the impedance mismatch.

Same problem.

My hope is to use a MFJ 33' telescoping pole and a cheap radio shack
rotator to hoist and spin this light weight antenna. Having the extra
gain from a phased array might finally let me work some of those far
off DX stations that I can never seem to hear.

One variant of the idea I'm considering turned up this morning. The
ZL-Special is two dipoles 1/8 wave apart with one dipole flipped over
and connected by a 1/8 wave connection. The combination of the flip
and the line provides an effective 135 degrees of phase delay. Very
close to what I came up with doing the computer modeling. The problem
I started seeing with the ZL-Special is that most designs show complex
matching networks or building folded dipoles to get the right
impedance.

You might take a look at the "Field Day Special" antenna. It's similar
to the ZL special, but there's software to allow you to actually design
one. You can get the info from ftp://eznec.com/pub/. Download fdsp~.exe.
It's a self-extracting ZIP file that includes instructions. I've been
using this antenna on several bands for many years, from home as well as
on Field Day. It works as designed.

I've pretty well hit the wall on my knowledge of matching antennas to
feedlines. The ARRL Antenna Book has been of help, but I feel like
I'm missing something in there that just isn't letting me finish the
idea.

Any information on this would be very helpful. Suggestions on how to
proceed would be most welcome.

Thanks.

73 de NQ4S

Roy Lewallen, W7EL