By Mark Connelly,
an electrically-short dipole presents a much higher impedance
than 50 ohms?
-- Read more --
hcdx mailing list, April 8, 2001
A dipole antenna normally is cut for approximately one half of a
wavelength. The customary formula is (468 / F) feet or (142.65 /
F) meters where F is the frequency in MHz. This works out to about
0.48 wavelength. The dipole antenna has seen little use at medium
wave because of its impractically large dimensions: about 883 feet
/ 269 m at the low end of the broadcast band (530 kHz). Furthermore,
for the antenna to perform well, it has to be a minimum of a quarter-wave
above the ground if mounted horizontally. If set up vertically,
its top end is a half wave (or greater) above ground and if sloped
at 45 degrees, the top end is at least 0.35 wavelength high.
In a quest for a compact omnidirectional passive antenna of reasonable
dimensions, I did some measurements on vertical and sloping dipoles
measuring only 33 ft. / 10 m total length (16.5 ft. / 5 m each side
of the center feed point). This antenna is less than 0.02 wavelength
at 530 kHz and 0.06 wavelength at 1700 kHz. Obviously, in this frequency
range, such an antenna is a very poor match to direct 50 ohm feed.
The way to get a reasonable amount of signal capture out
of such a compact antenna, without resorting to distortion and noise
producing amplification, is through the use of passive devices to
improve the match. A narrowband technique (a remotely adjustable
inductive - capacitive (L-C) tank at the feed point) is ultimately
the best approach in terms of efficiency, but it is difficult to
implement and does not lend itself to frequency-agile DXing.
Matching with a broadband balun transformer is the other approach
that can be used. The ratio to be used is fairly high because an
electrically-short dipole presents a much higher impedance than
50 ohms. I did tests with two transformers: one had a 36:1 impedance
ratio, the other a 16:1 ratio. The high impedance winding leads
went to the antenna elements and the low impedance winding to the
coaxial cable. I also had 9:1, 4:1, and 1:1 transformers available
for testing if results indicated the 16:1 better than the 36:1 unit.
This, however, was not necessary as the 36:1 transformer gave the
better results across the medium wave dial. I set up a 10 m (total
length) center-fed dipole in a nearly-vertical configuration suspended
by a nylon rope over a high branch on a black locust tree in the
backyard at my home location in Billerica, MA (GC =71.221 W / 42.533
N). The bottom of the antenna was about 1.5 m off the ground and
the coaxial feed ran about 20 m from the balun box (at the middle
of the antenna) into the house: I tried to keep the feedline close
to a right angle to the dipole as much as possible. Daytime signal
measurements on some groundwave locals were made using the Drake
R8A in PREAMP ON mode. Across the band, the 36:1 transformer showed
a 4 to 8 dB advantage over the 16:1.
Measurements with the 16:1 transformer (Mini-Circuits T16-6T-X65)
Measurements with the 36:1 transformer (Mini-Circuits T36-1-X65)
Homebrew transformers of various ratios may be tried in future
experiments. One possible design is a 25:1 transformer consisting
of a 40 turn high impedance winding on the opposite side of an FT-140-43
toroid from an 8 turn low impedance winding. The larger core than
what's used in the Mini-Circuits models would tend to reduce the
likelihood of harmonic and intermodulation distortion in strong-signal
areas. With the Mini-Circuits transformers I used, I did notice
a slight amount of WRKO-680 audio under WLYN-1360 (680 * 2). WRKO's
50 kW transmitter is about 5 km from my home location.
The short dipole picked up less local electrical noise (relative
to desired signals) than an active whip. If a 1:1 isolation transformer
is used at the "shack" end of the coaxial cable, noise can be reduced
a bit more in some circumstances. A broadband loop (square loop
at 3 m per side, coupled through 1:1 transformers on each end of
the coaxial feedline) was still quieter than the dipole in terms
of locally-produced TV / other electrical noise pick-up.
Signal levels produced by an efficiently-coupled short dipole are
adequate for typical DXing. When more gain is needed, a good high-Q
regenerative preselector amplifier will bring marginal signals up
out of the mud.
The primary niche that the short dipole has is a very space-conserving
antenna for small pieces of land. A broadband loop (as described
above) can be located close to the dipole and the two antennas can
be phased to produce a cardioid pattern. When more land is available,
you can phase two vertical dipoles spaced at approximately 60 m
on the desired peak-null axis.
Sloping the short dipole off true vertical only has a slight
effect on directivity. An antenna of such small dimensions does
not have much in the way of inherent nulls. If setting up two of
these antennas about 60 m apart for phasing, it might help to slope
one the opposite way of the other to make the antennas a little
"more different". Setting the antenna up in a horizontal position
probably does reduce the low angle pick-up some.
If a tree or other support much taller than 10 m is available, the
dipole can be scaled up in size. With a 20 m antenna length, the
transformer ratio can probably be lowered to something in the 4:1
to 12:1 range. As you approach a half wavelength, the required ratio
goes down to 1:1. The closer the size of the antenna gets to resonance,
the greater the sensitivity. Still I was quite surprised at how
much signal can be obtained from an electrically short dipole when
the correct matching transformer is used.
Ngai: I don't quite understand your article. Why an electrically-short
dipole presents a much higher impedance than 50 ohms? In fact, from
theory, small dipole has 0.2 ohm typically. Or probably you may
include matching to get the high impedance. Please clarify.
As an antenna gets progressively shorter, its dominant
impedance characteristic is that of a capacitance which gets smaller.
A short antenna may look like a 100 pF capacitor (shunted by a very
high, often negligible, leakage resistance value). An even shorter
antenna may look like a 25 pF capacitor: the corresponding reactance
term 1/(2*pi*f*C) would then be 4 times that of the other antenna
at the same frequency. In the limiting case (infinitely small elements),
your feedline is playing into an open circuit, not a short.
Tests here recommend
the use of an L-C tank circuit between the antenna and the 50-ohm
feedline for the absolutely best match. This is, of course, a narrowband
approach. Small changes in operating frequency necessitate retuning.
In receive-only applications, somewhat less efficient but more broadbanded
matching schemes may be preferable.
matching of an antenna considerably shorter than a halfwave
dipole, optimum signal transfer occurs with RF transformers (binocular
or toroid core) having the higher number of turns (high impedance
winding) connected to the two antenna leads and the lower number
of turns (low impedance winding) on the coaxial cable leads: center
and shield. A center-fed vertical dipole of total length 10 m
used at 1 MHz (300 m) has an electrical length of 0.033 wavelength
or 12 degrees. A stepdown transformer in the 25:1 range gives
best 1 MHz transfer to 50 ohms in several tests run here. This
would put the antenna's source impedance in the general range
of about 25*50 = 1250 ohms.
antenna looked like a 0.2 ohm source as has been suggested,
I would have the best match with the lower number of transformer
turns going to the antenna and the higher number going to the
coaxial cable feeding the receiver. In reality if I do this, there
is almost no signal transferred. Even with a 1:1 transformer,
which would be fine for an antenna in the halfwave overall length
range, signal transfer is very low. I have to use a transformer
similar to what would be used with a Pennant or Flag antenna:
high impedance winding to antenna wires and low impedance winding
to the coaxial feedline.
Connelly, WA1ION, USA, 24 August 2004