80 m and 40 m
80 40 and 20
20 and 15
There are a number of options, and the choice among them will depend on your individual circumstances.
Some of the more likely ones are:
Multiple dipoles on a common feedpoint. This is the simplest to set up and tune (as long as the wires
are not too close together.) Can be set up to cover any combination of bands. If all the wires share
the same end supports just tie them at different heights or let the shorter ones droop below the longer
ones. This is what I use for all my portable antenna kits - in some cases the wires were originally tuned
over 30 years ago and I have used them in many different configurations without a tuner.
Center-fed doublet with ladder line or twinlead to a tuner. To keep the broadside pattern the overall
wire length must be less than about 1.25 wavelengths ("extended double zepp"). Within that range,
the longer the wire the higher the gain or efficiency. So in each of the cases you give, you can just
calculate 1.25 wavelengths for the highest band and use it on the others. Note that in the case of
80 / 40 / 20m the wire will be about 0.3 wavelengths - this is getting pretty short, and is approaching
the point where the efficiency drop becomes noticeable. This approach requires a good antenna tuner
and readjusting it each time you change bands. Here is how W5DXP made this approach work for a
dual-band 80 / 40m antenna without a tuner:http://www.w5dxp.com/HEDZ.htm
Trap antennas: reasonably simple to install once tuned. Can use traps made from coaxial cable or
coils without discrete capacitors, depending on the design. Here is an example of a 40 / 80m version:http://vk1od.net/antenna/coaxtrap/8040.htm
Either of the dual-band antennas can be built this way with a trap frequency between the two bands,
which minimizes losses. The tri-band version requires more experimentation to do that way (probably
using a modeling program) because there are more interactions, but even using resonant traps the
efficiency is still pretty good.
A combination of the above options: My current 40 / 80 / 160m antenna system uses two dipoles
on a common feedpoint, one for 80m and one for 40m with loading coils on the end to operate 160m
in about the same length as the 80m wires. The same approach can be used for 20 / 40 / 80m.
Or you can use a ladder-line fed doublet that is long enough for good 80m efficiency, but with a
pair of 20m traps to prevent the pattern from breaking up on that band.
A TRUE fan dipole: by using multiple wires of the same length spread apart as a fan on each side
of a dipole you can get a low SWR over a very wide bandwidth. For example, with 5 wires each
40' long over about a 25 degree span (from memory) on each side of a 4 : 1 balun the SWR is
below about 2.5 : 1 from 40 through 10m. While the match isn't perfect, it should be well within
the range of most built-in tuners. The pattern breaks up beyond a certain length, just as it does
with a regular antenna, but it should work over a 2 : 1 frequency range. Such a dipole could
cover, for example, 20 / 17 / 15 / 12 / 10m with reasonably low SWR.
Then there are the more creative approaches, especially allowing the antenna to be 2-dimensional
(or more) rather than just a straight wire. A vertical full-wave loop on one band can be used as
a bisquare on the second harmonic by putting a quarter wave open stub at the top. (This will
probably need ladder line feed.) A folded dipole for the lower band can be twisted into a pair
of phased loops for the upper band - I've done this horizontally for NVIS on 80 / 160m and it
has matches well with a 4 : 1 balun. One of the more creative methods would be to build a
doublet using 102' of wire, letting 17' hang down at each end and attaching about the same
length of ladder line to the center. This will load up in the normal manner on 40 and 80m (just
like a G5RV if you switch in a bit of extra twinlead.) On 20m you can short the ends of the
ladder line and feed it as a bobtail curtain against ground. (In fact, you could try using that
extra piece of feedline you used on 40 and 80m as a quarter wave zepp feed on 20m...)
The bands are close enough together for the 20 / 15m antenna that something like a lazy-H
or other broadside array could work on both bands. The simplest would be to build a pair of
1.25 wavelength elements for 15m stacked one above the other (say at 60 and 30' above
ground) and feed them in parallel with equal lengths of ladder line. That would give you a
reasonable amount of bidirectional gain on both bands. Even a single (or double) 20m full
wave vertical loop will have a bit more gain on 15m without the pattern breaking up.
I'm sure we can come up with a few options given the chance...