I live in a home owners association with antenna restrictions. Dave Kelley,
owner of CCDAntennas.com, is a friend of mine. After hearing my complaints about
trying to use my roof flashing for an antenna he suggested I try a CCD antenna
(by the way I did make contacts on the roof flashing). It was
pretty brave of him because he knows I never take anything technical at face
value. I have a burning desire to understand the theoretical and practical in
everything I do. Dave's antennas are made to withstand the full legal power limit with thick wire and large capacitors. I had to hide my antenna and I only run 100 watts. So to make it as invisible as possible, I decided to build my own CCD with very thin wire and surface mount capacitors. I quickly discovered the formulas available on Dave’s web site were not intended for the small gauge insulated wire I was using. After building a few 40 meter CCD antennas that didn't work quite right I decided to play with antenna modeling software. It was when computer modeling the CCD antenna I discovered its potential to resonate on two adjacent ham bands. After hours of computer modeling and much wasted wire I had a 20/40 meter CCD made with 24 gauge insulated wire strung between the peaks of my roof and the block walls. It matched pretty well on 20 and was by no means perfect on 40 meters. It was, however, well within my tuner's range. With more computer modeling and less wasted wire (I was getting better at the modeling) I had a 22 gauge, dual band, 20/40 meter design that has an SWR under 1.5:1 all the way across both bands. I currently use this antenna on a 24 foot pushup pole while camping. Some more modeling with larger gauge wire and I had full legal limit designs for a 40/80 meter CCD and a 20/40 meter CCD. But there was no way I was able to test the designs on my little city lot. Dave built both of these designs on faith (you have to remember these antennas are not the easiest to build). When he put them in the air they both worked even better than modeled. They are now part of his standard inventory at CCDAntennas.com. 
In order to understand the dual
band CCD you need a little background in how the standard model works.
Everyone has seen series inductance used to lower the resonant
frequency of a short antenna. The screwdriver antenna is the first example
that comes to mind. But what happens if you introduce series capacitance? As you
might expect adding series capacitance raises the resonant frequency. A standard
length CCD is a full wavelength of wire with enough capacitance evenly
distributed throughout the wire to double its resonant frequency. Notice I did
not say “full wave” but said “standard length.” The capacitive reactance
redistributes the current so the longer wire acts like a half wave at the design
frequency. In my discussion of the dual band CCD I am going to use a simple visualization and talk about the capacitors “shortening the wire” much the same way some of us think about inductors lengthening the wire. So using this simplification we say the standard length CCD uses enough capacitance to “shorten” a full wave dipole to resonate like a half wave dipole. 

Current Distribution 
A normal dipole will resonate with a useable feed impedance at odd multiples
of its design frequency. A 40 meter dipole will be useable on 15 meters. This is
because the antenna becomes a wavelength and a half long. As long as a dipole is
an odd multiple of half wavelengths long the center feed point will be at a current
peak and exhibit a low impedance that is easy to match to your feed line and
transmitter. A CCD antenna behaves differently as you raise the frequency. Remember the capacitive reactance is “shortening” the wire. However as the frequency goes up the capacitive reactance goes down. You can visualize this as the capacitors slowly disappearing as the frequency goes up and the wire “growing” to approach its real length. I noticed on the computer models that this effect is large enough to place the second SWR dip of a standard length 40 meter CCD at 12.5 MHz. This is between the 40 and 20 meter bands! This observation made me wonder, if a 40 meter standard dipole has a second dip at 21.5 MHz and a standard length CCD has a second dip at 12.5 MHz could I find a length of wire in between that has the second dip at 14 MHz? The answer is yes. If you take just the right length of wire and add just enough capacitive reactance to resonate it on 40 meters it will have its 3rd order harmonic resonance smack in the middle of the 20 meter band. Some of the benefits of this design are:
Comparison to Single Band CCDsOn the lower frequency, or primary band, the dual band design behaves almost exactly like the single band design. The differences in the computer models are so slight they should be undetectable in real world operation. On the next band up the major difference is in the radiation pattern. The dual band design is operating at 1 ˝ wavelengths while the single band antenna operates like a half wave dipole. Because of this the dual band design no longer gives you the characteristic horizontal oval pattern of a have wave dipole. It exhibits a multilobe pattern as you would expect of a 1 ˝ wavelength radiator. For those of you that like to analyze radiation patterns here they are: Remember that the height above ground and amount of V angle will significantly affect the radiation pattern of any dipole. For all the patterns above I used 15 meters (about 49 feet) above ground. Other height and angle combinations can be used to favor a particular type of operation. For example here is how you might optimize this antenna to improve its 20 meter DX performance:

