Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.
This is a long story, but I will try to keep it short.
When you look to a HW dipole, you see two quarter wave transmission line sections with the source in the middle. The "loss" due to radiation causes the impedance between the two, quarter wave pieces to be non-zero.
If there were no loss, the impedance would be zero (property of quarter wave transformer, infinite load at end, gives zero input impedance).
For a half wave dipole, the input impedance at resonance (so electrically two quarter wave pieces of metal) is about 50…70 Ohms. The Z0 of the conductors is significantly higher (several hundred of Ohms).
You might know that the Q factor of quarter wave open-end resonator is proportional to Zo/Zres. Zo is cable impedance, Zres is impedance at resonance.
To make the dipole more wideband (=lower Q), you have to lower the Zo of the quarter wave pieces. This can be done by making them thick (with respect to wavelength). At a certain point, the Q will be that low that you get a more or like traveling wave pattern instead of a standing wave pattern.
So wide band antennas are by nature wide structures.
When you want to feed your very wide structure, you will get a step transition. Step transitions you can use in narrow band matching, but not in wide band matching. For wide band matching you need to distribute your steps, or even better, use tapering. You have to create smooth impedance transitions over certain wavelength along the line.
A ground plane under or close to your patch tries to cancel the radiation. Hence the wanted "loss" due to radiation reduces. So the input impedance (in case of a HW dipole) will reduce. The resonance impedance reduces more rapid then the reduction in Zo (because of the groundplane). So adding a ground plane almost always results in less bandwidth.
At the lowest frequency, a monopole UWB antenna behaves as a quarter wave radiator. At higher frequency it behaves more like a traveling wave antenne where the wave loses sufficient energy so that after back travel (towards the source), it is so weak that you will meet your VSWR/Return Loss requirements.
One can increase the loss due to radiation (this is the desired affect) by also bending the radiator (in fact a 3D taper). This technique is used in tongue-like UWB antennas that have some directivity.
Now we come to directivity.
You can have a good impedance behavior, but the radiation pattern versus frequency may vary significantly making your antenna less effective. So you have to study the amplitude and phase of current distribution to shape your antenna to get the desired radiation pattern over wide frequency range.
So in just some words: Wide structures, not close over a ground plane and make sure you have smooth impedance transitions.