Generally speaking most stations will choose to use an antenna with full-wave spacing between the bays. However, in certain instances, a half-wave antenna may be more desirable. A six-bay full-wave spaced antenna has a gain of roughly 3.3, which is similar to the 3.1 gain from a 10-bay half-wave assuming circular polarization on both. The half-power beamwidth of the main lobe in the vertical plane for the full-wave is around 4.2 degrees compared to roughly 5.2 degrees for the latter style antenna. This seemingly insignificant span of one degree can result in much better coverage over several miles of real estate. The drawback is that the half-wavelength spaced antenna will cost more due to a greater quantity of materials, and will result in additional tower loading.
Due to their radiation pattern in the vertical plane, however, half-wavelength spaced antennas are desirable in cases where non-ionizing radiation levels on the ground are a concern. In addition, some of the antenna manufacturers prefer half-wavelength spaced antennas for directional antennas as the pattern is controlled better than its counterpart with lambda spacing. Of course other spacings are permissible, and in some instances the use of 0.7 lambda spacing for a directional antenna may work better.
There is no doubt that the FM band is becoming more and more crowded every day. The 2007 NCE filing window, for instance, saw the submission of some 3,600 applications. Although many wound up being dismissed for various reasons, the end result is a large number of new signals were added to the FM band, many of which were required to use a directional antenna. Some of the patterns submitted were downright strange, and quite frankly are not realizable with a run-of-the-mill side-mounted antenna.
In such cases it may be necessary to consider other designs such as panel or Yagi arrays. Panel arrays, because of their size, can add substantial loading to a tower by their sheer size and surface area. Yagi arrays on the other hand, may not cause as much loading to a structure, but can be more delicate in their construction, and suffer more from icing.
In some community antenna situations, the use of typical side mounted antennas is not practical due to the amount of available vertical real estate. In those cases more exotic solutions have been implemented. For instance, in St. Louis the community site there uses a combined antenna system consisting of a panel array with several layers. In this case, several stations have individual modules on a combiner spline feeding a single antenna. A drawback to this type of scheme is that an issue with the antenna or combiner will affect numerous stations in the market simultaneously. By contrast Willis Tower in Chicago accommodates numerous FM stations through a stack of cavity-backed resonators. Under this scenario each station has its own antenna, thus a failure does not have a widespread effect.
As previously mentioned, half-wave spaced antennas typically have lower downward radiation resulting in a lower power density at ground level. The ring-stub type antenna tends to have the greatest downward radiation component, and will have the greatest chance of exceeding the exposure standards. The Commission uses this design as a benchmark in their RFR analyses. If a proposed site passes the exposure criteria with this style, then there typically will not be a condition on the construction permit requiring measurements at the site. The double-V and roto-tiller style antennas have lower downward radiation, and specifying one of these designs may allow you to skate by the measurements.
In the end, the antenna can be one of the most costly components in the system to replace. Not only do the base material costs have to be considered, but factoring in the costs for a tower crew must be considered as well. It is not unusual for quality antennas, if well maintained, to last 30 years or more. True some of these antennas tend to be more expensive, but attempted cost savings by doing it on the cheap has a way of coming back to nip you when least expected.
Ruck is the principal engineer of Jeremy Ruck and Associates, Canton, IL.