The subject of antennas is often relegated to a mere footnote in many enterprise wireless networking discussions. Yet it is the antenna - the conductor that converts electrical signals to radio waves for transmission and back again upon reception - that can optimise wireless application performance by altering how the signal propagates.

Omnidirectional antennas, for example, broadcast a signal in all directions. Certain types are used in outdoor point-to-multipoint bridging, while others are used in cellular and cordless phones, two-way radios, and for AM/FM radio broadcast reception.

Directional antennas, by contrast, are designed to focus their "beam" in a specific direction. Directional antennas have a number of sub-types. Here are a few and how they are often used:

  • Yagi: Point-to-point outdoor bridging across medium-range distances.
  • Sector: Point-to-multipoint outdoor broadcast to recipients in a certain direction.
  • Parabolic: For outdoor use, has an emitter mounted such that it is aimed into a bowl-shaped reflector, which focuses the signal into a tight beam. Used in terrestrial-to-satellite applications.
  • Circularised polar: Used indoors where a lot of reflective material, such as metal, shelves, and water, is present.

During antenna discussions, you might frequently hear the term "gain."

Gain refers to the ratio of the output signal strength of an amplifier to the input signal strength. This ratio is usually expressed in terms of decibels (dBi). In wireless LANs, the higher an antenna's gain, the higher its price usually is.

Some systems let you adjust the signal strength of your access points to shrink your coverage cell for better throughput; some now also automatically adjust depending on network conditions (part of a function called "radio resource management").

The term "antenna diversity" in WLANs traditionally refers to the use of dual antennas on a given access point to help relieve the effects of multipath interference. As you likely know by now, however, forthcoming 802.11n-standard WLANs use at least four antennas to actually capitalise on the effects of multipath to boost transmission speeds over 100 Mbit/s.

Radio signals are electromagnetic waves, and polarisation refers to the direction in which the the electric and magnetic fields fluctuate. These fields fluctuate in a direction perpendicular to the direction the wave travels (much like the motion of the water surface in a water wave).

Fairly obvious, if a wave comes from a straight antenna, it will be polarised so its electric field fluctuates in the direction parallel to the antenna. Also obviously, if it's detected by the same sort of antenna, the antenna will get the best signal if it is aligned with the polarisation of the wave (the electric field can then have the biggest effect on the electrons in the antenna, and generate the biggest current up and down it).

Given all this:

  • To form a wireless communications link, the polarisation of two communicating stations must match. This maximises received signal strength, while a mismatch will degrade signal strength and performance.
  • Cross-polarisation design strategies might eventually play a role in circumventing interference from neighboring systems as the airwaves grow increasingly populated.

Most antennas radiate either linear (including horizontal and vertical) polarisation or circular polarisation. Vertically polarised antennas' electrical field is perpendicular to the Earth's surface.

Horizontally polarised antennas have their electric field parallel to the Earth's surface. So a straight-wire antenna will have one polarisation when mounted vertically and a different polarisation when mounted horizontally, because its relationship to the Earth's surface changes with the different mounts.

Circular polarised antennas, by contrast, radiate energy in both the horizontal and vertical planes and all planes in between. Polarisation is also affected by reflections.

If one station has vertical polarisation, others that it communicates with must also have vertical polarisation. If the station has a horizontal polarisation, its mate must, too. Same with circular.

When two antennas do not have the same polarisation the condition is called "cross-polarisation." This can actually be beneficial to your design strategy in some circumstances.

If the antennas of link A are cross-polarised to the antennas of nearby link B, where links A and B aren't meant to communicate with each other, interference between the links is reduced and possibly prevented.