Last month's Essential Wi-Fi touched on the issue that signal range is affected by the type of antenna that is being used. No antenna puts out energy equally in all directions, so understanding the coverage pattern of the antennas that you use is crucial to ensuring that your users have good coverage. Being comfortable with the many types of antennas that are available, and being able to choose which antenna is right for each application is one of the marks of an expert RF system designer. This month, we'll go over some of the common types of antennas on the market and discuss their coverage patterns and typical applications.
In order to discuss antenna coverage, first, we must define horizontal and vertical beamwidth. A full definition of these terms is outside the scope of this month's column, so suffice it to say that an antenna's beamwidth describes a vertical and horizontal arc in which the antenna's signal will be strongest. Signal is still transmitted outside of the antenna's beamwidth, but it is weaker than signal within the antenna's beamwidth. Horizontal beamwidth refers to the width of the antenna's beam along a plane that is parallel to the horizon. Vertical beamwidth refers to the height of the antenna's beam along a plane that is perpendicular to the horizon. Keep in mind that, since beamwidth is expressed in degrees, the actual size of the antenna's coverage area increases as you get further from the antenna--just like the circumference of a circle increases as its radius increases.
Horizontal and Vertical Beamwidth
Consider the direct signal path from the middle of the antenna straight out to the point where the two lines for vertical and horizontal beamwidth meet in the picture above. With reference to any point surrounding that straight path it should be easy to imagine how the signal strength is always strongest when you're closer to the straight path. If you were to move away from the center, straight path at a 90-degree angle there would be a point at which the signal strength was reduced by half. It's that point, the Half Power Beamwidth (HPBW) point, that is used to define the beamwidth angle. Technically the "beamwidth" is the "Half Power Beamwidth."
Omnidirectional Antenna Signal Propagation
The signal is strongest in a plane perpendicular to the axis of a dipole antenna. Above, the green volume represents dense (strong) RF signal energy.
The "rubber duck" antenna that comes by default on nearly all access points is an example of a dipole antenna. A dipole antenna is omni-directional, which means that it has a 360 degree horizontal beamwidth. Put another way, if you imagine a disc with the antenna stuck up through the middle of the disc like the axis of a pinwheel, the signal will be equally strong anywhere on the disc.
So an omni-directional antenna puts out the same amount of energy anywhere along a 360 degree horizontal plane around the antenna--but it does not put out the same amount of energy above and below that plane. Typically, an omni-directional antenna puts out most of its energy along a horizontal plane around the center of the antenna and signal power decreases as the client moves above and below that plane. The signal is weakest directly above the antenna (looking down at its point) and directly below the antenna (looking up at its base). This means that an omni-directional antenna has a vertical beamwidth that is less than 180 degrees (the maximum possible theoretical value). The vertical beamwidth of an omni-directional antenna determines how much of its energy is transmitted above and below its central horizontal plane. Antennas with a high vertical beamwidth (a wide beam with lots of coverage above and below the antenna) are called low gain, and antennas with a low vertical beamwidth (a narrow beam with little coverage above and below the antenna) are called high gain.
When the term "gain" is used relative to an antenna it implies that the construction of the metal elements in the antenna cause the electromagnetic signal to propagate more in some directions than in others. It's not that the overall power increases with a "high gain" antenna, it's that the energy is focused into a smaller area in space and is, therefore, "denser".
Low gain omni-directional antennas are preferable when users will be located above or below the antenna and when they will move closer to and further away from the antenna. This is why low-gain dipoles are usually the default on access points: the assumption is that users will be located at an unknown position relative to the AP and they will be moving. The AP's low-gain omni-directional antenna provides the most equal coverage to all locations around the AP.
High gain omni-directional antennas are usually preferred when both ends of the connection are stationary and at roughly equal heights, such as a link between the roofs of several equal-height buildings. Although high-gain antennas have less vertical beamwidth, they give this up in exchange for increased range, so high-gain omnidirectional antennas might be the only way to make a longer-distance link work. In many cases, a high-gain antenna's low vertical beamwidth is a liability--requiring more precise alignment, for example--but in some cases, it is an asset. For example, the antenna's low vertical beamwidth means that the transmissions are more focused and therefore a little bit harder to spy on and less likely to interfere with other transmitters.
The dipole antenna is the only type of omni-directional antenna that is in wide use in 802.11 today. Another type of omni-directional antenna that can be found is called a slotted waveguide, but they are much less common than dipoles in 802.11 applications. You should also be aware that some vendor's specifications refer to an antenna as being "omnidirectional" when, in fact, it is "almost" omnidirectional. That is, the antenna may have close to equal power in all directions but there may be "notches" in the coverage pattern that will make correct installation orientation an issue.
One of the keys to designing and troubleshooting an 802.11 network is understanding how RF signals propagate and how they can change as they do. This month, we'll look at various ways in which an RF signal can be transformed as it travels through the air, and what this means for network engineers.
Attenuation is the most basic way in which an 802.11 signal changes as it is transmitted. As the signal travels away from the antenna, its wavefront expands and its energy is spread out over a larger and larger area. As a result, the amount of energy in any given area decreases. As an analogy for this, think of a balloon being blown up. The surface of the balloon represents the RF wavefront. As the balloon expands, it gets thinner. This is analogous to the RF energy being weaker the further you get from the antenna--it's not that there's really less energy, but that the energy that's there is spread "thinner". It's important to keep in mind that this type of attenuation is inherent to signal transmission and would occur even if the signal were transmitted in deep space, where nothing else could interact with it. This type of attenuation is referred to as Free Space Path Loss, or FSPL.
Absorption refers to attenuation of the signal as it passes through matter. Some amount of absorption happens as the signal travels through air molecules, but the most significant RF absorption occurs when the signal passes through metal objects or objects containing lots of water (water is highly absorptive to 2.4 GHz RF energy). Examples of metal objects include sheet metal walls, metal stairwells, metal studs in walls, and the metal reinforcement in concrete rebar construction. Examples of water-containing objects include aquariums (obviously), deciduous trees and other broad-leafed plants, paper (when stacked such as in a paper mill or in filing cabinets), and last but not lest, people. Absorption is usually a function of the type of object and the thickness of the object. The amount of absorption for an object can be measured by placing an access point near the absorbing object and measuring signal strength on both sides. Once you know the absorption of common objects, you can more accurately calculate your system operating margin.
Obstructions will absorb some of the RF signal energy and reduce the strength of the propagation wave
The Effect of Frequency on Signal Absorption
You may have heard that 802.11a doesn't have as much range as 802.11b or 802.11g because 802.11a operates in the 5 GHz (Gigahertz) frequency band. The reason that higher frequencies (like 802.11a in the 5 GHz band) are more severely degraded when passing through an obstruction than lower frequencies can be understood by studying the picture to the left.
Here you see two signals passing through an obstruction (perhaps a wall or a door). One of these (the black line) is a lower frequency signal (imagine 802.11b or 802.11g) and the other is a higher frequency (imagine 802.11a or 802.16 WiMAX). Notice that the signal energy must change from low to high more often for the higher frequency. The lower frequency has gone through a single cycle: from the zero-point (on the left), to its maximum value, down to its lowest value, and back to the zero-point on the right. That's one single cycle (of which, by the way, there are 2.4 billion per second in 802.11b and g). The higher frequency has gone through three cycles in the same space.
As the signal propagates through the obstruction some of the energy of the electromagnetic field is transferred into the medium from which the obstruction is built (wood, drywall, trees, etc).
You can imagine that the signal energy must "wiggle" around; up and down as it passes through the obstruction. Both signals travel at the same speed (the speed of light in the medium) so you see that the higher frequency signal must "wiggle" around more during the time it's inside the obstruction. Hence, more of its energy is lost to the obstruction. Higher frequencies are absorbed more than lower frequencies in any particular medium.
Reflection refers to a signal bouncing off of an object. We're all familiar with reflection of light from looking in a mirror and anyone who has tried to make a bank shot in a pool game has also dealt with reflection. RF signals experience reflection too, although it's not always obvious what types of objects will absorb the signal and what types of objects will reflect it. In general, large, flat metal objects, such as sheet metal sides of hangars, cause reflection.
Metallic objects, such as tinted glass on skyscrapers, can also cause reflection, and the flat surfaces of still lakes are also well-known as causes of reflection. Reflection is not inherently bad for the signal, and in some cases, it can create coverage in areas that wouldn't otherwise have it (by bouncing the signal around an obstruction), but if reflection creates multipath cancellation, it will do more harm than good. In general, RF network designers should attempt to avoid reflecting objects if at all possible, if only because of their unpredictable effect. In point-to-point links, reflection can be minimized by using high-gain antennas that focus their energy tightly; in indoor environments with omni-directional antennas, reflection really can't be avoided, but it tends to help more than it hurts by creating coverage in areas that don't have line-of-sight to the access point.
The direct path and the reflected path are actually different lengths and this is the basis of "Multipath Fading" - the two signals can cancel each other when they arrive at the receiver.
Attenuation, absorption, and reflection are the most easily-understood types of RF signal distortion. Next month, we'll discuss less intuitive ones: diffraction, refraction, and scattering.
The promise of the 802.11n standard with its 100 Mbps data rates is making some decision makers wonder whether they should wait for newer equipment before they implement Wi-Fi today. The answer is a resounding NO. And here's why...
First of all, there are some "Pre-N" products on the market today. None of these manufacturers are willing to say that what they ship today will be able to work with the final 802.11n standard because that standard isn't yet finalized. Moreover, there's a possibility that some of today's pre-N chipsets won't operate at all above 802.11g speeds with the coming breed of true 802.11n systems. So, don't go with the vendor propreitary "Pre-N" equipment for commercial, business applications.
The 802.11n standard is, as of today, not expected to be finalized until November 2006. The Wi-Fi Alliance has an excellent FAQ that details much of where the 802.11n marketplace is today; you should read this FAQ.
Here's the bottom line: The 802.11 system you install today will be providing the same levels of service in two or three years. That means that, for example, an office that uses 802.11g (at 54 Mbps with an effective throughput of around 20 Mbps) today, or a public access HotSpot that provides DSL-level speeds to multiple users today, will still be providing those same levels of service in 2 or 3 years.
Just because newer, higher-speed standards are coming in 12 or 18 months doesn't mean your current investment is going to have be thrown out or that your equipment will have to be replaced. Think about the fact that it took almost 10 years for Twisted-Pair Ethernet to displace its coaxial predecessor. Think about how long folks still had their Windows 95, and the list goes on. All of the Intel Centrino notebooks and other 802.11b equipment that's out there today is not going to get thrown in the trash just because 802.11n comes out in 2006 or 2007. And, 802.11n is BACKWARD COMPATIBLE with 802.11b so even then the newest equipment will continue to work on today's Wi-Fi networks.