I recently completed my Ekahau Design training and certification. During the first day of class the instructor went over wireless fundamentals. Now as an experienced wireless engineer most of the topics were ideas that I had already known, however there were a few idea’s that changed the way I understand wireless for the better. Lets walk though each of these idea’s together!
Channel width means nothing if you don’t have the available airtime to take advantage of it.
For those who don’t know, channel width is the range of frequencies a wireless devices is listening on. Generally, the wider the channel width the higher bandwidth you can achieve. Now in a lab environment where you have one client connected to a single access point this is absolutely true, however in the real world we have multiple clients connecting to one of many different access points. Because wireless is considered half duplex, meaning only one device can transmit data at a time, the amount of time any one client has to transmit data is shortened as more clients join that channel.
Many customers of mine have increased the channel width, thinking that doing this would increase the wireless throughput, but saw a decrease in throughput. They failed to take into account the number of clients per access point, and/or if increasing the channel width would cause co-channel interference. If increasing channel width causes co-channel interference with neighboring access points then the number of clients on that wider channel has increased, and reduced the total airtime for any one client. While a client can now take advantage of higher bandwidth when it can transmit data, it now has to wait longer for it next turn to transmit, lowering the overall throughput.
2.4ghz and 5ghz bands travels the same distance in free space.
I have always been told, and believed, that the 2.4Ghz band travels farther than the 5Ghz band. So I was shocked to learn that both frequencies travel the same distance, but are perceived differently by the receiving radio antenna. Not surprising, each band has a different frequency length. The length of the 2.4Ghz band is more than twice the length of the 5Ghz band.
As a way of optimizing the antenna for each band, vendors choose an antenna length that is proportional to the frequency length, which results in the 2.4Ghz antenna being twice the size of the 5Ghz antenna. The surface area for the 2.4Ghz antenna is twice that of the 5Ghz antenna, so it can pick up a weaker signal on the 2.4Ghz band and perceiving a stronger signal farther than the 5Ghz band. This would be like taking a sheet of plywood and a 2×4 into a strong wind. The speed of the wind doesn’t change, however holding the plywood sheet you feel a lot more of the winds power than if you held the 2×4. Surface area is key. This is also why laptops tend to have better signal than mobile phones, as their antenna tend to have a larger surface area.
So what is the point in knowing this? Well each of these antenna has an offset that we can use to tune the power levels on the access point to make both bands behave similarly. On average the 5Ghz antenna has an offset of -6dBi compared to the 2.4Ghz antenna. We can compensate for this offset by either increase the 5Ghz radio or decreasing the 2.4Ghz radio by 6dBm.
Antenna’s don’t add to the power, but focuses the power to send the signal further and can receive a weaker signal back.
This one may seem wrong to someone who looks at the datasheets for wireless antenna, as all antenna advertise additional gain. What is important to understand is that this gain is passive gain rather than active gain. The antenna passively focuses the energy it is given by the access point rather than actively adding to it. To illustrate this, lets take into account the following scenario.
Say we have an outdoor access point with an omni-directional antenna that covers an outdoor eating area. We are getting reports that users can’t connect to the wireless on the far side of this eating area. Our first instinct would be to increase the power level on the access point, however that won’t solve the problem. The reason is because access points can easily increase power, however the receiving devices (that is often battery powered) doesn’t have enough power to send the data back. We create an asymmetrical link, where the client can receive data from the AP, but can’t send data to the AP.
What is the solution? The solution is to change out the low gain omni-directional antenna with a higher gain directional antenna. The more focused antenna will focus the power from the access point so that the client can detect the signal, and it can better amplify the weaker return signal coming back from the device. A good way to think of this focused amplification is with one of those cones used to amplify someone voice. Using one allows your voice to be more focused and travel farther, and by placing that cone around your ear, you can hear the response a lot clearer as well.
Placing access points in hallways leads to a worse wireless experience.
When deciding where to place access points it is important to consider the wireless capacity of the network not just the coverage. A single access point can only support a certain amount of concurrent clients at a time, and as we know from the first idea, as more clients join a channel the overall airtime decreases. So to decrease this client per channel issue we add more access points into a given area. Assuming zero to little co-channel interference this approach works great, however often times in dense deployments we often see a lot of clients joining a few of the access points, leaving some access points without any clients. The reason for this issue a lot of the time is due to access point placement.
In an open space with little to no obstruction wireless signals can travel extremely far. Without walls and other obstacles in the way the free space path loss algorithm tells us that signal attenuation happens at a very slow rate. (6dB for every doubling of distance) Without walls in the way clients connecting to the network and the roaming around the building may still see the signal from the original access point a suitable and not trigger the client to roam to a closer access point.
By moving access points from the hallways into enclosed offices, or conference rooms the signal from that access point will be attenuated and won’t travel as far as the signal from an access point in a hallway. This means that clients close to the access point will be able to connect, but as they move away from the area they will be forced to roam to a closer access point.
Wireless design is both an artform and a science. I highly recommend use a predictive tool like Ekahau to get the most accurate placement as you can. It will make you a better wireless engineer, and provides you with a start place to tune your wireless environment. Hopefully while reading this post at least one of these ideas has the same impact on you as it did on me and you walk away a better engineer.