Very important question that what should be the power setting of an access point. All vendors suggest that maximum transmit power is not a good idea because it increases the chances of co channel interference. During my Ekahau ECSE training by Mr Keith R. Parsons I was able to discuss this question and his answer was very simple.
“Design RF for what you want and do not for what you do not want”
So what does mean by design RF for what you want and do not for what you do not want. End of day as a WiFi engineer our goal should be to design WiFi network which meet client’s requirement which must be determined in initialization phase of the design.
During design process we determine that how many access points and what power we are going to use. Imagine If you have 1 access point and large area to cover so probably you will be using maximum transmit power.
On other hand if you have 4 Aps then 3 APs in 2.4Ghz on channel 1,6,11 and 4th AP will have to reuse 1 of the channels. In this case you will have to consider that what power you want to use for the Aps because great power will leave longer RF tail which will cause interference.
Note: reducing the AP transmit power reduces coverage greatly as compare to RF tail
What I am trying to say is reducing the power doesn’t necessarily mean we are reducing co channel interference. I will go through some examples which shows it’s very important to understand that reducing power will affect coverage greatly.
As Keith suggested that we should break down the area in 3 different areas.
2. Don’t want
3. Don’t care
Want area is easy because its coverage area. This is the value you will be able to get in requirements by locking down least capable most important device or will be given by client. For example, you are design for a cisco mobile device which needs minimum RSSI -67 dBm.
So If we place an access point and achieve -67dBm on some point in the area. It means we met our first requirement, but it does not mean RF stopped there. RF is still going, and this area is called as Don’t want area until it hit our don’t care area.
So question is what is the problem in this situation?
Don’t care area:
Let’s talk about don’t care area. This is the area where RF exists but below device usage threshold. So devices don’t consider this area as RF interference and will continue with normal Wi-Fi mechanism. Don’t care area is -85dBm but do be on safe side you can consider as -90dBm which could be difficult to achieve in 3D RF design.
Don’t want area:
This is the area which cause RF interference if seen by station on same channel. This area is within stations RF threshold and stations will consider this RF energy before transmitting. As long access points are on different channels there is no issue but RF energy travels far and this area will cause co channel interference for access points and the devices which might be far away.
I have taken this graphic below from Keith Parsons lectures. It’s the best way you can demonstrate that how RF energy travels and how it might cause issue for other stations on same channel.
Access Point power:
I am using Ekahau to demonstrate that how power might not help to reduce co channel interference, but good design would do. In the scenario below we have floor plan with 2 access points on channel 36. Both access point configured with 18dBm transmission power. So lets suppose our required or want area is -65dBm, with 2 access point I have achieved my coverage goal of -65dBm on this floor.
But both APs are on same channel we must have some co-channel interference which can be seen below. We can see that both APs can see each other, now we can get rid of interference by changing the channel because we only 2 APs. But what if we have to reuse our channels where we have more APs on same floor or different floors.
Can we rely on power to get rid of co-channel interference?
lets test by simply reducing the power to minimum 1dBm and see if it has reduced our co-channel interfernce. By changing the power from 18dBm to 1dBm we have reduced the interference but we can see interfernce still exists.
by reducing the power it must have reduced coverage?
screen dump below showing how coverage was effected by reducing power to 1dBm. We can clearly see that our want area is greatly affected. It shows that you can not always fix bad Wi-Fi design with tweaks such as power in our scenario.
Always use best practices for Wi-Fi design and then validate by post installation survey. Try to use appropriate antennas for every deployment. With directional antenna you can control RF much better as compared to Omni antennas. Higher transmit power will give you longer RF tail and we can decrease the long RF tail by reducing access point power but same time its necessary to check that how it will effect coverage.
So remember the rule:
“Include the coverage area you want and do not cover the area you do not want”