Creating Cells - Power Settings - How determine?

  • 1
  • Question
  • Updated 11 months ago
Background - We've been using Aerohive for many years, most of that has been with sparse populations of APs and limited # of clients.  We have always left auto-power to manage signal strengh as our APs were widely spaced.  We are now moving into density installs with single APs (3x3 models) in each classroom, and presently we're installing 5 APs in a new fire house.  My question is general in nature, but a  present project of a new fire house is my working example-test case.

It's our understanding from Aerohive training.other that for high density one wants to in essence create cells in rooms or areas, to avoid bleed between rooms/areas with other APs.  And 5ghz and 2.4ghz being different in their penetrating characteristics.

So here is the question(s) -  What is the best way to determine good power settings?  Actually measure with a mobile device (smaller antenna), laptop (larger antenna), use the RSSI or predicted via the maps?  Use a standard setting for room size & wall construction?

And regardless of the measurement method, what are good dbi levels to avoid bleeding over?

In our fire house example I have adjacent rooms of drywall, wood studs, single floor (garage cinder block walls).  We have one room we want the AP to be dedicated too (training & community room - AP230), and some APs (AP 130) to serve 2 rooms, and 1 AP to serve  multiple small sleeping rooms clustered off a hall.

All the documentation out there talks about the concepts, but nowhere is it ever mentioned how to determine how to best implement the concepts.

Any expertise, tips, tricks, best practices would be welcome!  We can go so far as to provide a floor map if the discussion does get that detailed.  Our goal is to learn this to get this fire station properly configured and use it as a learning case to then apply in larger installs like our schools.

Photo of Kevin Berrien

Kevin Berrien

  • 1 Post
  • 0 Reply Likes

Posted 11 months ago

  • 1
Photo of Jonathan Hurtt

Jonathan Hurtt

  • 98 Posts
  • 48 Reply Likes
A few concepts that will help you tackle this is that when you think about Power settings, you first must look at the Channel. The need to adjust power is two fold, prevent clients from associating or staying connected to AP at further distances (roaming issues) and secondly preventing Co-Channel or Adjacent Channel interference which can degrade performance.  

So I would suggest you first tackle the channel before power. When looking at the channel we will have to treat the spectrums differently due to the number of non-overlapping channels. Typically I would recommend using 2.4 GHz for coverage and 5.0 GHz for capacity, so in a very high density environment (as in number of Access Points) you will most likely  be disabling 2.4 GHz radios, turning them into sensors or if you have the AP250/AP550 changing the in to 5 GHz Radios. What ever you do, It is important to come up with a proper channel plan. You can do this manually or you can allow the Auto Channel to perform this task. If you do chose to use auto channels on 2.4 GHz you are going to have to make sure you have disabled/converted significant # of radios since 2.4 GHz has limited # of channels.

Once you get a channel plan, the power is the second part to tweak. It is recommend to try to match your output power to the output power of your clients. So most clients will be in the range of 11 dBm to 14 dBm. I would recommend to set in the Radio Profile the Max TX power to 17 dBm (5Ghz) or 14 dBm (2.4Ghz) as starting point and see how the Auto power adjust, if you see may access points dropping below that Max you might want to revisit your channel plan or you can decrease the max transmit power another 3 dB (which is 1/2 about of power) to 14 dBm (5Ghz) or 11 dBm (2.4Ghz). The one thing about lowering your power on AP too much (say like 5dBm) is that the clients might still be using higher power (>11dBm) output and causing CCI when they send frames. 

I would also look to adjust your data rate and remove 1 & 2 Mbps and 5.5 & 11 Mbps (if you have no 802.11b clients that need to connect). If you have a higher density environment you can even look to remove 6 & 9 Mbps (12 Mbps lowest basic Data rate) and potentially 12 & 18 Mbps (24 Mbps lowest basic Data rate) data rates on both 2.4 & 5 GHz spectrums. The lowest basic data rate is helpful as that its he rate which broadcast and multicast frames are sent at, so if you have a chatty network this will help speed things up.

Once you get the channel and power settings where you want, you will want to verify it with a active/passive site survey or you can spot check certain areas. The idea is you want the strongest signal to be on a different channel than the second strongest signal. Typically you want a delta of 30 db between the strongest signal on a channel and the second strongest signal. This will ensure the highest SNR needed to obtain highest MCS data rates (for 20 MHz wide channels) 

If you use auto channel for 5 GHz you will most likely avoid ACI as well as CCI. There are enough channels where you wouldn’t want Channels 36 and 40 being the two strongest signals. With 2.4 GHz it will be hard to avoid that but it could be done, but you might have to manually set the channels depending on the physical environment and number of 2.4 radios you have disabled/converted. 

Note that the 802.11 standard states that a client must backoff from sending a frame if it hears another 802.11 signal at -82 dBm or stronger or any RF signal at -62 dBm or stronger, so as long as the neighboring signal (from perspective of the sender) is weaker than -82 dBm the client should be able to send the frame and not be affected by neighboring access points. 

The key here is that that signal is on the same channel the client is using, not signals on different channels. 

Sorry for long response, but hope this helps
Photo of Crowdie

Crowdie, Champ

  • 960 Posts
  • 269 Reply Likes
A couple of additional points:

  • Using a standard "one access point per x rooms" rule across a school doesn't commonly work.  You are concerned with airtime and utilisation and this is not consistent across a school.  Keith Parsons wrote an excellent article Why One Access Point Per Classroom Approach Is Wrong on this topic.
  • You need some overlap of adjacent access point's coverage cells so clients can roam.  If the overlap is too small then clients will fail to roam.  If the overlap is too large than clients will become "sticky" (reluctant to roam).
  • It is not the roll of the radio to direct signal - that is the role of the antenna.  If you are going to place a number of access points in an open plan area, like a hall, then using multiple access points with internal (omni-directional) antennas will not work.  What tends to happen when multiple access points with internal antennas are placed into an open plan area is that 80% or so of the wireless devices will associate to the same access point.  To avoid this use access points with external antenna connectors mounted on the ceiling with directional antennas pointing down at the floor.  If you need four access points for capacity then you need external antennas that cover just over a quarter of the floor.  However, if you have an open plan area with a low ceiling (say 2.5 to 3 meters high) then you are not going to be able to use external antennas.