I have recently been reviewing placement of the WAP’s on our network, and the channels that we use.
Until very recently, we had a mix of fixed AP’s in some areas, and mobile AP’s on laptop trolleys. In short, this didn’t work well. If someone moved a mobile AP into an area where it interfered with a fixed AP – well then, kiss goodbye to any reliable wireless connectivity in that area. This wreaked havoc with a few lessons, meetings and other events.
So – to solve this, we looked at covering the entire of one end of the college, permanently, with fixed AP’s. We had a couple spare, so could afford to do this.
With the AP’s in their new fixed locations, we ensured that the channels were set apart from each other using channels 1, 6 and 11. As recommended in the HP Wireless Fundamentals Guide. (A great wireless networking read, BTW. Can’t find the link again on their website though.)
This improved the connectivity and reliability somewhat, but still not as good as I was hoping that it would be for a fixed installation.
Using NetStumbler, I found out that there were a lot of other wireless networks operating on the same channels that we were, particularly 1 and 6 though. I referred back to the Fundamentals guide, and a few other resources on the Internet, but everything kept telling me that 1, 6 and 11 were the channels to be using. I started looking at the frequency plot of the channels to try and understand why this is.
I have now taken a gamble. I have not yet read anywhere (maybe I’ve looked in the wrong places?) that says that 3, 8 and 13 is a valid combination and will work. I have deployed it though, and as of yet, my network is the only one operating on those channels. Also, it seems to be much more solid connections. I will post if anything comes up which stops it working, but so far, 1 week down, so good.
The chart above also got me thinking about the channels, and at a glance, it seemed that channels 1, 5, 9 and 13 were all non-overlapping. This is a bit of an illusion though. Drawing out a frequency plot in Excel highlighted a small (2Mhz) overlap between these channels. I have included the table below for completeness:
|Channel||Lower Boundary||Centre Frequency||Upper Boundary|