Sunday, September 13, 2015

Should I re-design my Wi-Fi, or "swap them out"?

WLAN Site Survey – Should I redesign my WLAN, or just “swap ‘em out” and be done with it?

The purpose of this exercise was to design a WLAN for an office building with either the Cisco 3602i or the 3702 series access point.  We wanted to know if we could survey with either one of those access points, and deploy the other, if it came to that, because we didn’t have a 3702 on the shelf.  After mentioning this to Sam Clements, he immediately came to the rescue and let me borrow some gear.  Our new goal was to survey with both APs, configured to the same power level, so we could compare cell sizes.  You can follow Sam here: http://sc-wifi.com/

We (@WiFi_Princesa and I) rigged up a 2504 controller running 8.x code and a mast-like structure that would hold two access points up to the ceiling.  We don’t normally use a WLAN controller to do surveys – we usually put autonomous code on them and configure them as survey APs.  Since this was borrowed gear, I decided to leave it in its CAPWAP state and use the 2504.  At the last moment, we decided to toss a 3502 into the equation, since we had one available.

We configured the three access points as follows:

3702, power level 3, channel 36

3602, power level 3, channel 40

3502, power level 3, channel 44

This is the 3702’s  “don’t want” coverage area.  Notice how far out that -85 dBm boundary goes! 

This is the 3702’s “want” coverage area.  The -67 dBm boundary is about where we expected it would be.  This is an office building with offices around the perimeter and cubicles in the center. 

 

Now let’s compare to the 3602 series AP:

This is the 3602’s  “don’t want” coverage area.  Similar to the 3702.

This is the 3602’s “want” coverage area.  The -67 dBm boundary is about the same as the 3702. 

 

Now for the 3502i series that we added at the last minute:

This is the 3502’s  “don’t want” coverage area.  Notice how far out that -85 dBm boundary goes!  Something is fishy here!

This is the 3502’s “want” coverage area.  The -67 dBm boundary is a lot larger than I thought it would be.  But we put them all on UNII-1, power level 3. 

The only thing we can think of is that power level 3 is not the same across the board with all three APs in the UNII band.   We quickly figure out how to do this by surfing here:  http://mrncciew.com/2013/03/25/rf-power-terminology/

So, let’s get started!

(Cisco Controller) >show ap config 802.11a Cisco_3602i

<output omitted> 

Tx Power

      Num Of Supported Power Levels ............. 5

      Tx Power Level 1 .......................... 14 dBm

      Tx Power Level 2 .......................... 11 dBm

      Tx Power Level 3 .......................... 8 dBm

      Tx Power Level 4 .......................... 5 dBm

      Tx Power Level 5 .......................... 2 dBm

      Tx Power Configuration .................... CUSTOMIZED

      Current Tx Power Level .................... 3

      Tx Power Assigned By ...................... User

    Phy OFDM parameters

      Configuration ............................. CUSTOMIZED

      Current Channel ........................... 40

 

(Cisco Controller) >show ap config 802.11a Cisco_3502i

<output omitted>

Tx Power

      Num Of Supported Power Levels ............. 7

      Tx Power Level 1 .......................... 17 dBm

      Tx Power Level 2 .......................... 14 dBm

      Tx Power Level 3 .......................... 11 dBm

      Tx Power Level 4 .......................... 8 dBm

      Tx Power Level 5 .......................... 5 dBm

      Tx Power Level 6 .......................... 2 dBm

      Tx Power Level 7 .......................... -1 dBm

      Tx Power Configuration .................... CUSTOMIZED

      Current Tx Power Level .................... 3

      Tx Power Assigned By ...................... User

    Phy OFDM parameters

      Configuration ............................. CUSTOMIZED

      Current Channel ........................... 44

 

(Cisco Controller) >show ap config 802.11a Cisco_3702e

<output omitted>

Tx Power

      Num Of Supported Power Levels ............. 5

      Tx Power Level 1 .......................... 15 dBm

      Tx Power Level 2 .......................... 12 dBm

      Tx Power Level 3 .......................... 9 dBm

      Tx Power Level 4 .......................... 6 dBm

      Tx Power Level 5 .......................... 3 dBm

      Tx Power Configuration .................... CUSTOMIZED

      Current Tx Power Level .................... 3

      Tx Power Assigned By ...................... User

    Phy OFDM parameters

      Configuration ............................. CUSTOMIZED

      Current Channel ........................... 36

 

Looks like we have three different dBm power levels associated with these three access points, all in the UNII-1 band, and on WLC power level three.  We have the 3602i at 8 dBm, the 3702 at 9 dBm, and the 3502 at 11 dBm.

This information seems to not be readily available on Cisco’s website, however, I can point you to Brian Long’s blog, as he is working on a matrix for us: http://blong1wifiblog.blogspot.com/

After reading Brian’s blog and seeing all of the channels, I think abandon the 2504 WLC for surveying with the 3702 and us an autonomous image on it.  Here’s how to do that on a 3602i.  https://ciscotophat.wordpress.com/tag/3602/

Conclusion:

At least we know the 3602 and 3702 are somewhat similar for coverage patterns.  I also answered my own question – if I had WLAN with 3502 series APs deployed properly and met my requirements at the time it was designed, I believe I would still survey/re-design with today’s requirements in mind, and not just “swap them out”.