site survey

[marcinwojcik]

Hi

What is the best way of doing a site survey when APs are managed by a controller? What I mean is that the AP can be set to different power level when there is a low and high number of clients - in such case a survey may not deliver correct data (when it's done with a small number of clients associated to APs), or I got it wrong?

Any suggestions? Best practices?

Thanks.

[Scott Fella]

Have you done a site survey prior to the install?  If you are doing a post site survey, there are various ways you can go about doing it.  If you had a site survey done and they surveyed using 50mw for example, then you can globally set the power for the AP's and then I would freeze the DCA so the AP's won't change while doing your survey.  This way you can see if the survey that was done, provided you with the coverage required.  You can also just survey as-is and see how RRM affected the coverage or not.  Depending on your density and if the WLAN is in production, that will sort of limit you on what you can or cant do.  As long as you understand the software you are using to perform the site survey, you should be fine either way.

[marcinwojcik]

Thanks for the reply. I do not have much experience with WLANs so I do appreciate your help.

Have you done a site survey prior to the install? 

- I was told the survey was done. I need to contact the company and get the information.

If you are doing a post site survey, there are various ways you can go about doing it.  

- This is what I like to do. I want to get the picture of the running WLAN. I think the passive survey is the way to go in such case - am I on the correct track?

 If you had a site survey done and they surveyed using 50mw for example, then you can globally set the power for the AP's and then I would freeze the DCA so the AP's won't change while doing your survey.  This way you can see if the survey that was done, provided you with the coverage required.  

- I can see that we had 2 channel changes in last 7 days, so freezing the DCA shouldn't cause any problems (right?).

- Can I trust that the WLC sets the APs to the best settings in terms of power level and channel? Is it worth coming on weekend and set APs to specific settings and how it reflects on the survey?

- How can I globally change the APs power level?

The exercise needs to help me to decide if I shall add APs or not.

Thanks.

[Scott Fella]

- I can see that we had 2 channel changes in last 7 days, so freezing the DCA shouldn't cause any problems (right?).

It will not be an issue

- Can I trust that the WLC sets the APs to the best settings in terms of power level and channel? Is it worth coming on weekend and set APs to specific settings and how it reflects on the survey?

I would trust the WLC in setting the correct channel and power. If they surveyed right, you will not have any issues☺

- How can I globally change the APs power level?

On the WLC GUI, click on the WIRELESS tab on top, then 802.11a/n or 802.11b/g/n, then click on TPC and change the FIXED to 1 (100mw), 2 (50mw), or 3 (25mw). Depending on what AP’s you have, the 5ghz will vary as some start off at power level 1 (50mw). Just look at the AP spec sheet.

[marcinwojcik]

Scott,

What do you think about passive vs active survey? Which do you carry out more often (especially in the case of a post survey)?

In the case of active survey what roaming aggressiveness you use?

I would trust the WLC in setting the correct channel and power.  If they surveyed right, you will not have any issues☺

Is there a point of changing the setting (freezing DCA, setting up the power level) if you think that what WLC sets is trustworthy? I do want get a picture of the coverage during production day.

Thanks for all advice.

Marcin.

[Scott Fella · Accepted solution]

What do you think about passive vs active survey? Which do you carry out more often (especially in the case of a post survey)?

I normally do a passive, just because I’m looking at coverage. Active survey varies and it doesn’t really give you the roaming identical to various devices out there. Some devices roam well and other roam poorly. Focus on your RF coverage for now.

In the case of active survey what roaming aggressiveness you use?

Again… I wouldn’t worry about this, if you want to test it out, just leave it at default.

Is there a point of changing the setting (freezing DCA, setting up the power level) if you think that what WLC sets is trustworthy? I do want get a picture of the coverage during production day.

Well if your AP changes channels, it might show up as poor coverage. It doesn’t matter what channel the AP is on, you just want to make sure you have coverage for now.

[jerryard1]

I generally survey to the power level of the weakest client the wireless network will be supporting. If the client has a max transmit power of 40dBm, then I will set the APs to the power setting closest to that (preferably lower like 25mW but no more than 50mW.)

Having said that, the Cisco default answer of "it depends" generally holds true. Your wireless budget will determine the number of APs you can buy which will pre-determine your AP density allowance. But you can't go wrong surveying to the lowest client power level and completing a full spectrum analysis of both the 2.4GHz and 5GHz ranges.

[jerryard1]

Correction: 40mW for the client power.

[marcinwojcik]

Hi

How can I check the power level of the client? Shall I check a vendor's spec for the wireless card?

As I mentioned in the reply to Scott  - I'd like to get the picture of my WLAN as-is and maybe after that start making changes.

Thanks.

[Scott Fella]

You can check the vendors website for the specs, but usually for data, I would just survey at 50mw and for voice, 25mw as a general rule of thumb.

Verify with the vendor who initially performed the site survey.  Take a look at their report and you should see what they have set the power levels to.  If they did survey at 100mw, well... that might be an issue or not, depending on how dense their deployment is.  Like I mentioned, if you do not have any wireless VoIP phones, then setting the AP's globally to 50mw prior to your survey will get you the best info regarding your coverage.  All you need to do then is to perform a passive survey like you mentioned and then analyze your findings.

[marcinwojcik]

Thanks for your help guys.