7925G vs IBM laptop - RSSI and SNR vastly differ - Same AP, Same Location?

Answered Question
Jun 4th, 2009

Hi Guys,

I am trying to troublshoot an issue.

I have an IBM notbook attached to a 1242AG AP on WLC1.

I have a Cisco 7925G phone connected to the same AP. Both devices are next to each other.

Both connected to different WLANs, Data and voice respectivley.

Data runs G only.

Voice runs A only.

No rates below 12Mbps are enabled.

Is it normal to see the phone with a much worse RSSI and SNR than the data device because of the high freq?

Example,

Data laptop.

RSSI -35

SNR 63

(thus noise is -98)

Voice phone.

RSSI -75

SNR 18

(thus noise is -93)

I can see from this, that the noise is OK, but RSSI is the problem.

Also, can anyone tell me how often when you do a show client detail <mac>, these values update? Is it every 5 minutes? Is it configurable?

Many thx indeed,

Ken

I have this problem too.
0 votes
Correct Answer by Robert.N.Barrett_2 about 7 years 5 months ago

Given that the phone is on 802.11a, the results you show are not unexpected. 802.11a does not propagate as far as 802.11g (assuming the same EIRP and radiation pattern for both radios at the AP). What other AP's do you have using 802.11a on the same channel? That could also be part of your noise problem.

Are you getting RSSI values from the CLI, or from the devices? There is no standard for calculating RSSI, so comparing the RSSI reported by different devices is not apples-to-apples. Comparing it at the controller might be OK, but I don't know if the controller is reporting an RSSI that it calculates, or if it simply displays what it is receiving from the devices. I believe the controller is simply displaying what RSSI value the ap receives from the devices, so using "show client detail" from the CLI would also not be apples-to-apples when comparing two different devices.

Also - the laptop probably has a better antenna (probably actually two antennas in the display panel) than the 7925 has.

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 5 (2 ratings)
Loading.
migilles Thu, 06/04/2009 - 17:21

Well you may have different tx powers for the 2.4 and 5 GHz radios.

Also we can't guarantee 1 to 1 parity with other clients as they may use different radios with different rx sensitivity / characteristics as well as different antennas in play.

So would suggest to put both on the same band and AP first.

Correct Answer
Robert.N.Barrett_2 Sat, 06/06/2009 - 10:30

Given that the phone is on 802.11a, the results you show are not unexpected. 802.11a does not propagate as far as 802.11g (assuming the same EIRP and radiation pattern for both radios at the AP). What other AP's do you have using 802.11a on the same channel? That could also be part of your noise problem.

Are you getting RSSI values from the CLI, or from the devices? There is no standard for calculating RSSI, so comparing the RSSI reported by different devices is not apples-to-apples. Comparing it at the controller might be OK, but I don't know if the controller is reporting an RSSI that it calculates, or if it simply displays what it is receiving from the devices. I believe the controller is simply displaying what RSSI value the ap receives from the devices, so using "show client detail" from the CLI would also not be apples-to-apples when comparing two different devices.

Also - the laptop probably has a better antenna (probably actually two antennas in the display panel) than the 7925 has.

migilles Sat, 06/06/2009 - 13:17

Yes signal will be hotter on 2.4 even if using the same tx power. The 7921/7925 does use dBm for the RSSI. There is also no gain in the 7921/7925 antenna, where typically there may be up to 2 dBi gain on the laptop. The 7921/7925 can tx up to 16 dBm for OFDM (802.11a/g) and up to 17 dBm for 802.11b rates.

kfarrington Mon, 06/08/2009 - 01:09

Many thx Rob, That is fantastic. Most helpful indeed. I will check this out :)

kfarrington Mon, 06/08/2009 - 01:57

Ahh, I am just looking at my heat maps, and where I have 19 channels avalable, 36 all the way thru to 140, some APs that are adjacent, seem to be on the same channel.

How bizarre is this?

Any thoughts?

Thx

Ken

migilles Sun, 06/21/2009 - 10:00

There is no need to enable so many channels. Typically would need more than 12 channels (UNII-1, UNII-2, UNII-3). UNII-1 can have limited power as well as some of the UNII-3 channels. Not sure if UNII-3 is available for you or not though as not all countries support it. UNII-2 should be able to do 17 dBm. It looks like you have all of UNII-1 (4 channels), UNII-2 (4 channels) and UNII-2 Extended (11 channels) enabled. All of UNII-2 are DFS channels. Would suggest to have some UNII-1 through the deployment. Can use the 4 UNII-2, then 4 of the 11 UNII-2 Extended to utilize 12 channels.

The more channels enabled, the longer it could potentially take to discover an AP even if in the high priority scan list. Also DFS channels must be passively scanned first per 802.11h.

George Stefanick Fri, 06/12/2009 - 08:48

It is always best to use your lowest common device when you survey, most cases your VoIP client. So you can make sure you are meeting the needs of those clients. I attached a pic of my laptop with a cisco phone...

I can always check what my laptop is recording and what the handset is seeing. Handset always wins!

You will also see the delta between your survey client NIC and your phone.

Attachment: 
kfarrington Mon, 06/22/2009 - 02:53

Thx guys, That is great info.

The WLCs automatically with DFS enable all channels so maybe I should look at chaning this?

Many thx indeed,

Ken

Actions

This Discussion