I'm interested in hearing feedback on 'Active' versus 'Passive' survey results (in particular, using AirMagnet) in regards to accuracy and precision. It is my understanding that true accuracy can only be verified by testing the survey software alongside a true Spectrum Analyzer; and that the values we do see reported by the survey software are some algorithmic voodoo that accounts for RSSI and card calibrations. The problem is the difference in results that I see when comparing an 'Active' survey next to a 'Passive' one.
For example: I did a test last night for 20 minutes total, 10 minutes using each survey type. I was located about 10m away from a Cisco 1252 in a static environment (my laptop and access point were stationary, there was no movement in the room). I then pulled the data from the AirMagnet survey files and summed it up:
Survey Type Signal Strength
Active Survey (5 GHz) -49.1 ???????¡À1.116
Passive Survey (5 GHz) -51.8 ???????¡À1.899
Active Survey (2.4 GHz) -50.0 ???????¡À1.665
Passive Survey (2.4 GHz) -43.7 ???????¡À5.388
What does everyone thing about these results? The differences aren't night and day, but certainly substantial. The listed signal strength is the average of all the logged data ???????¡À the standard deviation.
Anyone else have something to add on the topic?
I believe that passive surveys just calculate information from beacons and probe responses, while active surveys look at information from the data packets that it sends to the AP. I would assume that alone has something to do with the differences.
The channel interference calculations are definitely a bit of voodoo, but I am not sure about how Airmagnet's signal strength calculations might differ from the card by itself.