Collecting verifiable and reproducible Wi-Fi signal strength statistics

[Halcyon]

As part of my studies, I'm looking for a scientific method (or as close as possible) to quantify Wi-Fi signal strength/quality which can be presented in a report and reproduced by someone else. I don't know if there is an "international standard" way of conducting those types of measurements?

I realise the measurement of signal quality is highly dependent of a number of factors, particularly how manufacturers implement such measurements in their equipment.

Am I better off just recording the sensitivity of the receiver, gain of the antenna and the dBm value?
Is it possible to translate dBm values into loose "excellent, good, poor" terms?

[dmills]

What exactly are you trying to measure?

Signal strength (per channel, or per access point), or SNR, or BER? They all make sense in different ways and need different instruments to measure.

Excellent/good/poor depends as much on signal to noise ratio as raw field strength, and for most WIFI in most (inner city) locations, interference from other ISM band devices is likely as much a limiting factor as pure received power is.

You will find that the wavelengths are short enough (and the environment sufficiently reflective in many cases) that received power can fluctuate wildly with movements of only a few tens of cm.

A calibrated aerial and spectrum analyser will tell you about power in each channel, but says little about which APs are generating that power moment to moment, the tools depend on just what you want to measure....

Regards, Dan.

[Halcyon]

Hi Dan, thanks for the reply.

Basically, I'm a network engineer. I understand how 802.11 works from OSI layers 2 and up. I do have a fairly good understanding of how RF works, limitations in terms of 2.4 and 5 GHz bands, channel widths, designing a channel plan etc...

But I'm stuck between a rock and a hard place. I basically need to show a room of non-technical people that if I take a measurement, say I use my phone and it measures -50dBm in a particular area relative to a particular access point, then it's a pretty darn good signal. On the flip side, -90dBm is a fairly average-to-bad one and something needs to be done about it.

How do I quantify it? I'm thinking worst case here... let's just say that room of non-technical people don't trust the figures I've provided and wanted to get someone to repeat my tests and verify what I've said as being true, how would I go about ensuring my tests are transparent and repeatable? Obviously not everyone is going to get exactly the same results as me, but it will be fairly close.

But say I measure -90dBm and someone else gets -85dBm, well that's "five better" than I got so I must be wrong.

It's actually a really difficult question I'm asking because the science doesn't translate well to laypersons.

[ogden]

But I'm stuck between a rock and a hard place. I basically need to show a room of non-technical people that if I take a measurement, say I use my phone and it measures -50dBm in a particular area relative to a particular access point, then it's a pretty darn good signal. On the flip side, -90dBm is a fairly average-to-bad one and something needs to be done about it.

How do I quantify it? I'm thinking worst case here...

Find some industry-best WiFi survey tools - software and hardware, read user manual/recommendations/papers All the argumentation is there. Signal strength measured in dBm is signal power coming into receiver and it is antenna gain/quality/orientation dependent. Antennas and their performance differ. Signal measurement repeatability using simple devices with random antennas is very poor. Two phones or laptops can report reading that differ by 10 dB easily - just because [lot of reasons]. If you want better than "ballpark signal strength figures", then you shall use calibrated receiver/antenna.

BTW WiFi receive performance depends not only on signal strength but also on signal to noise ratio (SNR). You can have "good" -60 dBm signal but nothing works just because of (let's say -60 dBm) in-band interferer signal.

[edit] You can calibrate your "test receivers" in-house. By properly doing that you can get into +/- 3dB spread/repeatability which is more or less good for WiFi surveys