Signal strength/signal-noise-ratio question

black02ep3 · 09-08-2004

Hi,

I'm new at this, so please bear with me:

We are running sniffers in a room at various locations, trying to read signal strength/noise ratio at those spots at different times and trying to build up a simple pattern for the signal strengths.

We noticed that signal strengths are always varying by quite a bit at the some locations (variation being from 5 to 10 units). Is this normal? And if so, is there a way we can normalize the data so that they become more consistent? And how can we use the signal-noise-ratio alongside the signal strength? (currently we're disregarding the signal-noise-ratio values)

Thank you very much,

James

Thorn · 09-08-2004

Quote:

Originally Posted by black02ep3

Hi,

I'm new at this, so please bear with me:

We are running sniffers in a room at various locations, trying to read signal strength/noise ratio at those spots at different times and trying to build up a simple pattern for the signal strengths.

We noticed that signal strengths are always varying by quite a bit at the some locations (variation being from 5 to 10 units). Is this normal? ...

Yes.

Quote:

Originally Posted by black02ep3

And if so, is there a way we can normalize the data so that they become more consistent?

Define "normalize" in this context.

Quote:

Originally Posted by black02ep3

And how can we use the signal-noise-ratio alongside the signal strength? (currently we're disregarding the signal-noise-ratio values)

Please rephrase the quesiton or define how would you like to use the SNR.

black02ep3 · 09-08-2004

"Normalize" as in, some sort of algorithm that will cause two sets of signal-strength and signal-noise-ratio to end up (nearly) the same

h(ss1, snr1)~=h(ss2, snr2)

In essense, if I get different signal-strengths at the same place, I want to have some way that tells me that even though they differ a little bit, they are really the same, through either the application of signal-noise-ratio or any other ideas.

As to "how" I want to use SNR, I really am not sure. That's the whole problem: it's there for a reason, I believe, so how do I use it? Use it as part of some calculation? Use it to determine whether to discard a signal-strength reading?

Thanks,

James

Thorn · 09-08-2004

Quote:

Originally Posted by black02ep3

"Normalize" as in, some sort of algorithm that will cause two sets of signal-strength and signal-noise-ratio to end up (nearly) the same

h(ss1, snr1)~=h(ss2, snr2)

In essense, if I get different signal-strengths at the same place, I want to have some way that tells me that even though they differ a little bit, they are really the same, through either the application of signal-noise-ratio or any other ideas.

The mean Signal is probably the best then.

Signals (and Noise) always fluxuate to a certain extent, due to a number of different influences.

Quote:

Originally Posted by black02ep3

As to "how" I want to use SNR, I really am not sure. That's the whole problem: it's there for a reason, I believe, so how do I use it? Use it as part of some calculation? Use it to determine whether to discard a signal-strength reading?

S - N = SNR

Unless your a RF engineer, Ham, or some other "freq geek" it's probably not going to mean too much to you.

Typiclally you want a SNR to be about 10dBm or better to ensure a clear signal.

Casisiempre · 09-08-2004

Quote:

Originally Posted by Thorn

Typiclally you want a SNR to be about 10dBm or to ensure a clear signal.

You want it to be at least 10dBm, and generally, the bigger the SNR the better (if nothing else then less chance of interference cutting your signal).

netzst · 09-20-2004

The operative word in SNR is 'ratio': there might be great SNR, but if the signal is so weak (low signal strength) that you can't hear it, it won't do you any good.

On the other hand, the signal could be quite strong, but if the noise is quite strong as well (low SNR), you won't be able to detect the signal out of the noise.

That's why both are reported.

Maybe.

Another way of looking at it might be, SNR tells how high the signal stands out above the noise. Higher SNR is better. But if the signal is weak, only expensive radio receivers are going to be able to detect it. High signal, high SNR, even cheap radios can pick it up. (Real high signal: even radios that don't want to hear it will hear it--but since they don't want to hear it, it'll go into THEIR SNR as 'noise'.)

w2004 · 09-21-2004

I am new in this field. Please bear with me.

Is it possible to increase the sampling rate for SNR to 100 samples/s in the NetStumbler?

Thanks for your comments.

W2004

09-08-2004	#1 (permalink)
black02ep3 Registered Member Join Date: Sep 2004 Posts: 8	Signal strength/signal-noise-ratio question Hi, I'm new at this, so please bear with me: We are running sniffers in a room at various locations, trying to read signal strength/noise ratio at those spots at different times and trying to build up a simple pattern for the signal strengths. We noticed that signal strengths are always varying by quite a bit at the some locations (variation being from 5 to 10 units). Is this normal? And if so, is there a way we can normalize the data so that they become more consistent? And how can we use the signal-noise-ratio alongside the signal strength? (currently we're disregarding the signal-noise-ratio values) Thank you very much, James

09-20-2004	#6 (permalink)
netzst Registered Member Join Date: Sep 2004 Posts: 2	The operative word in SNR is 'ratio': there might be great SNR, but if the signal is so weak (low signal strength) that you can't hear it, it won't do you any good. On the other hand, the signal could be quite strong, but if the noise is quite strong as well (low SNR), you won't be able to detect the signal out of the noise. That's why both are reported. Maybe. Another way of looking at it might be, SNR tells how high the signal stands out above the noise. Higher SNR is better. But if the signal is weak, only expensive radio receivers are going to be able to detect it. High signal, high SNR, even cheap radios can pick it up. (Real high signal: even radios that don't want to hear it will hear it--but since they don't want to hear it, it'll go into THEIR SNR as 'noise'.) Last edited by netzst : 09-20-2004 at 01:40 PM.

09-21-2004	#7 (permalink)
w2004 Registered Member Join Date: Sep 2004 Posts: 1	Can SNR be sampled at higher rate? I am new in this field. Please bear with me. Is it possible to increase the sampling rate for SNR to 100 samples/s in the NetStumbler? Thanks for your comments. W2004

Thread Tools
Show Printable Version Email this Page
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

09-08-2004	#3 (permalink)
black02ep3 Registered Member Join Date: Sep 2004 Posts: 8	"Normalize" as in, some sort of algorithm that will cause two sets of signal-strength and signal-noise-ratio to end up (nearly) the same h(ss1, snr1)~=h(ss2, snr2) In essense, if I get different signal-strengths at the same place, I want to have some way that tells me that even though they differ a little bit, they are really the same, through either the application of signal-noise-ratio or any other ideas. As to "how" I want to use SNR, I really am not sure. That's the whole problem: it's there for a reason, I believe, so how do I use it? Use it as part of some calculation? Use it to determine whether to discard a signal-strength reading? Thanks, James