Page 1 of 1

Noise Floor

Posted: Tue Jun 11, 2013 11:35 pm
by hidagar

I have a Noise Floor -104dbm, I read in some post that if the signal is above - 100 dbm it's not good means that I can have RF. That is true?


Re: Noise Floor

Posted: Wed Jun 12, 2013 1:08 am
by killersoft
I hope the following helps you understand nose floor vs signal level(or strength). Just remember you are dealing with negative numbers and you will be all good...
Quoted FROM : ... ratings.3F
How do I read signal and noise ratings?

These numbers are given in decibels (dB) and are expressed as negative numbers. The more negative the number, the less strength it represents. Thus, -40 dB represents more strength than -70 dB. The values are logarithmic. A signal amplitude change of 3 dB is equivalent to a factor of two; 10 dB is a factor of ten.

Based on this Forum-post

Signal: (in dBm) A small negative number is good (-40 is good, -98 is bad)

Noise: (in dBm) A large negative number is good (-98 is good, -40 is terrible, -70 would be pretty bad in the real world)

SNR: (in dB) High is good (should be the same as difference between noise and signal, a difference of 20 would be great, a difference of 1 may barely work)
SNR(dB) = Signal(dBm) - Noise(dBm)

Signal Quality: High is good (somewhat like SNR but indexed to 100 with noise as the base, percentage of the best theoretical ideal quality in regards to your local-noise)

Signal - Noise = SNR
-82 - -98 = 16

Signal / Noise * SNR = Signal Quality
-82 / -98 * 16 = 13.4%

Typically, noise will be -92 which means you should get a clean connection with a signal as low as -92. However, expecting to hold a good connection with a signal lower than -85 (e.g. -90), is expecting too much. The signal can be improved by -3 dB by doubling the power setting at the transmitting radio, e.g., 100 mW increased to 200 mW would improve your signal from -85 to -82. Antennas with increased gain will also help. Say you had the standard 3 dB antenna and changed it for a 12 dB antenna, that's a 9 dB increase, so your signal would increase from -82 to -73 which would be an excellent signal, probably capable of 54 Mbps. Using the term excellent in terms of running a WISP, it would probably be only 3 bars on a 5 bar signal strength meter. Don't worry if, as a WISP your signal quality is low, like 14%. It's not really a problem since -82 is considered acceptable.
[edit] How does the SNR impact the speed and range of my wireless connection?

SNR, range and speed (data rate) are tightly interdependent. Users often notice that higher data rates do not "travel" as far as lower data rates do - and frequently they think that increasing the power on the router will take the signal further (increase the range).

It is not the power of the router, it is Signal-to-Noise ratio (SNR) that dictates the data speed and the range of the signal. SNR determines which data rates can still be correctly decoded in a wireless connection - as data rates increase from 6 Mbps towards 54 Mbps, more complex modulation and encoding methods are used for transmission and that requires much higher SNR to properly decode the signal back to the data stream on the receiving side.

Using full 54 Mbps data rate requires at least 25 dB of SNR - and getting that much SNR is achievable only if router and client are relatively close together. As the signal travels further away from the transmitter, a path loss occurs (the signal gets attenuated) and SNR is getting lower and lower. Lower data rate transmissions can be decoded from much weaker signals (low SNR) and as a result the signal appears to travel further.

Increasing the power of the transmitter will often affect the listening side of the same device as well, affecting much higher noise levels (and worsening the SNR ratio). It is frequent occurrence with beginners to see their routers tweaked so they generate the highest possible wattage of signal, raising the floor of the noise as well - thus keeping the SNR at the same level, as if the router hadn't been tweaked at all.

[Data Rate] [Minimum SNR] [Modulation/Encoding]
6Mbps 8dB BPSK 1/2
9Mbps 9dB BPSK 3/4
12Mbps 11dB QPSK 1/2
18Mbps 13dB QPSK 3/4
24Mbps 16dB 16-QAM 1/2
36Mbps 20dB 16-QAM 3/4
48Mbps 24dB 64-QAM 2/3
54Mbps 25dB 64-QAM 3/4

Re: Noise Floor

Posted: Wed Jun 12, 2013 11:13 am
by hidagar
Then if i Have -104 dbm is better than -97dbm

Best regards

Re: Noise Floor

Posted: Wed Jun 12, 2013 1:35 pm
by ndor
For noise,yes.for signal,no

Re: Noise Floor

Posted: Wed Jun 12, 2013 4:55 pm
by lfs0a
killersoft,learned a lot from your post,thanks.

Re: Noise Floor

Posted: Thu Jun 13, 2013 1:12 am
by killersoft
No worries. I edited the last table of info, I had the brackets in the wrong spot, which made the table header read wrong!..
The same principal's applies to ADSL over phone lines as well as digital satellite services.(Though Satellite tends to use the Term 'EbN0' for signal to noise value of a signal depending on modulation scheme encoding/decoding).

Good luck..

Re: Noise Floor

Posted: Mon Nov 20, 2017 12:34 pm
by el berto
Hi guys, I have one question: on same link I get two different noise floor levels at same frequency (similar result also on all other frequencies)
- CM9 on old radio = about -96dB
- R52H = about -104dB

Is it normal to earn about 10dB just with new wireless card?
Or -96dB means CM9 module is damaged?

Re: Noise Floor

Posted: Wed Nov 22, 2017 12:17 am
by troffasky
Do you actually get better speeds/lower packet loss?

A few possibilities:
- Better RF design with improved filtering and rejection will get you a better noise floor
- Perhaps old card wasn't seated/connected properly
- Cards may not be calibrated the same
- Different cards may have different firmware