Noise Floor calculation

(menno faber) #1

I know that the Noise Floor depends on the Bandwidth and in the datasheet from Semtech an1200.22, there is the following equation to calculate the Noise Floor:

If I use 125kHz for the Bandwidth (B), I get the following result: -122.9 dBm
When I receive LoRa packets at the Gateway and calculate the average measured Noise Floor, by “RSSI - SNR” i get the following result: -100dBm more or less. This creates a gap of 22.9 dB.

Where I live we use the 915MHz, if I use the whole band for my Bandwith “928 MHz - 902 MHz = 26 MHz = B”, and calculate the Noise Floor I get -99.78 dBm, that is much closer to the measured Noise Floor.

Is what I do correct? Can someone refer an article or book explaining this?



Not sure why “RSSI - SNR” should give you the average measured Noise Floor ?

(LoRaTracker) #3

I dont think you will have a lot of success trying to relate the reported RSSI values to a external calculation.

The RSSI reported by the LoRa device, is in my opinon, only an estimate, an indication and a long way from a precise measured figure. Indeed the formulae used to provide the ‘reported’ RSSI level has changed considerably between the different revisions of the SX127x data sheet.

If you are actually trying to work out how close the packets you are seeing are to failure then the reported SNR seems to be a good indicator. For instance if at SF12 you see a reported SNR of -10dB, then you probably have 10dB of link margin left, since SF12 operates down to -20dB SNR.