I’m about to do some range testing on gateways in my local area, to these the distance at which nodes can connect and transmit data to The Things Network successfully.
Currently I have the follow plans:
Test a variety of locations of differing elevation and distance in a somewhat circle shape around the gateway from a few dozen meters to about a kilometre and a half away.
Test two different nodes, these being a TTGO ESP32 SX1276 and a RN2903 connected to a Adafruit Feather M0.
I will be recording if the node believes it is connected, both in the serial line and on TTN and see if the gateway has picked up the node (as I have access to it on TTN). I will also be recording the spreading factor and transmission time.
I was seeing if anyone had any recommendations and advice as to the way range testing should occur and if my process needs any refinement?
@ElectronicallyE, if you decide to do your own thing then look at recording and plotting SNR. The approach you are describing does not identify if a “successful” connection is strong or weak. What you need to record is the margin above the minimum level where a connection can just be made. For example, if you are connecting at SF12 then it operates down to -20dB SNR and if your record a SNR of -10dB, then you have approx. a 10dB link margin. (look up the min SNR for other SFs) Hopefully your node should be capable of outputting the SNR from a transmission. (I fully understand SNR is not the only parameter is a digital transmission system, but its something simple to record and gives a “feel” for the link performance)
@andrewl Sounds like a plan. I’ll take note of it during the testing.
@DavidG Thanks for the suggestion. I think I’ll use my RN2903 to obtain values I get from the range testing and upload them as defined under the contribution section of the website.
I found that the TTGO ESP32 SX1276 can record both SNR and RSSI. I’ll get the device to present me this information during my range testing and as you suggested, record and plot it.
Now for the questions as I didn’t know about this until you posted about it…
Would you be able to clarify the difference between the SF and SNR (signal-to-noise ratio). From my current understanding, the higher the SF the longer the transmissions are and as a result the further the signal can reach whereas the SNR refers to the strength of the connection. Therefore is there a connection between the two measurements? For example, is the SNR value produced as a result of analysing the transmission on a set SF?
And RSSI (received single strength indicator). Is this somewhat the inverse of SNR? For example, is this based on the receiver end of a signal, such as the gateway sending a signal back to the node?
Finally, these measurements are new to me, so I was seeing if you would have any recommended resources about this I could take a look at?
@ElectronicallyE Lets see if we can explain this, let’s start by separating the transmission end from the receiving end. The transmitter is programmed to send data at a specific Spreading Factor. Selecting a SF has a number of outcomes, one is the speed of transmission and for a specific number of bytes, it changes the transmission time and therefore the power consumed. By changing the SF also changes how well the receiver can work. Receiver performance changes with SF, but if the link is short with low losses or there is very low background noise there is no need to select a SF that un-necessarily wastes energy. BTW, I’m am being deliberately vague with the terminology as I’m focusing on the concept without being distracted with jargon. Now at the receive end you not only receive the wanted signal from the transmitter (Signal) but also other unwanted signals (Noise). This noise comes from other transmitters on the same channel (eg Lora neighbours and other ISM transmitters) and other sources of unwanted signal such as RF radiated from electrical equipment etc. In addition transmitters are not perfect, they not only transmit in their designated channel but also radiate much lower levels of energy in the adjacent channels. If this transmitter (on another channel) is close to your receiver will also see this as noise. Hence the term Signal to Noise ratio has for a long time been an important measure, particularly with analogue radio systems. There are other measures used as well, one of these being RSSI. I’ve seen reports that RSSI in a Lora Receiver is not directly measured but calculated and so should be considered as an estimate. (Anybody able to share their knowledge on this?) Receiver performance is based on some of these measures. The example I used in my earlier post is SNR. When transmitting using SF12 the receiver operates down to -20dB SNR and if your receiver records a SNR of -10dB, then you have approx. a 10dB link margin. You then look up the SNR levels for the other SFs. Hope this high level overview puts some of these terms into perspective.
Tony, just a quick question. Is the RSSI based on the the strength of the signal transmitted versus that received or is it just simply based on the strength of the signal received or something else? Just curious, if only one of those is the basis for RSSI, then is the other one I described another form of signal strength measurement and if so, what is it? Hopefully I make sense.
RSSI is the received signal strength. The receiver does not know the power level of the transmitter and so there is no receive to transmit power ratio. This in fact would be path loss which is what a terrain models tries to predict. The other measure is Signal to Noise Ratio (SNR) as strong signal is ineffective if unwanted signal is also being received. Have a look at http://www.loratracker.uk/?p=670
In digital system we also start to look at other parameters, one of these is Bit Error Rate as they also are useful ways to characterise the performance of a system
