I'm trying to improve my point-to-point communication between two Lora device (one MKR 1310 and one ESP32 with Lora module). My goal is to transmit over about 1 km (which works fine), but with a small hill between the two locations. With standard settings, the communication is OK a few meters from my ideal locations and I feel that an optimization of the Lora parameter could make the communication feasible at this ideal place.
In order to evaluate the best parameters, I'm recording the SNR (Signal to Noise Ratio) as indicated by the Lora receiver using the library function and tried to optimize from a shorter location (this is not practical to go from one place to the other to change the lora settings of each device).
Unfortunately, this is quite hard to find a reliable indicator. For example, I switched the SF (Spreading factor) from 11 to 12 and the SNR decreased, whereas this should be the opposite (verified by moving the two devices appart, the SNR indeed goes down). Is the SNR comparable when changing SF?
Other question, I'm considering changing the bandwidth from 250 kHz to 125 kHz but I'm not convinced this would have an impact of the achievable transmission. Would that have a strong impact on the transmission capabilities (length and potential to deal with obstacle in the line of sight)?
The signal to noise ratio has nothing to with receiving noise. The noise of concern is the noise generated in the receiving components themselves. If the test also includes the received noise, then the ratio is signal to noise+noise. The receiving noise generation is from ALL the components in the receiving circuit, including all resistors, all wiring, the antenna and the solid state devices. The data sheet for the receiving device should give an appropriate dB number for the noise in that device.
Thanks for the answers on SNR. I assume this will be hard to use as a meaningful indicator.
"If there are 'obstructions' then that is a standard UHF comms issue, there is no LoRa magic that can get around such issues"
Well, if I wanted the signal to be send through a mountain, I would agree with this statement. Now I assume there is a continuum between "no transmission" and "perfect transmission" depending on the amount of obstacle to deal with. A reed would create manageable disturbances as opposed for example to an oak tree forest.
So in other words my question would be : are parameters advised for long range also advisable for "small obstacles". I tend to think yes, so would look at 125 kHz or even 62.5 kHz bandwidth and a SF of 12. Does that make sense?
Your reasoning is way too narrow. Weather also affects your transmission path. That includes all types of moisture and dust, as well as temperature. There are also reflections that occur and can be single or multiple.
Your best signal path will be found using directional gain antennas.
The biggest factor is noise caused by man-made devices, like LED lights that use switching power supplies.
Obstacles can massivly reduce the signal level seen by a receiver, often so much that there is no reception at all.
In such circumstances you can try the longest range parameters which might work. But long range also means much slower transmissions, with longer air times and potential legal issues because of duty cycle limits.
SX126X LoRa devices normally have a TCXO to set the frequency and can be used down to bandwidths of 7.8khz, but air times become really long with more chance of interferance corrupting reception.
The antenna(s) will have a big impact on your range. The antenna must fit to the frequency. Don't expect to much from the antennas delivered with the modules.
To conclude this topic, I managed to have the transmission I was aiming for (800 meters with a hill just on the boundary of the signal path). I changed several parameters so I do not know which one is the most important:
bandwidth : 250 kHz => 125 kHz
SF still 11 (for some reasons I could not communicate with 12 with my devices even on short distance)
One important thing also is that with some power supplies I was using to power the ESP32 (standard phone charger with USB cable), the transmission did not work. This could have been the problem, I will check later. I am using rapsberry pi power supply with the appropriate usb connector.
I also changed the receiver antenna. Not sure this made a difference or not neither.