1
\$\begingroup\$

So after being informed on the Amateur Radio Exchange this isn't quite HAM/Amateur Radio, posting here.

So I'm, well, very new to general RF/radio in general other than basic WiFi, but I'm looking at the feasibility and cost of a long range (>200M between nodes, sometimes probably 500+M) in flat terrain, relatively low power (solar + 12V SLA, short term usage) and probably fairly low bandwidth needs (couple of hundred Kbps?), but bi-directional/fully shared. So in starting doing some searches/research, I came across SX1276 based module (https://www.adafruit.com/product/3072) that's relatively cheap, even if I need it for a few dozen or more nodes.

My concept is to use a fixed mesh, or star mesh topology as the distance between a centrally located node and the furthest away may be as much as 2+ km, possibly 3+. That's long term planning, my first attempt would likely be a handful of nodes perhaps 400-600m from a central point. The advantage of this area is flat as a pancake, although lots of varying height temporary structures, although the antennas would not be placed very high, 6-10 feet off the ground so they may or may not have perfectly clear line of sight.

Based on my reading of LoRaWan, it's really designed to be a central star network topology without much bi-directional communication, in general. And LoRa could, in theory, perhaps, do a mesh network but it'd need a non-trivial amount of software layered on top of the transceiver access code.

I did come across https://ieeexplore.ieee.org/document/8048465/ which is very interesting. In my naive reading of it, it implies that if you simply immediately re-transmit each packet once received (and ignore it once you've re-transmitted it) you essentially can flood a mesh network of LoRa devices reasonably efficiently despite collisions and such. I did see there's more to it than that with seeing to do offset concurrent transmission and other bits that I didn't get to. If that concept actually works reasonably well, it shouldn't be too difficult, in theory, to put a basic layer on top of a LoRa transceiver to spread data through a mesh (that doesn't actually know about each other really) in an Easy™ to implement fashion which spans several square kilometers.

So...disabuse me of this and show me how naive and unknowing I am and I should tuck my tail between my legs and run away :)

\$\endgroup\$
  • 1
    \$\begingroup\$ You could start by checking what kind of bandwidth you actually need. You say "low bandwidth" but a couple of hundred kBps. That's pretty high bandwidth in my world of low power radio. If I'm not mistaken, LoRa is designed for speeds measured in hundreds of bytes per second, tops. \$\endgroup\$ – pipe Jun 8 '18 at 18:11
  • 1
    \$\begingroup\$ LoRaWAN typically assumes the hubs are far more expensive gateway-class radios able to monitor 8 frequencies at multiple spreading factors, something the cheap sx1276 node-class radios cannot do. To build a mesh, you'd likely need mains power to the nodes to permit long receive windows, and either well coordinated clocks for a tightly coordinated frequency agility, or to reduce your usage rates to what would be permitted without changing frequency as much as a LoRaWAN implementation is typically designed to. \$\endgroup\$ – Chris Stratton Jun 8 '18 at 18:18
  • \$\begingroup\$ So, I'm definitely not looking to do LoRaWAN, but will have a relatively large power budget per node. Think RPi hooked up to a solar panel and sufficient SLA battery. And each as a guess right now, but each message might be 3000+ bytes (spread across packets?) that need to be sent/received by each node. Appears I need to learn a LOT more about LoRa, as per electronics.stackexchange.com/questions/277722/… to figure out the data rate. \$\endgroup\$ – Drizzt321 Jun 8 '18 at 19:00
  • \$\begingroup\$ Also, did you read ieeexplore.ieee.org/document/8048465 from above? On my naive reading of it, it might work fine with concurrent transmissions to avoid having to have tight send/receive and well coordinated clocks. At least that's my skimming through the paper thought as to what their findings are. \$\endgroup\$ – Drizzt321 Jun 8 '18 at 19:04
  • \$\begingroup\$ Regularly exchanging 3000 byte messages doesn't really feel like a fit for LoRa. \$\endgroup\$ – Chris Stratton Jun 8 '18 at 19:14

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.