Very nice. This post gave me incentive to restart my LoRaWAN testing with some new equipment. On Thursday, January 19, 2023 at 8:30:47 PM UTC-5 Graham Eddy wrote:
> i posted some time ago about my trusty old river level monitor. i am still > receiving occasional enquiries about it - yes, unfortunately that product i > used is no longer available. whilst it is still working for me, i have now > deployed a new river level monitor in parallel, some details below > > *old river level monitor* > > Aquagauge provided a radio receiver with serial interface and a number of > sensor types with radio transmitters. a receiver could support up to 8 > transmitters. the radio and serial protocols are proprietary > > i wrote a weewx service that read the serial interface and inserted the > data values into weewx packets. later i migrated this to a daemon that > emitted the data as mqtt messages, and weewx reads the mqtt messages. this > works fine; it is still in use > > this river level sensor is pressure-based i.e. the probe sits on river > bottom (in my case wrapped around a brick and tucked into a river bank > nook) and measures the water pressure. there is a very inconvenient cable > from the probe to the transmitter. the latter measures atmospheric > pressure, the pressure difference is the weight of the water above the > sensor thence deriving water depth. it proved to be quite accurate > > the transmitter has suffered many travails, the latest being going for a > swim during a flood when the bridge it was on was swept downstream. it is > very robust but now failing more often and i fear one day soon not being > able to rescusitate it → motivation for replacement. this product is no > longer sold or supported > > *new river level sensor* > > i decided to try an ultrasound sensor i.e. the probe measures round-trip > time for a signal reflected from river surface back to probe with known > speed of signal. this means there is no cable needed to run into the river > > my river level sensor must be a commercial off-the-shelf product, for > which i am happy to write interface software. it also has to be wireless > connection to RPi at least 50 meters away and be powered by long-life > battery > > i picked a Dragino LDDS75 Distance Detector ( > https://www.dragino.com/products/distance-level-sensor/item/161-ldds75.html). > i mounted it under the sparkling new bridge i built over the river, higher > than the old bridge. it sends quite large and complex messages via mqtt, so > i wrote a daemon that plucks the salient data and re-publishes as simple > message, imported into weewx. my daemon also configures a fix-point height > for the sensor, so the simplified mqtt message includes a ‘level’ field as > well as ‘distance’ for easy digestion > > this is a LoRaWAN device so provides the long distance and battery > longevity i require > > *radio receiver server* > > the new river level sensor is the first of several sensors on my new open > radio network, as opposed to the closed proprietary radio network per > Aquagauge, so i need a radio transceiver to talk to them. i assume > familiarity with LoRaWAN ( > https://www.thethingsnetwork.org/docs/lorawan/architecture/) > > for the local LoWRaWAN gateway, i put a RAK 2245 LoRaWAN HAT ( > https://www.rakwireless.com/en-us/products/lpwan-gateways-and-concentrators/rak2245-pihat) > > on an RPi 4B and installed RAK’s gateway software from their GitHub > > i tried using the chirpstack LoRaWAN stack running on local server but > couldn’t get it to work reliably, it seemed to work only intermittently. i > think i was having trouble with not picking up the radio packet preamble > but, not being a radiohead, gave it up as too much like hard work. i tried > The Things Network stack ( > https://www.thethingsindustries.com/docs/getting-started/ttn/) and, to > paraphrase steve jobs, ‘it just worked’. i am using the ’community > edition’, which is a subscribed but free service. the downside is that the > network server and join server run on TTN’s hardware over the internet so i > lose device connectivity when internet link is down. they provide the > software to run your own copy of the stack, which i plan to do in future > > so, the river level sensor emits radio packets, my RPi gateway receives > the packets and sends to TTN stack, the TTN stack processes and emits them > as complex mqtt messages, which my dameon picks up and simplifies, then > weewx digests the simplified mqtt messages. (i have chosen to put LoRaWAN > gateway on a separate RPi from the weewx RPi) > > the TTN stack’s mqtt broker is on their remote server. i have chosen to > bridge that mqtt broker to my RPi gateway so that, from my home network’s > perspective, all that LoRaWAN stuff is happening on my RPi gateway. this > fits my plan to move the TTN stack to the RPi gateway in future > > the RAK 2245 includes an itty bitty antenna. i get a good signal and > reliable connection at 100 meters, haven’t tested further distance. LoRa > supports 10 km with a decent antenna > > *additional sensors* > > the LoRaWAN gateway supports a lot of LoRa devices simultaneously, it is > just a case of joining them to the radio network and processing their dat > > i have also deployed a river temperature sensor (replacing the VP2 > wireless temperature station crushed in a storm a year ago). i picked a > multiple-probe Dragino LSN50v2-D23 Waterproof Temperature Sensor ( > https://www.dragino.com/products/temperature-humidity-sensor/item/193-lsn50v2-d23.html) > > to see how well multiplexing on single device is incorporated and, as you > would expect, it is just additional fields in the mqtt messages for each > probe > > *conclusion* > > reading back, it sounds more complicated than it is. i grabbed a separate > RPi, stuck a bit of RAK hardware on it, installed RAK’s software, > subscribed to TTN, configured my RPi into the TTN service, and i had > connectivity. i wrote a daemon to make the sensor data more digestable for > weewx > > what i got was an open, extendable radio network that supports long > distance wireless sensors, for which there are multiple vendors competing > to sell you their sensors > > see https://wx.geddy.au, pull up ’today’ charts > > cheers > *⊣GE⊢* wx.geddy.au > > > > *⊣GE⊢* > > -- You received this message because you are subscribed to the Google Groups "weewx-user" group. To unsubscribe from this group and stop receiving emails from it, send an email to weewx-user+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/weewx-user/0896737f-bcce-4b08-9e6d-420fc62976b7n%40googlegroups.com.
