LoRa under water

Dear TTN-fellows,

I’ve spent a couple of hours going through relevant posts on this forum and I found the bargain threads very useful:

For one of my application, I need the LoRa endnode to be as tiny as possible since we want ot put it ontoa mussel with a biology chair from some university.

So there are for sure small lora modules:

I also like this idea of using the opposite side of a hope module:
And the STM32WLE chip has been announced but only very few people got their hands on it:

So right now, the solution of taking a hoperf module and make another pcb on the opposite side with a MCU and a battery seems to be the smallest configuration.

Looking forward to your ideas!

1 Like

Presumably you’ll need something very waterproof. And with a teeny tiny antenna that can get a signal to a gateway …

as antenna I was thinking of just taking a cable, that would be acceptable, the cable doesn’t have to fit into the housing. Or I can use a standard 868MHz antenna wiht an ipex connector.

I guess the RF traces of the hope module are well documented so when designing an IPEX connector, there should be some design guidelines.

I’ve followed the antenna topic also, will need to take a look at it again.

Thank you!

How will you be coping with the underwater signal losses?

Are you looking for an end node (as in a complete device?) or just a LoRaWAN module?

You might also take a look a Onehorse’s TIlera Sensor Tile or Gnat https://www.tindie.com/stores/TleraCorp/

Also look at e.g. ACSIP S76S as small footprint module

…havent suggested any SX126x products as may be a bit early yet wrt LoRaWAN Stacks/Development environments etc but keep your eye (and favourite search engine!) open looking there also.

The signal losses are severe, I’ve estimated them already. Of course, they strongly depend on the conductivity of the water but we are fine with bridging 1-2m already, which shouldn’t be a problem. Another solution would be getting as close to the water surface as possible with the antenna of course.

Thank you Jeff, that’s really a nice finding:

I think I’ll either go for the hoperf module and make a little custom PCB to solder on the backside or continue working with the STM32WLE Nucleo board that I have. However, I’m just starting with the STM32CubeIDE and the MCU isn’t yet officially supported.

Thanks for your time and effort again.

1 Like

There’s already a great thread here on TTN: Small LoRaWAN nodes

Maybe better close this one then.

Is this for use with LoRaWAN or LoRa point to point?

In LoRaWAN terminology the official name is ‘end device’, in practice also called ‘node’ (but not ‘endnode’).

Realistically you’re not going to get meaningful range from underwater to any conventional LoRaWAN gateway at a distance type of installation.

If you want to place “gateway” specifically to cover a shellfish bed, there are probably more suitable schemes; in particular, you probably want to look at the lowest available frequency, as you’ll get better penetration of water. Maybe even something with long range RFID type stuff.

Though it’s a bit unclear what sort of information you hope to obtain from an individual mollusk. Underwater conditions are likely better obtained by some sort of stake driven into the mud with sensors below and radio above, well clear of the water. Tracking the actual mollusk population is probably numerically and economically best done by some sort of scanning or tagging, not trying to put a fairly expensive radio on each one.


20 x 20 mm. HopeRF module on the backside. AtMega328 processor.


everyone will ask - where to get one, when, how much, design open source?..etc! :wink:

Price depends on volumes, € 25.00 ex. VAT for one-offs, around € 17.50 ex. VAT for smaller amounts (10-100). The design is not open source in the way that we provide gerber files and such to produce your own PCB, but obviously we do provide electrical schematics needed to write software for it. It’s Arduino compatible. 4 programmable I/O-lines have been provided on the outside, and selected in a way that you can create an i2c interface, have an analogue input and we also provide an interrupt input. And of course: it has an LED :wink:

Such a thing is necessarily larger than the integrated MCU+Radio modules already on the market, because the RFM95 itself is larger at 16mm x 16mm than something like the Murata at 12.5 x 11.6. Also an ATmega328 is drastically limiting compared to the MCUs in integrated modules and typically means people cut corners and only partially implement LoRaWAN.

But its unclear that any LoRaWAN solution is going to work well for the asker’s underwater application…

@cslorabox I wasn’t claiming it’s top of the line :wink: but it fulfills our purposes and it’s cheap. And also important for us: the GND plane is still big enough to use a helical antenna on it and have an acceptable RF performance.
As to underwater use, the formula gives that the theoretic maximum underwater penetration for 868 MHz for salt water is 8 mm, and for fresh water it’s 31 cm. That of course also requires that the geometry of the antenna is adapted to the dielectric constant of salt resp. fresh water.

1 Like

If you put the gateway antenna right over shallow water you might get something, but you won’t get any meaningful range to a gateway at a distance in the usual LoRaWAN manner.

As such, LoRaWAN on conventional frequencies makes no sense for such an application.

If someone wanted to put a local reader over a shelfish bed, they should probably look at something like low frequency RFID schemes or something like that.

@ecosoph reach out to R. Daniel Harris. He’s done similar work with snails and oysters.

Thank you @salvagione I’ve already sent him a mail. He focuses more on the biology part, which is great.

@bluejedi it is or use with LoRaWAN

@cslorabox I would be happy with a few meter 3-7m under water already, which depends strongly on the water conductivity, as I’ve estimated. It might also be an option to pull the antenna larger to the water surface (sth. like a floating antenna).
I’ve looked for low-frequency alternatives, ther’s a group at the Iowa university that uses 315MHz transmission:

RFID would for sure be the best option but to my knowledge the wire would need to be extremely long and if badly designed the losses there are heavier than for a bad designed 868MHz antenna. Anways, I see the advantage of using 13.56MHz or even lower frequencies with 133kHz that is used in pet tagging.

I’ll be thrilled to hear your feedback :slight_smile: you seem to have some practical experience.

@rharte that’s cool and extremely small. I would prefer a 64KB RAM ATMEGA processor over the ATMEGA328P but if your whole code its onto it, then it is just great.

Underwater needs are pretty clearly not an application for LoRaWAN or UHF frequencies, but rather for the low MHz to more likely high Kilohertz range.

“Antennas” are more likely to be coils than wires.