The non functional requirements for this compost measurement system are fairly simple - it should measure the temparature deep inside the compost heap the whole year round; in a setting where there is now power, WiFi and so on.

The approach is based around an Helltech CubeCell AB01 LoraWAN module with a few Dallas Semiconductor DS18B20 temperature sensors. The main one is in the top of a metal pole stuck deep into the compost heap; the other two are in the shaft and on the PBC.

Asmall 5V solar panelkeeps things a LIPO batterycharged through the circuit on the AB01. The schematics are just the connector for the solar panel, the connector for the chain of temperature sensors and a 4k7 pullup for the data signal of the sensors.

See also the full schematics in EasyEDA at OHWLabs. We are using the switched Vext; so that the code on the HTCC can power on/off the power to the DS28B20 sensors. We're careful in the code (see the .INO below) to power them off when we're not measuring.

The business end is a metal pole (pink in below drawing) into which POM insert (green) is ftted; with again a metal tip (blue). With, for some extra sturdyness a threaded rod (hellow) with a hole that exactly fits a DS28B20 sensor (with some heat-sink compound to keep it in place).

The plastic is there to give us a reasoable tip temperature without too much heat leaking away via the steel pole. All is a bit overdimensions to allow us to ram the rod into the compost heap without too much ado. All this thanks to the Lathe at my local makerspace Leiden.

The relatively large thermal mass of the tip is not a large concern; as we're only sending the data every half hour or so.

We feed the wires through the metal rod up.The metal rod is then mounted in a 3D enclosure - with the wires coming out into the inside of the enclosure.  We feed the wire antenna wire through the handle, far away from things metal.

The board is mounted on the lid at the back/bottom of the top bit (blue in above diagram). The pink horizontal pin is simply a bolt that goes through the metal rod to keep it in place; and bring the any force to the plastic body. This is to prevent it causing the shell to bulge under the glass solar panel.

The small indents under the solar panel are to be filled with silicon caucking - this makes keeping things waterproof a lot easier. All this is printed at my local makerspace; on our 3D printer in plain dark red ordinary PLA. That seems to have no UV issues with the realtively feeble sun in the Netherlands.

All this finally looks like this when installed in the compost heap:

Software is pretty much 'stock' Heltech CubeCell example. The setup atthe things network is very standard and follows the wizard for a "HTCC-AB01(Class A OTAA)" setup to a tee.

The one special setup is adding CayennePP as a payload format:

The application has a single webhook configured; with a base URL set to the the MRTG/RRDTool setup behind a trivial cgi bin script in Apache; and a URL set explicitly for the uploads:

This then ultimately yields nice graphs: