Legionella temperature monitoring hardware, designed and built for you.
If you want to offer continuous temperature monitoring for water systems and have no embedded team, we design the sensor node, write its firmware and set up the radio and cloud pipeline behind it. We have taken a four-channel LoRaWAN node from schematic to a field trial.
Just want monitoring deployed, not a custom build? See Paqett →
The problem
Water hygiene guidance under ACoP L8 and HSG274 includes temperature checks. The published HSG274 limits are hot water at 50 °C or above within one minute and cold water at 20 °C or below within two minutes. Checking that by hand means a visit and a reading, and anything between visits goes unseen.
Continuous monitoring closes that gap, but it needs a battery-powered sensor with working firmware, a radio link and a cloud pipeline behind it. That takes hardware, firmware and radio engineering at once, which a company without an embedded team rarely has in house. We design the device and everything behind it, and we tell you plainly how far each part has been proved. Our own trial is an example. One node went silent for seven days after it entered a sleep state it never left, because it had no watchdog to reset it. We found that on a real site, fixed it and wrote it up in the case study.
What we've built
4-channel LoRaWAN temperature node
A battery-powered node with four temperature channels, built on a LoRaWAN module and sent through a gateway to a network server and database that we also set up. It is a trial of one unit, not a product, and there is no fleet or customer deployment behind it. Mechanical and enclosure design for your site would be scoped with you at the start.
Designed from schematic to field trial; trialled in an operational plant room
| Layer | What we did | Detail |
|---|---|---|
| Electronics | A custom, battery-powered PCB around a RAK3172 LoRaWAN module (STM32WL). | Four temperature channels · STC3100 battery fuel gauge |
| Firmware | RUI3 firmware built with arduino-cli. It is written to recover on its own if the node stops responding (verified with a forced-hang test), because a sensor nobody visits has to look after itself. | Independent watchdog (32 s) · sleep capped at 20 s · join back-off from 30 s up to 1 h · progress markers in backup registers, read over ST-Link · two vendor patches, to sleep and timers |
| Radio | LoRaWAN, with an indoor gateway running Basic Station into the network server. | LoRaWAN · indoor gateway · Basic Station · after a gateway outage, a single reset rejoined the node in 5 s |
| Cloud | A ChirpStack network server and a telemetry database, on a self-hosted server. | ChirpStack network server · telemetry database · both on one self-hosted server |
How working with us goes
- 01 Scope We would typically start with which points need measuring, what battery life you need, what radio coverage exists on site and what you want to see in the end.
- 02 Prototype Schematic, board layout and firmware. We would typically bring the board up on the bench and then with a gateway and network server behind it, so the whole chain is tested and not only the board.
- 03 Trial A unit running on a real site, so that faults show up and can be fixed before anything is repeated. On this node, progress markers in the firmware let a hang be read back afterwards.
- 04 Production Design for manufacture and a production run, if the trial results justify one.
Related
Damp and mould monitoring ·Water leak detection ·New-build specification ·IoT product design ·LoRaWAN sensor design ·Embedded firmware development · All sectors →
Need a monitoring device designed and built?
Custom hardware, firmware and software, from first sketch to production.
Talk to us about your project alessio.grespi@gmail.comWant it deployed, ready to run?
Paqett, our own monitoring company, deploys and runs white-label water hygiene monitoring, for companies that would rather buy it than build it.
See Paqett →