I have an issue with an IOT project covering the city of Enschede. I have six Sodaq Ones gathering temperature data using the DS18B20 digital temperature sensor. I use a 4.7k pull-up resistor as is indicated and connected it to 3.3v. The system worked fine, but now they are starting to read -127 and are completely unable to communicate with the DS18B20 sensors. I have tried connecting different temperature sensors (hijacking still working nodes' sensors) without avail. Once a Sodaq stops communicating with the sensor, it is incapable of reconnecting.
Things I tried:
Checked the connection of the sensor
Checked the sensor
Checked battery voltage (it's 4.3v)
Checked the software (even the most basic version does not work)
Checked other boards (sensor works)
This is the second batch of sensor nodes I have created, the first batch used the Sodaq One revision 2. This has been running the same basic code for the past year and a half without ever losing a beat. But having switched to revision 3; the whole thing is unstable. Three sensor nodes have stopped being able to interface with the temperature sensor (they have not lost power, they are still sending the data over LoRa)
The only change I have made is using the integrated solar charging circuit instead of having created my own. It is possible that this does not have overcharging protection? The sun has been very bright the past days. Could this break the Sodaq's internal logic? It might also have been the heat, but one Sodaq broke after having brought the sensor nodes back in to a airconditioned environment. So these things might not be the culprit at all.
I'd appreciate new ideas, I can't think of anything new to try.