Your level sensor might be measuring the weather.
We recently tested a level sensor where the actual water level wasn’t changing – the measured level was.
Overnight it moved 15 mm peak-to-peak, closely following temperature. Analysis showed a temperature dependence of approximately +1.17 mm/°C. And this was indoors, with only a 5.5°C temperature swing.
Rather than change the sensor, we characterised the error and added a temperature compensation lookup table in a Senquip script.
The result: a significantly more stable level measurement.
Sometimes getting better data doesn’t require a better sensor – it requires understanding the sensor you already have.
