Water Sensor Behavior - Full Immersion Test

I’m hoping to use the Grove water sensor to measure how the water in my tank rises and falls. But I’m not getting consistent results, especially if I immerse the entire sensor. Am I using it correctly?

The plot below shows the results of a simple test I performed today using a Spark Core to measure the water sensor’s output. I’m using 3.3V (filtered) as Vcc, and putting the signal directly into an analog input of the Spark Core (no buffering, or pullups/pulldowns.) For testing purposes I’m using tap water in a glass next to my desk.

It works great if I just immerse the tip (tree traces) of the sensor. But I get wildly inaccurate results when I immerse the entire sensor. That’s unfortunate, since in my application the water level will go well above (and below) the sensor.

Is there something I should be doing differently?

Any ideas? Or how to get a better signal?


— Malcolm

Just to be clear. The first and second segments are perfect. Third segment (fully immersed) is getting weird, but I can live with that. But things fall part when I try to go back to the dry (4th) or partial immersion (5th segment).

— Malcolm

Hi Malcolm,

This Grove water sensor is not meant for water level detection. It can only indicate if the sensor is in dry state and wet state.
If the sensor in a fully immersed in water it will short Vcc and GND.

Please feel free to ask any other questions.


OK, thanks. Wasn’t clear from the specs (and the water level design that first directed me towards your site.)

– Malcolm