This repo is dedicated to the baby turbidity sensor. This sensor compromises on accuracy for low-cost and it was developed for a class on environmental sensing.

This very simple turbidity sensor consists of the DFRobot turbidity probe, it is housed in a very simple PVC housing, and it's programmed in the Arduino framework. The "anatomy" of the sensor is shown in the figure below (forgive the German).

The senseBox and batteries (figure below - LEFT, MIDDLE) are inside of a home-made PVC housing. The housing was closed with a big cork, which is not the best for waterproofing. The DFRobot probes are housed in a transparent housing and these transparent plastic pieces were glued to our PVC housing (figure below - RIGHT ... I wish I had a better photo of this but I don't). This was done so that the turbidity probes can be directly connected to the senseBox while staying dry and still taking measurements through the small DFRobot transparent housings.

Below you can see some results obtained using these sensors for three different time periods in April 2020 in the Holderbach river in Zurich. The top panel in each figure shows the turbidity (in NTU) measured by both sensors (blue and orange curves) since our device had two DFRobot turbidity probes.

The middle panel in each figure shows the solar radiation (yellow curve) and the precipitation in mm (blue bars). This data was provided to the students.

The bottom panel in each figure shows again the solar radiation (yellow curve) and the average NTU of both probes (blue curve). We see some very interesting findings int eh figures below.

The figure to the LEFT shows that the NTU measurements are influenced by solar radiation but also around April 12th we see some spikes in turbidity which correspond to precipitation. The MIDDLE figure shows no precipitation but there is anyways a constant increase in NTU. Perhaps this is because the sensor is getting dirty? The last figure to the RIGHT shows increases in NTU corresponding to precipitation, also in the absence of solar radiation.

These results show that the sensor is able to measure changes in turbidity, but also that these measurements are affected by solar radiation (see section below). The students also measured the turbidity in two more Zurich rivers (the Sihl and the Limmat) and those results can be found in the folder "Results."

The senseBox uses the Arduino framework so the Arduino library is a good place to start with example codes for SD card usage. Or the Adafruit library for the ADC. SenseBox also has example codes for each of their products. But I have also included the (very basic) code to obtain the turbidity in the folder Codes called "turbidityCalibrated_sensebox.ino.

If you would like to read out from the temperature and/or humidity sensor then you can check out Codes/temperature_humidity.ino. And if you would like to save to an SD card then check out Codes/SD_temp.ino.

To calibrate the sensor(s), follow the procedure outlined in Calibration/calibration.pdf. Below you can see an example of the calibration curve our students obtained (LEFT) and the polynomials they fitted to their curves (RIGHT).

As the figure below demonstrates, these cheap DF robot sensors are affected by sunlight which is why we calibrate in dark-room conditions (please forgive my hand-drawn units... the students forgot to add the units and I don't have access to the original figure).

• This sensor doesn't have a wiper so biofauling would be a problem and the sensor shouldn't be installed long-term without regular maintenance
• The cork isn't the best for a waterproof housing

For questions, please email me at [email protected] or @rivertechjess on Twitter 🐦