Testing the prototype in soil

Testing the prototype in soil

.Soil Moisture Sensor Project

Testing the prototype in soil

Other people may test something bit by bit.

They may test it under ideal conditions for 24 hours to see whether it is stable.

These people may then change one environmental variable at a time to see what effect that variable has on it.

Once they have developed a model for how it responds under ideal conditions to changing environmental variables they will test it.

But not me. Not because I am cavalier, or because I know better than them but because I haven’t got the time to run and write up that many tests.

So, here goes.

This is half of the experimental setup. Soil from the garden, with some water poured over the top, and a few seeds of “kitchen greens” (whatever they are) dropped on top. The sensor is attached to the two yellow wires that run from the oscillator into the soil.

The plan was to use the radio receiver to scan across a range of frequencies every 10 minutes. After a few days of the soil slowly drying out, I hoped to see that the frequency was slowly increasing over time.


After a few days I saw this….

Yes, this is the result of almost 72 hours of waiting.

Three things stood out straight away.


The first thing that stands out are three noticeable decreases in frequency that I’ve circled in red. What is interesting about them is that they are roughly the same amount of time apart. And there were three of them in three days.

The second thing that was interesting was the fact that after the drop, the frequency was a little bit higher. I have drawn a red line to show that the overall frequency seems to increase over time.

The third interesting point is the fact that there are two periods where the signal appears to jump around a little bit. I’ve pointed these out with arrows.

Possible explanation

Over time the frequency seems to increase. But given how much the measured frequency jumps around, I think it is too early to claim that this trial has been successful. We need to try and explain the noise in the data.

The fact that there are three decreases in the frequency over three days suggested that the decrease had something to do with the time of the day. What I forgot to mention is that the experimental setup is in an east facing room, on the south wall, next to a window. This means that it gets hit by morning sunlight (southern hemisphere, currently winter as I write this). I started the test at about 10pm one night and the first decrease in frequency was several hours later. This suggests that the effect occurs in the morning.

Looking at the raw information from the sensor supported this observation. The decreases all occurred between 8am and 10am. This is when the sunlight enters the room and would hit the experimental set up.

But I can’t blame this effect on sunlight. The third interesting thing was that there were periods where the frequency seems to jump around. These occur in between each of the large decreases in frequency. Looking at the raw data revealed that they occurred between 5pm and 10pm. This has been the time of the day that we’ve needed to put the heater on in the house.

From these two observations, I would hypothesise that the frequency is being affected by heat. But I can’t tell you whether the heat is having an effect upon the oscillator or upon the sensor.

That will be a test for another day.

Main Menu