[55] Overkill sensor

A few years ago, I built a sensor to help me water my plants (see project n°33).

It recently broke, so rather than repair it, I decided to design a new one — something elegant and easy to repair.

I started by taking the old sensor apart and measuring each of its components.

I wanted to avoid adding a physical switch, to keep the design as minimal as possible.

The idea was simple: a sliding shell that would serve two purposes:

• It protects the electrodes when closed.
• It switches the device on when opened.

To see whether the idea would work, I started with a prototype.

Have I mentioned that prototyping is my favorite part of the process? It’s when ideas come to life, and when anything still feels possible.

With all the dimensions defined, the concepts validated, and the overall layout clear, I was ready to model everything in Fusion.

I tried to keep the entire design in a single file. It made things a little more difficult at first, but much easier to maintain in the long run.

Then came the part I was dreading most: the electronics. Soldering isn’t exactly my strong suit, and my tools don’t make the job any easier.

For this project, I had to rebuild a few boards, resolder some connections, and remake some connectors.

Get Project 67’s stories in your inbox

Join Medium for free to get updates from this writer.

Remember me for faster sign in

It could have been more compact, but I wanted the finished device to be fairly chunky. Making it too small would have made it harder to handle.

The device has two LEDs: a white one to indicate that the soil is dry, and a blue one to show that it’s wet.

To protect them, I decided to try using hot glue. The result was better than I expected: the glue gives the LEDs a glass-like finish, but can still be easily removed and replaced.

As I mentioned in the introduction, the sliding mechanism also acts as the power switch:

• Two electrodes rest on small rubber pads on the fixed shell.
• A metal rod is mounted on the moving shell.

When the moving shell is open, the rod bridges the two electrodes, completing the circuit and switching the device on.

I also did a little refactoring of the original code. Apparently, I had initially planned to use three LEDs instead of two.

But I can’t find any trace of that plan in the code. My memory clearly isn’t what it used to be.

Assembly was by far the easiest part: I just had to screw a few things together and plug in a couple of connectors.

The design turned out to be very easy to repair!

I also like how minimalist the finished object looks: just a monochrome tube, with no visible switch.

All in all, it’s definitely overkill for a problem nobody had — and I love it!