Dielectric Plant Guard


So I started this project several years ago, in 2013 if I remember right. I got a plant from my friend and it almost DIED, this was not good, I had to fix this in some way. So I started to read a lot about soil moisture measurement, several Phd articles. And started to build several prototypes, and wrote a lots of code. It ended up with 5 prototypes that worked well, but they were to expensive and the year after several plant monitors showed up on the market. To bad, I thought.

But now in 2017, when I have worked as a Developer engineer for some time and have learned a lot in the area of PCB, programming, construction, mechanics and theory. I started up this project again. 9 days after I had the new Mark-III prototype, and it works!, but there were some small misses that I had to fix. And the code was not optimal.

But I post my project here I case you want to build your own plant monitor or just lying in your bed, surfing on some strange things and end up here, then you are just welcome to read, but you should be sleeping, you have work to morrow (fuckers)!

Several types of measuring the soil moisture

  1. With your finger (or toe)
    1. Good, due to the fact that you almost always have your finger and toes with you
    2. Bad, it gets dirty, and you need to go deep for it to be a valid measurement.
  2. Resistive mesurent with two electrodes
    1. Good, it easy and cheap!
    2. Bad, it poison the plant, due to electrolyte, and metal eroding from the anode part.
    3. This is the one that is sold as a Arduino soil sensor, it so BAD, and you will notice that the metall on the senior will erode dur to electrolyte, and in the end you will have metals-poisoned your plant!
  3. Capitative mesurment (most common way)
  4. Pro: It is cheap, easy, saves the plant, and understandable.
  5. Cons: Not as accurate as I want
  1. Frequency analysis
    1. Pro: Accurate, saves the plant, cool way to measure, power saving.  (Its my way to measure 😉 )
    2. Cons: Hard to program, PCB layout is sensitive to this method. More expensive.

Frequency analysis

So, its simple, the frequency changes due to the Dielectrium in the soil, hence there for the name of my product, Delectric Plant Guard (DPG). I sample the frequency and measure the dry, sand an soil, then water. Thees to measurements will be stored as upper and lower thresholds, (0-100% moisture)

Im using a base frequency of 166kHz, and according to Nyqvist, I have to sample at least twice as fast in order to yield a safe measurement. without any folding. Using Atmel MCU, using there timer to count the pulses from the freq sources and then calculate the amount of water.

First I started with the MCU Atmega 328p, and all worked well, one can use Arduino code to make life simpler. But I wanted to optimize the product so now I’m using ATtiny, lower cost and smaller, and started to use AVR studios and C-code in order to go deep in the MCU.


Hardware Design

First breadboard design

One must start some where, so just with simple breakout boards and whirring the first test circuit was made.

Using arduino UNO as a MCU.

Bread board development of the Dielectric Plant Guard


PCB Design


Well it’s a little lie that it is Mark-I, cos Mark-I, i did a mistake on that one… It did not work at all… So it’s actually Mark-II on the image… Shhhh, no one needs to know 😉

Here is a image of the first PCB, as you can see the important thing is to yield a good antenna. This PCB was ca 120mm long and was mor or less Mark-I and Mark-II. It worked well and gave a good ground to optimize the product. As I stated i used Atmega 328p, and used the Arduino Bootloader in order to get the project and coding going fast forward. This first design had a test-button on it, but as we all know, buttons are expensive and not really necessary.

The first PCB design Mark-I

I mean, it looked grate and worked well, but the cost, damn, so expensive to produce

Mark-I in the action, and one of the better images of the project



 So, Mark-III i did now in 2017, but just 5pcb for prototype, after tuning in the antenna the latest one is Mark-IV, and thats the one I will review deeper on this project side.

Compare to Mark-I and -II Mark-III and -IV vad designn with Altium Designer, I have the software at work, due to the fact that I work as a  developer Engineer. Now, the cobtrol of liberarys and footprints are so much bettee in Altium. Though, Altium is bit yet Mac compatible…

Anyway, Altium have standard 3D visulation, which is awsome!

3D visulation in Altium Designer



Image below shows the  base-frequency, 166.7 kHz. And the next image show the frequency at 100% moisture (in a glas of water)

0% of water, shows 166.7kHz



50% of water, shows 7.3kHz


Here is a video, showing how the period changes due to various exposer to water. Note that one can measure water level with the Dielectric Plant Guard also.



Another video of the complete product, showing how the LED changes color when watering. Note that in this video I have sat the sample time to 4.8 seconds, but when in ordinary use, the sampel time is every 15 minutes.


[Bild på ocilloscope]




E-postadressen publiceras inte. Obligatoriska fält är märkta *