Here
      are some initial concepts for testing soil moisture for the robot
      to determine if the plant is to be watered. Here is the full write up of
      the moisture computer!


As a starting
      point, a circular plant tray, numerous small pots and soil, and
      a plant moisture meter was purchased. This meter has four levels
      of moisture and a chart on the back for determining which level
      for which type of plants. Measurements are made by plunging this
      meter into the soil next to the plant and the analog meter reads
      in four zones. It works by generating a small amount of measureable
      current in a bi metal sensor at the bottom when in the conductive
      soil. So its like making a battery using two dissimilar metals
      in a cup of salt water - a small battery is made with a tiny
      amount of current. Now this is not the same as measuring the
      soil resisitance like I will be doing but is directly proportional.
      My goal is to pinpoint the four zones and find the equivelent
      resistance for measureing with a micro controller.
 

      
The proto type
      probe is an active system, which you apply 5 volts from the PCB
      which contains the processor. It generates a current through
      the resistor and converts to a voltage to be remotely measured.
      Output is 0 - 3.5 v depending on moisture conductivity. It is
      very sensitive and just by holding the probes by hand, you can
      get nearly a volt out of it.

      
Probes are stainless
      steel #6 screws, three inches long. This avoids electrolytic
      corrosion deposition common with such probes made from brass
      or common steel that can rob your sensor of its accuracy. The
      goal is to measure the soil, display with 4 LEDs the level equivelent
      to the meter probe and transmit via optical link to the robot
      which will be parked near by.



      

Click to enlarge!

      
Schematic of my
      moisture probe. I may change the sensitivity by varying the NPN
      emmitter resistor, but this works pretty well so far.

      
Early test results:
      The soil was mixed with various amounts of water to yeild readings
      on the purchased moisture meter from 1 - 4. Then each time I
      took a both a resistance and circuit output measurement. The
      results were quite variable, the readings varied from 3M Ohm
      to 2.5 M Ohm. But there was little correlation to soil moisture!
      In other words, nearly dry soil was unreadable as expected, but
      moist soil compared to total muck was nearly the same values.
      More testing....

      
And now a complete
      write up on the plant moisture sensor can be found HERE.


BACK TO ROBOT PAGES

HOME