Moisture Sensors

The moisture sensors were described in the phase 2 research

Construction

I created them by first machining a mold out of machinable wax.  This was the first part I've machined on my CNC mill.  It is a two part mold.  It creates a simple cylinder with a rounded end.  The mold is shown below.

I created the machinable wax myself by mixing 75% paraffin wax with 25% LDPE (low density polyethelene).  I don't think I let it stew long enough; I cooked it for about an hour.  It is recommended to be cooked for 3-4 hours..  The whitish colouring is small bubbles, probably from the lack of cooking.  They tend to make the plaster stick to the mold a bit making it difficult to remove.

To make the sensor wires consistent, I created small PCB spacers to hold the wires the same distance apart.  These are shown below.

Then the wires were placed in the mold and filled with plaster of paris, according to the directions on the packet.  Below is an action shot of the mold at work.

Performance

Aggh!  Not again.  Something isn't quite right with the sensor.  It has been installed, two actually, at 10cm and 40cm depths.  Below is a snap shot of the first few days of sensing captured from the live online monitoring system.

The two green lines are the moisture lines.  The solid line is at 10cm, the dotted line is at 40cm.

The first problem is that it took more than 2 days for the moisture level to settle.  I'm not sure if this is normal.  Maybe the gypsum blocks hadn't dried enough?  Maybe they are too wide (the new blocks are 25mm wide, the old one 20mm)?  Maybe the wires are spaced too far apart?  Maybe the PCB spacers and the soldering is having an effect?  Maybe my concerns about grounding are a problem?  Maybe I need to change the load resistor?  Plenty to play around with. 

The next problem is the reaction to moisture.  The small glitch between 23 Mar and 24 Mar is when the garden was watered.  The moisture increased, but then quickly settled.  Again, I have no idea what went wrong here, if there is a problem.