All of the capacitive methods listed on my previous log have something in common, they need a capacitive probe where the soil acts as the dielectric core.
I found that coplanar electrodes is the best way to emulate a capacitor using cooper pads on a printed circuit. In the case of a double sided pcb, we need to use the same polarities on both sides of the pcb, in order to ignore the material that the pcb is made of (FR4, paper fiber, etc).
Is important that the geometry of the board allows it to be placed easily on the ground. Also, all the probe body needs to be water isolated because the material the pcb is made of can store moisture on the long run (FR4).
Doing this can result in other problems, the same material isolating water from the board, can begin to store water, causing drifting problems measuring capacitance.
More info about this, here: http://www.deepdarc.com/2011/09/08/square-one/