IC1 ……. CD74HC132
R3 ………. 3.9M
……… Two AAA alkaline cells, with holder
R5, R6 ….. 680
C1, C3 … 1nF (0.001uF) or 2.2nF (0.0022uF)
R7 …….… 15
C2 ……… 100uF/16V electrolytic
R8 …….… 47K
C4 ……… 220nF (0.22uF)
D1 …….… 1N4148 or
R1 ……… 470K (all resistors 1/4W, 5%)
D2 ………. MV8191 or
R2, R4 … 100K
Q1, Q2 ….. 2N4403 or 2N3906
Using a Veroboard
mother-board about the same size as the battery holder, a
daughter-board was added to hold the remaining parts:
1) IC1D is a CMOS Schmitt trigger oscillator at about
2KHz. It starts and continues to oscillate with a supply
down to 1.24V (the lowest output voltage of my LM317
variable power supply) or less.
2) IC1A is an inverter.
3) IC1B is a Schmitt trigger NAND gate. Its output is low
only when both inputs are at, or higher than the upper
Schmitt trigger threshold voltage. With 47 ohms or less
between the probes, an input is always low, so the output
is always high. With a resistance of only R8 between the
probes, the voltage across C3 is high most of the time,
so the gate’s output is low for ½ the oscillator’s
period. With a resistance that is halfway, then C3 is
charged high by that resistance when the oscillator’s
output is high, then is discharged when the oscillator’s
output is low. When C3 is being discharged, then pin 12
of the gate is high, and pin 13 is also high until the
discharging voltage of C3 reaches the lower Schmitt
threshold voltage. During this time, the gate’s output is
low. So the low time of the gate’s output depends on the
value of the resistance between the probes. This is
Pulse-Width-Modulation of the low output of the gate.
4) IC1C is another CMOS Schmitt trigger oscillator at
about 2Hz. D1 and R4 discharge C4 quickly so that its
output is low for only about 15ms with a 3V battery, and
about 25ms with a 2V battery.
5) The series connection of Q1 and Q2 performs like a NOR
gate, so that the LED lights only when both inputs to the
transistors are low.
6) R7 is a current-limiting resistor for the 1.8V LED.
With a 3V battery, the LED current is about 35mA.
1) When the soil is very dry, the LED flashes brightly,
since the soil’s resistance is very high.
2) When the soil has been watered a few days before, but
is drying, the LED flashes dimly,
3) When the soil is damp because it has been recently
watered, the LED is off.
Note that different soils have a different resistance.
Also, sometimes, watered soil will continue to have a
high resistance until the soil absorbs the water, a delay
of about one hour.
Although the LED’s current is 8mA with a 3V battery, it
is lighted for only a maximum of only about 1/64th of the
time, so its maximum average current is only 550uA. The
remainder of the circuit draws 200uA. The total is 750A
for new batteries, and about 250uA for run-down
batteries. Therefore the exponential current of 300uA
will continue with 1000mA/hr batteries for 2000 hours, or
about 4.6 months.
The LED’s current is logarithmic with the soil’s
resistance, so that when the resistance is one-half, then
the LED’s current is one-tenth. If you water the plants
when they need watering, then the average LED current
will be very low, and the batteries should last for about
1) Try to obtain the very bright and wide-angle LEDs that
are listed. Samples are available from Fairchild.
2) Use tinned copper 1.5mm diameter buss-bar wire about
8cm long for the probes.
3) Use silicone caulking to attach and seal the Veroboard
to the battery holder, and to seal the battery holder’s
4) Perhaps the project can be mounted in a plastic bottle
for pills, available from a pharmacy (chemist?), with the
probes sticking out of its lid.
A list of plants and their watering requirements is here:
Download this project in .doc format
Also check the conversation about this project at the
community. Post you questions here.