What it does

Two exposed metal tracks on a fork-shaped probe, pushed into soil. Wet soil conducts better than dry soil, so the resistance between the tracks falls as the soil gets wetter, and a small comparator board turns that into something an Arduino can read.

It is the same voltage divider as the LDR, with soil in place of the photoresistor. If you have done that tutorial, the electronics here will already make sense, and this page can spend its time on the two things that actually catch people out: what the sensor really measures, and why it destroys itself if you leave it switched on.

Before you begin

Required parts and tools

• An Arduino Uno or compatible board.
• One soil moisture sensor set. This is Siratek product 2841, which comes as two pieces: the fork probe and a small comparator board, with two jumper leads to join them.
• Four more jumper wires.

The two boards

The probe is the fork with two gold-coloured tracks and a two-pin header marked + and -. It contains no electronics at all. It is two electrodes.

The comparator board is the small blue board carrying an LM393, a blue trimmer, two indicator LEDs and a four-pin header labelled:

AO DO GND VCC

The supplied leads join the probe’s two pins to the two pins on the end of the comparator board. Either way round; the probe has no polarity that matters.

Analog or digital

This board gives you both, and the choice matters more here than anywhere else in the library.

• AO is the raw divider voltage. Wetter soil, lower reading, with everything in between.
• DO is HIGH or LOW depending on whether AO has crossed a threshold set by the blue trimmer.

Use AO. Soil is not a consistent material, and “wet enough” depends on your soil, your pot and your plant. The analog reading lets you watch the real numbers and choose a threshold from them. The trimmer makes you guess by eye and hides what is happening.

Fit DO at the end if you need the Arduino pin back. Not before.

What it actually measures

It measures electrical conductivity, not water.

That distinction is not pedantry. Conductivity depends on dissolved salts as well as water, so feeding a plant can raise the reading with no change in moisture at all. Different soils give different numbers at the same dampness. Compost and sandy soil are not comparable.

So a number from this sensor means something only against other numbers from the same probe, in the same pot. Every threshold must be measured where it will be used. A value copied from a tutorial, including this one, is worthless.

It corrodes, and this is the real limitation

Running a current through two metal electrodes in damp soil is electrolysis. The tracks oxidise and eventually the plating lifts off, and the readings drift the whole way through.

Left permanently powered in a damp pot, a resistive probe like this can degrade visibly within weeks.

The fix is to power it only while reading. Instead of wiring VCC to the 5V rail, wire it to a digital pin: turn the pin on, wait a moment for the reading to settle, take it, turn the pin off. A plant does not dry out in a minute, so reading once every ten minutes is plenty, and the probe spends 99.9 percent of its life unpowered.

The board draws only a few milliamps, well within what a pin can supply. The example below does this, and it is the main reason to prefer it to the simpler version.

If you need a probe that will last outdoors for years, the capacitive type is the answer, because its electrodes are not in contact with the soil at all.

Pin reference

Wiring steps

1. Unplug the USB cable.
2. Join the probe to the comparator board with the two supplied leads.
3. GND on the comparator board to the Uno’s GND.
4. VCC on the comparator board to D7. Not to 5V.
5. AO to A0.
6. Leave DO unconnected.
7. Plug the USB cable back in. The power LED stays dark until the sketch drives D7 high, which is correct.

Software and libraries

No library is required.

Minimal working example

Measure your own range first. Nothing else in this tutorial is useful until you have.

// Siratek: find the real range of your probe, in your soil.

const int sensorPin = A0;
const int powerPin = 7;

void setup() {
  pinMode(powerPin, OUTPUT);
  digitalWrite(powerPin, LOW);

  Serial.begin(9600);
  Serial.println("Siratek: try the probe dry in air, then in water, then in your pot.");
}

void loop() {
  digitalWrite(powerPin, HIGH);
  delay(200);                       // let the divider settle
  int reading = analogRead(sensorPin);
  digitalWrite(powerPin, LOW);

  Serial.print("reading = ");
  Serial.println(reading);

  delay(1000);
}

Do three things with this and write the numbers down: hold the probe in dry air, stand it in a glass of water up to the line on the board, and push it into the pot you actually care about, once just after watering and once when the plant is ready for more.

Air and water give you the two extremes. The two pot readings give you the range that matters, and the threshold belongs between them.

Note the 200 millisecond settle after switching the power on. Without it you read the divider before it has stabilised and get a number that is too high.

Watering only when it is needed

// Siratek: check every ten minutes and report when the soil is dry.

const int sensorPin = A0;
const int powerPin = 7;

// Measure your own. Higher reading means drier on this wiring.
const int dryAbove = 700;
const int wetBelow = 450;

const unsigned long interval = 600000UL;  // ten minutes

unsigned long lastCheck = 0;
bool dry = false;

int readSoil() {
  digitalWrite(powerPin, HIGH);
  delay(200);

  long total = 0;
  for (int i = 0; i < 8; i++) {
    total += analogRead(sensorPin);
    delay(10);
  }

  digitalWrite(powerPin, LOW);
  return total / 8;
}

void setup() {
  pinMode(powerPin, OUTPUT);
  digitalWrite(powerPin, LOW);
  Serial.begin(9600);
  lastCheck = millis() - interval;   // check once at startup
}

void loop() {
  if (millis() - lastCheck >= interval) {
    lastCheck = millis();

    int reading = readSoil();
    Serial.print("soil = ");
    Serial.println(reading);

    if (!dry && reading > dryAbove) {
      dry = true;
      Serial.println("dry: water the plant");
    } else if (dry && reading < wetBelow) {
      dry = false;
      Serial.println("watered");
    }
  }
}

Two thresholds again, not one, for the same reason as the LDR: a single threshold makes the state flap whenever the reading wavers. Here it would be worse than untidy, because this is the sketch that eventually drives a pump.

Averaging eight readings smooths the noise. The probe is in damp earth, which is not a quiet electrical environment.

Expected result

A reading that is high in dry air, low in water, and somewhere between in soil. The exact numbers are yours alone.

The power LED on the comparator board should blink briefly once per reading and be dark the rest of the time. If it stays lit, VCC is still on the 5V rail.

Troubleshooting

Next steps

• Easier first: Measuring Light with an LDR, which is the same divider with light.
• Related products: the rain and water level sensor, which works the same way, and a relay for switching a pump.
• Next project: an automatic watering system. Build it with a low-voltage pump and read the relay tutorial’s safety section first.

Sources and attribution

All text, wiring guidance and code in this tutorial are original Siratek work.

• The two-part construction described here, a passive fork probe plus an LM393 comparator board with an AO DO GND VCC header, a trimmer and two indicator LEDs, was read from the Siratek product photographs for item 2841.

Every threshold in this tutorial is presented as something the reader must measure in their own soil rather than copy. The corrosion behaviour of resistive probes is a property of the electrochemistry rather than a fault, and the switched-power approach given here is the standard mitigation. No third-party tutorial text, diagrams, photographs or code has been reproduced here.