In order to make a pause or wait for the task in Arduino, the most widely used function is delay(), but sometimes it was cause some problem, in this article it will discuss it, and shows how to use function millis() to replace it.

Example code

For example, we'd like the board to flash the LED precisely every 0.5 second.

It will be very easy to write the code with delay function:

int ledPin=2;

void setup() {
    Serial.begin(115200);
    pinMode(ledPin, OUTPUT);
}
 
void loop() {
    digitalWrite(ledPin, HIGH); // Turn the LED on
    delay(500); // Wait for 0.5 seconds 
    digitalWrite(ledPin, LOW); // Turn the LED off 
    delay(500); // Wait for another 0.5 seconds
}

Now, if we insert some codes, let the board send some unused message to the serial port.

int ledPin=2;

void setup() {
    Serial.begin(115200);
    pinMode(ledPin, OUTPUT);
}
 
void loop() {
    digitalWrite(ledPin, HIGH); // Turn the LED on
    Serial.println("LED is ON");    
    Serial.println("LED is ON");
    Serial.println("LED is ON");
    delay(500); // Wait for 0.5 seconds 
    digitalWrite(ledPin, LOW); // Turn the LED off 
    Serial.println("LED is OFF");    
    Serial.println("LED is OFF");
    Serial.println("LED is OFF");
    delay(500); // Wait for another 0.5 seconds
}

You will find the LED flash time is more than 0.5 seconds. That is because it needs some time to execute the Serial.println() function. But we don't know how long it will take.

Then, we will use millis() function to implement it:

int ledPin = 2; 
unsigned long previousMillis = 0; // Store the time of the last LED toggle

void setup() {
  pinMode(ledPin, OUTPUT); // Set the LED pin as an output
}

void loop() {
  unsigned long currentMillis = millis(); // Get the current time
  if (currentMillis - previousMillis < 500) {    
    digitalWrite(ledPin, HIGH);
    Serial.println("LED is ON");
    Serial.println("LED is ON");
    Serial.println("LED is ON");
  } else if (currentMillis - previousMillis < 1000) {
    digitalWrite(ledPin, LOW);
    Serial.println("LED is OFF");
    Serial.println("LED is OFF");
    Serial.println("LED is OFF");
} else{
    previousMillis = currentMillis;
}
}

In above code, it will compare current CPU time to saved previousMillis to get the delta time. While the delta time:

  • 0~0.5S, turn on the LED
  • 0.5~1S, turn off the LED
  • At time >= 1S, it will reset the previousMillis.

Different between delay() and millis()

  • delay() Specifies program pauses a number of milliseconds.
  • millis() On the other hand, it returns the number of milliseconds elapsed since the program started.

So we know that delay() is a relative time clock. While millis() is an absolute time clock. And the most important things that delay() will pause the execution of other codes.

There are two main advantage to use millis other than delay:

Get the exact time

The first advantage we will discuss is accurate timing. With millis()us, we can ensure that the loop executes as many times as needed, regardless of the execution time. With delay()this it's impossible because we don't know how long one loop takes to execute.

Such precise timing is useful when sampling or running filters at a certain frequency.

No blocking

Another benefit millis()is that it doesn't prevent us from running code while "waiting".

delay() would pause the entire code until the time is reached.

micros() and overflow

Just like delay()there is a microsecond version called delayMicroseconds(), millis() has micros(). If you need better resolution, you can use micros().

But, be aware that micros() will overflow about every 70 minutes, and millis() about 50 days. This means that the function's return value will start at zero again.

Summary

millis() and micros() is a very convenient function, which can be used when processing timed tasks. If you don't know these things, you'll probably end up just using them delay(), which may also works well.