Measure square wave frequency

Hi Matt, I don't know about this situation but usually sampling over a longer time period will improve accuracy.

John

John is right, there I said it!

You might be able to get more accurate timing by latching the wdt upon interrupt.

Reset the Arduino Due board using the watchdog - Arduino Forum

I tried setting mainPeriod to be longer  but it still didn't measure accurately lower frequencies.

Regarding the wdt, I don't understand how can it help, can you elaborate?

thanks.

Here are some DSs for early timer chips.  This functionality is typically integrated today, but is should give some basic insight as to what is going on in uP timing.  The more sophisticated your OS, the more difficult timing with your CPU becomes:

http://liris.cnrs.fr/~mmrissa/lib/exe/fetch.php?media=mc6840.pdf

http://www.z80.info/zip/z80ctc.pdf

try checking this out, sparkfun has a whole project but it is open source so yo can use and implement in your own way.

key thing to look at is the sketch as it will have a working solution even if you display it in a different way

https://www.sparkfun.com/products/10140

Here is a great cookbook article on the topic of square waves.  No direct answer to your question here, but everyone who enjoys this site could probably learn something from it, just came across it today.  From our great friends at EDN:

http://www.edn.com/design/test-and-measurement/4432218/Boost-and-modify-square-waves

This is the type of circuit where it may be good to maybe offload the core of it to external logic. Basically an external counter to measure the period, and the Arduino can read the counter once it is latched. There used to be tons of 10MHz and 50MHz frequency counter circuits doing exactly this, in 74xx logic gates and no Arduino, and displaying the output on LED displays. If you can't find anything and wish to pursue this method, let me know - I'm sure I can dig up such a circuit either on the Internet or in a book.

EDIT: Although, just noticed, you're only interested in low single-figure kHz range? If so, Arduino should be sufficient.

// mills() reference http://arduino.cc/en/Reference/millis
// attachinterrupt() reference http://arduino.cc/en/Reference/attachInterrupt
// Due pin 0,1,2 corresponding to pin interrupt 0,1,2 respectively
// All the best from RAVI BUTANI
// this code measures frequency only accurate for >100Hz
// for less frequency minor modifications needed


unsigned long time_stamp;
unsigned int count0 = 0;
unsigned int count1 = 0;
unsigned int count2 = 0;


unsigned int output0 = 0;
unsigned int output1 = 0;
unsigned int output2 = 0;


void setup()
{
  Serial.begin(9600);
  time_stamp = millis();
  attachInterrupt(0, sensor0, RISING); // sensor 1 on mega pin no 2 Due pin 0
  attachInterrupt(1, sensor1, RISING); // sensor 2 on mega pin no 3 Due Pin 1
  attachInterrupt(2, sensor2, RISING); // sensor 3 on mega pin no 21 Due Pin 2
}


void loop()
{
  if(millis() - time_stamp > 50)
  {
    detachInterrupt(0); // disable interrups
    detachInterrupt(1);
    detachInterrupt(2);
    output0 = count0 * 20;
    output1 = count1 * 20;
    output2 = count2 * 20;
    Serial.print("Sensor0 Reading = ");
    Serial.println(output0);
    Serial.print("Sensor1 Reading = ");
    Serial.println(output1);
    Serial.print("Sensor2 Reading = ");
    Serial.println(output2);
    delay(40);
    count0 = 0;
    count1 = 0;
    count2 = 0;
    time_stamp = millis();
    attachInterrupt(0, sensor0, RISING); // we have done with this mesasurenment,enable interrups for next cycle
    attachInterrupt(1, sensor1, RISING);
    attachInterrupt(2, sensor2, RISING);
  }
}


void sensor0()
{
  count0++;
}


void sensor1()
{
  count1++;
}


void sensor2()
{
  count2++;
}

Thanks for your reply.

I tried that, it measures 1500Hz and higher quite accurately but anything below it and it fails.

Do you know what else can be done to fix this?

Thanks again.

did you checkout the spark fun link I provided, what was the result

Thanks