[whiskey]

Hi all, I'm attempting to use a modified version of the speedy basic distributor code to read rpm of a motor.
But I get weird results when I approach zero and go back up from there. If i bring the speed up slowly the rpm oscillates between what it should read and half of what it should read. If i do it quickly the reading is perfect

No idea what I've done wrong, Im not very good at programming.
Any help is appreciated.

Code: Select all

unsigned long curTime;
unsigned long toothLastToothTime = 0;
unsigned int triggerActualTeeth = 8;

volatile unsigned long triggerFilterTime = 0;
volatile unsigned long curGap = 0;

volatile unsigned int rpm = 0; // Initial rpm value
volatile unsigned int cur_rpm = 0;
const byte rpm_in = 2;

bool hasSync = false;
volatile unsigned long toothCurrentCount;
volatile unsigned long toothOneMinusOneTime;
volatile unsigned long toothOneTime;
volatile unsigned long syncLossCounter;
volatile unsigned long startRevolutions;
volatile unsigned long toothLastMinusOneToothTime;
volatile unsigned long currentLoopTime;
unsigned long MAX_STALL_TIME = 500000;

void setup() {
  // put your setup code here, to run once:
  Serial.begin(115200);

  pinMode(rpm_in, INPUT); //Interrupt 0
  attachInterrupt(digitalPinToInterrupt(rpm_in), triggerPri_BasicDistributor, FALLING);
}


void loop() {
  // put your main code here, to run repeatedly:


  Serial.print("RPM =  ");
  Serial.println(rpm);
  //Serial.println(" ");




  currentLoopTime = micros();
  unsigned long timeToLastTooth = (currentLoopTime - toothLastToothTime);
  if ( (timeToLastTooth < MAX_STALL_TIME) || (toothLastToothTime > currentLoopTime) ) //Check how long ago the last tooth was seen compared to now. If it was more than half a second ago then the engine is probably stopped. toothLastToothTime can be greater than currentLoopTime if a pulse occurs between getting the latest time and doing the comparison
  {
  noInterrupts();
  if (hasSync == true)
  {
    rpm = 60000000UL / (curGap * triggerActualTeeth);
  }
  interrupts();
  }
  else
  {
    //We reach here if the time between teeth is too great. This VERY likely means the engine has stopped
    toothLastToothTime = 0;
    //toothLastMinusOneToothTime = 0;
    hasSync = false;
    startRevolutions = 0;
    rpm = 0;
  }


  } void triggerPri_BasicDistributor(void)
  {
    curTime = micros();
    curGap = curTime - toothLastToothTime;
    if ( (curGap >= triggerFilterTime) )
    {
      if (hasSync == true) {
        triggerFilterTime = (curGap * 3) >> 2;  //Recalc the new filter value
      }
      else {
        triggerFilterTime = 0;  //If we don't yet have sync, ensure that the filter won't prevent future valid pulses from being ignored.
      }

      if ( (toothCurrentCount == triggerActualTeeth) || (hasSync == false) ) //Check if we're back to the beginning of a revolution
      {
        toothCurrentCount = 1; //Reset the counter
        toothOneMinusOneTime = toothOneTime;
        toothOneTime = curTime;
        hasSync = true;
        startRevolutions++; //Counter
      }
      else
      {
        if ( (toothCurrentCount < triggerActualTeeth) ) {
          toothCurrentCount++;  //Increment the tooth counter
        }
        else
        {
          //This means toothCurrentCount is greater than triggerActualTeeth, which is bad.
          //If we have sync here then there's a problem. Throw a sync loss
          if ( hasSync == true )
          {
            syncLossCounter++;
            hasSync = false;
          }
        }

      }
      toothLastMinusOneToothTime = toothLastToothTime;
      toothLastToothTime = curTime;
    } //Trigger filter
  }

[jonbill]

What are the code modifications? Are you running a speeduino hardware and firmware, or have you just taken the decoder to use in a different context?

Also, what is "approach zero" in numbers? Less than 200rpm?

[whiskey]

Only minimal code modifications, taking out parts that aren't relevant. It's in the post above. Using V0.4 hardware but just measuring rpm of an electric moter

[jonbill]

Tbh, if you're not confident coding, why not just run speeduino as is?
(Although below 200rpm I wouldn't be surprised if there are problems, since engines never really turn slower than that)

[JHolland]

If you only want to read the RPM of an electric motor then read up on input capture, the way that Speeduino works is not the best way to read RPM.