Teensy 3.6: interrupt configured for falling edges fires on both edges

[paoloboatto]

I'm working with Teensy 3.6 and I set the interrupt function with the following simple code:
attachInterrupt(digitalPinToInterrupt(RPM_Pin), RPM_PulseTimeDifference, FALLING);
I'm using a signal generator to send a square wave to the RPM_Pin, and check it with the oscilloscope. I found that the interrupt on that pin, fires on both the rising and falling edges, as if I had set the interrupt with the CHANGE option instead of FALLING.
Am I missing some info about the interrupts on Teensy 3.6?

 

[el_supremo]

That "shouldn't happen", but how do you know it is also triggering on the rising edge?
Post some demo code.

Pete

 

[paoloboatto]

That "shouldn't happen", but how do you know it is also triggering on the rising edge?
Post some demo code.

Pete

I check the time differences with arduino micros function and compare with the oscilloscope's signal

 

[el_supremo]

I'd still like to see the code you are using to compute the time differences.

Pete

 

[manitou]

does the tick count from this sketch increase by 1000 each second? (jumper pin 12 to pin 23)

volatile uint32_t ticks;

void ding() {
  ticks++;
}
void setup() {
  while (!Serial);
  analogWriteFrequency(23, 1000);    //jumper 23 to 12
  analogWrite(23, 128);
  attachInterrupt(12, ding, FALLING);
}

void loop() {
  Serial.println(ticks);
  delay(1000);
}

try counting ticks with your signal generator.

 

[PaulStoffregen]

I'm running that test here. Indeed it is incrementing 1000 times per second, so only 1 interrupt per cycle.

 

[paoloboatto]

I found that there is a systematic jitter on the signal; this is why I'm counting two falling fronts instead of one