IR2110PBF only shorts when switching phases

Question

Here is my circuit:

When I start on phase 1 by PWMing LIN & HIN of U2 and send a HIGH signal to LIN of U1, then the circuit works perfectly and only one LED lights up.

When I try to switch to phase 0 (after starting on phase 1) by PWMing LIN & HIN of U1, and send a HIGH signal to LIN of U2, then the MOSFET bridge suddenly shorts and breaks both of my IR2110 ICs.

BUT if I turn my DC power supply down to 0 V, reset the Arduino to start on phase 1, and turn the DC power supply back up to 13 V then the other LED lights up without a problem.

For some reason my MOSFETs are only shorting when I try to switch phases; if I start on either phase 1 or phase 0 then it works perfectly.

Does anyone know why this would happen? I am implementing 10 μs of dead time between switching phases so I don't think shoot-through would be the problem.

Here is the code I am running:

byte preg = 2;
byte sec = 0;
byte pres = 0;

void setup() {
  pinMode(D7, OUTPUT);
  pinMode(D6, OUTPUT);
  pinMode(D12, OUTPUT);
  pinMode(D11, OUTPUT);
  digitalWrite(D7, LOW);
  digitalWrite(D6, LOW);
  digitalWrite(D12, LOW);
  digitalWrite(D11, LOW);
  // A5 is connect to a button so I can switch phases when 
  // pressing the button
  pinMode(A5, INPUT);
  // I have an LED attached to A0 to help me keep track of 
  // button presses
  pinMode(A0, OUTPUT);
  digitalWrite(A0, LOW);
}

void loop() {
  // This code is to debounce a button press when for switching 
  // phases
  sec = digitalRead(A5);
  if (sec == 0) {
    // Wait 1 second to check if button is still pressed
    delay(1000);
    sec = digitalRead(A5);
    if (sec == 0) {
      pres = digitalRead(A0);
      if (pres == 1) {
        // Set phase 0
        preg = 0;
      } else if (pres == 0) {
        // Set phase 1
        preg = 1;
      }
    }
  }
  
  // Switch to the new phase when button is pressed
  switch(preg) {
    // Phase 0
    case 0:
      digitalWrite(D7, LOW);
      digitalWrite(D6, LOW);
      digitalWrite(D12, LOW);
      digitalWrite(D11, LOW);
      digitalWrite(A0, LOW);
      delayMicroseconds(10);
      analogWrite(6, 100);
      for (int i = 0; i < 10000; i++) {
        phase0();
      }
        break;
    // Phase 1
    case 1:
    digitalWrite(D7, LOW);
    digitalWrite(D6, LOW);
    digitalWrite(D12, LOW);
    digitalWrite(D11, LOW);
    digitalWrite(A0, HIGH);
    delayMicroseconds(10);
    analogWrite(11, 100);
    for (int x = 0; x < 10000; x++) {
      phase1();
    }
      break;
  }
}

// PWM U1 LIN & HIN alternatively to charge C4 & C5
void phase0() {
  digitalWrite(11, HIGH);
  digitalWrite(12, LOW);
  delayMicroseconds(10);
  digitalWrite(11, LOW);
  digitalWrite(12, LOW);
  delayMicroseconds(1);
  digitalWrite(11, LOW);
  digitalWrite(12, HIGH);
  delayMicroseconds(150);
  digitalWrite(11, LOW);
  digitalWrite(12, LOW);
  delayMicroseconds(1);
}
// PWM U2 LIN & HIN alternatively to charge C2 & C3
void phase1() {
  digitalWrite(6, HIGH);
  digitalWrite(7, LOW);
  delayMicroseconds(10);
  digitalWrite(6, LOW);
  digitalWrite(7, LOW);
  delayMicroseconds(1);
  digitalWrite(6, LOW);
  digitalWrite(7, HIGH);
  delayMicroseconds(150);
  digitalWrite(6, LOW);
  digitalWrite(7, LOW);
  delayMicroseconds(1);
}

Answer 1

Finally figured this out after a lot of trial and error. Turns out the problem is my code. If you do analogWrite(pin, value), then try to turn the PWM off with port manipulation using PORTD / PORTB then the analogWrite will stay on and it is not effected by the port manipulation.

You need to instead do digitalWrite(pin, LOW), or analogWrite(pin, LOW) and it will turn off the PWM.

Edit:

In the answer i'm not using PORT manipulation in my code because I was trying to make it easier for people to read, but on my actual project i was using PORT manipulation so it was a problem.

Answer 2

The IR2110 cannot maintain a continuous "HI" output, the bootstrap capacitor will quickly (less than 1 second) discharge through the 10k resistor.

You've used a blocking delay so the MCU goes off into never-never land for at least one second when the switch is pressed and the previous MCU outputs (whatever they are) are maintained,

I don't know if that's your problem, but I think it's something that needs to be fixed. Better to use a hardware timer for debouncing. Also, 1 second is very long for debouncing, 10ms is adequate, but I think you'd need different logic- I think the way you have it now it just toggles back and forth every second as long as the switch is pressed.

Answer 3

For reference.

EE&O ...

The H-bridge driven by IR2110 (and others) has four "phases" for working.
Can be simplified .. Link.

LIN1 and LIN2 are High. HIN1 and HIN2 are Low. "Brake" phase LOW.
LIN1 and LIN2 are Low. HIN1 and HIN2 are PWM_ON. "Brake" phase HIGH.
LIN1 is High and HIN2 is PWM_ON, LIN2 is Low and HIN1 is Low: motor "rotate" "left".
LIN1 is Low and HIN2 Low, LIN2 is High and HIN1 is PWM_ON: motor "rotate" "left".