What is shoot-through, and why is dead-time necessary?
Shoot-through is when the upper and lower switch on the same leg of an H-bridge conduct together for an instant; a short circuit forms between the supply and the chassis and the high current pulse can destroy the transistors. Dead-time is the dead interval placed between the PWM edges, and with silicon switches it is typically 100–300 ns.
The problem arises because switches are not ideal: after a MOSFET receives the turn-off command, going fully into cut-off does not take zero time. If the turn-off command to the upper switch and the turn-on command to the lower switch are given at the same instant, the lower switch turns on while the upper one is still partly conducting, and a low-resistance path forms between the two rails.
The consequences:
- A short current pulse that can reach hundreds of amps, and sudden junction heating.
- Noise on the supply line, excessive stress on the bus capacitor.
- With repetition, cumulative fatigue and eventually failure of the switch as a short circuit.
The solution is to wait a short dead interval after turning one switch off and before turning the other on. During that time both switches are off and the current in the motor inductance freewheels through the body diode of the MOSFET. That is why the reverse recovery behaviour of the body diode matters: a slow-recovery diode increases both the switching loss and the risk of shoot-through.
If the dead-time is chosen longer than necessary, the output voltage is distorted and torque ripple increases; so the dead interval is balanced between safety and accuracy. It is not a parameter for the user to set, it is fixed in the driver design.
The full topic: DC Motor Driver Selection Guide