The relationship between dead time and minimum pulse width

[fullmoon]

Hello

Currently, 5777C is in use.
The PWM dead time is given at 3us, and the minimum pulse width is given at 3us.
When operating like this, PWM does not go out.
However, if 3.005us, which is 5ns more than the dead time, is applied to the minimum pulse width, PWM is good.
May I know the detailed reason?
Also, may I know why the minimum pulse width should be longer than the dead time?

Thank you.

[petervlna]

Hello,

Also, may I know why the minimum pulse width should be longer than the dead time?

Well, the dead time is the time where the transistors are switched of. So no PWM is generated. So your dead time should not be same length as the active period of PWM.

Please do a little due diligence for example here:

https://granitedevices.com/wiki/Dead-time_distortion

Best regards,

Peter