Dead time needed for Full Bridge converters?

[cupoftea]

Hi,
The max duty cycle of this UCC28085 full bridge driver is 49% (translates to 98% "real" duty cycle).
Do you agree, this is absurd?
In no way should you ever operate a Full bridge that could, at any time, go to 98% duty...because it would mean one fet turning on very soon after the other fet turned off....and the ringing in the leakage inductance could mean the on-coming fet turns on whilst the diode of the off-going fet is conducting the leakage ring ....giving a massive reverse recovery current spike...which will eventually lead to power supply failure.
You can get round it by massive snubbing of the primary, but surely adding a duty cycle clamp is the way to go? (unless of course, dare i say, the aim is to make PSU's that fail sooner, then they are re-purchased, quids in!)
Do you agree with this?

UCC28085

https://www.ti.com/lit/ds/symlink/ucc28083.pdf

 

[KlausST]

Hi,

A couple of issues..

* You say it´s a full bridge driver.. but datasheet says "push-pull" or "half bridge".
A full bridge needs to drive 2x HIGH side and 2x LOW side. But the device can drive just 2 x LOW side MOSFETS.
* duty cycle is specified to be less than 50%. a dead time of 110ns typically
--> better post the schematic you are taking about to get a more focussed discussion.
* you say "massive reverse recovery". How massive it is depends on dead time, rise time, current, reverse recovery behaviour.
100ns is not uncommon.

I don´t see a generally "reduced lifetime"

Klaus

 

[Easy peasy]

You clearly have not undergone a detailed study of the full bridge, when one diagonal leg turns off the transformer voltage will begin to swing the other way - it is entirely possible to trigger the other mosfets at any tine after this transition begins - but ideally when the volt swing has reached its maximum for lowest turn on losses ( at full power ) this may be only 50nS - which is 98 - 99.5%