Half Bridge SMPS, gapped transformer ?
 

I know in fly-back SMPS I need a gapped transformer but do I need one for a half bridge SMPS?

no, you want as high as possible magnetising inductance . a transformer shall only transfer energy, not store it as the flyback coupled inductors.

thus no air gap. typical primary inductance ie magnetising inductance is in the range of mH for switching transformers operating off-line but it depends of course on A lot of factors.

General rule the magnetising current should not exced 10% of max load current (primary side) but again , it depends.

I will dig up a few links where this is explained more in detail, if you are serious about going to transformer driven converter it is important to understand them to prevent the magic smoke from escaping or worse...

I have come up with a circuit based on IRS27951 app notes and have laid out a pcb but was just wondering about the transformer gap.
The app notes give turns for transformer but it didn't mention a gap.
I intend to use a ETD54 core.


My previous project was a fly-back SMPS.

Any help with stopping smoke would be appreciated.

Air gap is only needed when a continuous DC magnetizing field is inside the core: Fly Back, Buck, Boost, polarity inverter, and forward converters using 1 or 2 transistors. And it is sometimes used a very small gap in push pull topology if there exists risk of DC core imbalance because of different on times, off time, DC wire resistance or RDSon in FET´s or VCEsat in BJT´s or IGBT´s.

But in HB mode converters, either hard or resonant, the blocking cap ussually connected between the center main bulk caps and the transformer, prevents core imbalance, so no gap is needed, in fact make it a worse transformer.

This design has a DC blocking capacitor.

well that is a resonant converter and they are different from ordinary forward topology, the series inductance is typically made from a large leakage inductance (I am not 100% shure about this) but nevertheless there should be no air gap.

Ok, then no gap at all.

The "cold" wire of the transformer, is returned via a non-polarized cap to the middle point of two bulk caps in series, or one to each rail?

Quasi resonant usually uses the leakage inductance, normal or deliberately included in the transformer.

No. But some people end up with one because they have the wrong number of turns.

Even with a capacitor, the complete absence of air-gap can lead to some form of periodic flux walking.
The simplest cure is to include a very marginal air gap, even if it looks counter-intuitive.
See this document p 4-6:
http://www.ti.com/lit/ml/slup126/slup126.pdf
Damping can also be a solution but a small airgap is free from a technical and BOM perspective.
IEEE Xplore - Transformer saturation and unusual system oscillation in capacitively coupled half-bridge or full-br...

Although very very small, in normal ferrite cores there exists a gap between halves of the core. Except in toroids with no distributed gap.