Full Bridge LLC converter, 4kW cannot be done with offtheshelf ferrite?

[cupoftea]

Hi,
Do you agree that a 4kW Full Bridge LLC converter, with vout = +/-180Vdc and Vin = 390Vdc (PFC output) cannot be done with any offtheshelf gapped ferrite core sizes? (That is, cores that can be bought ready_gapped)

AYK, the biggest gapped ferrite core available offtheshelf is the ETD59 which is offtheshelf gapped to 2mm (AL = 311). However, if that is used as resonant inductor or magnetising inductor of such an LLC, then its core losses will be too high (approx. 10W).

Would you agree with this? Or can you design one which doesn’t suffer such high core losses when only offtheshelf gapped cores are used.

The attached shows the design document for such an LLC. Done at 70kHz….the core losses come out at 10W for the resonant inductor, and 10W for the magnetising inductor, which in both cases uses the ETD59 TDK core with N87 , ready gapped with 2mm gap. These core losses are too much, would you agree?

 

[mtwieg]

I certainly wouldn't say it "cannot be done". 20W of core loss for a 4kW supply seems quite small (maybe unrealistically small), especially if the TX functions as both Lr and Lm. But how much power dissipation the core and windings can tolerate depends on other factors such as ambient temperature and airflow, which you didn't mention.

Increasing fsw is another option; 70kHz is not very high for an LLC. Depends on your semiconductor limitations of course.

Using discrete Lr/Lm would also help, of course. You're asking the transformer to do much more than transformers generally do in other topologies.

Also, how did you estimate 10W for Lr? In principle, Lr should not induce core losses, since it's not coupled to the other windings.

 

[cupoftea]

Also, how did you estimate 10W for Lr?

Using the graphs on core loss for N87 material, with volume of ETD59

https://www.tdk-electronics.tdk.com/download/528882/71e02c7b9384de1331b3f625ce4b2123/pdf-n87.pdf

In principle, Lr should not induce core losses, since it's not coupled to the other windings.

Thanks...this uses an external Lr.....AYK, the current induces B (delta B) which acts in core loss.

I have heard on this forum, that 6W is max core loss for an ETD59...unless it has a fan right on it. (actually , it was said, 6w core loss and 6w winding loss, max)

There are fans in this supply, but they are on the semiconductors, not directly on the ferrites

 

[cupoftea]

BTW, the LLC transformer is not a problem with core loss, as we can select a big core as it doesnt need gapping......it doesnt need gapping because we are using an external magnetising inductor......it is the external magnetising and resonant inductors which mean that its not possible to do this with available offtheshelf gapped ferrite cores.....would you agree?

AYK, offtheshelf gapped cores are only available for ferrite cores of size ETD59 downwards.....you cannot buy , for example, a gapped E70 core.

This is the problem.

And this is why a 4kW LLC full bridge resonant converter, from a Vin of 390Vdc+ cannot be done with offtheshelf gapped ferrite cores....would you agree?.....the core loss is simply too high in Lr(ext) and Lm(ext).

 

[Easy peasy]

I have seen a very plausible and build-able design for 112.5kW using standard OTS cores ...

 

[cupoftea]

Thanks yes, though were those cores (any of them) gapped?...and if so, were the gaps offtheshelf available?
(....eg like an ETD59 with 2mm gap, AYK, its available to buy gapped like that off the shelf.)