Current density in smps transformer windings

[eem2am]

Hi,

Is 6A/mm^2 ok for smps transformer windings.....84KHz switching frequency.....natural convection cooling

 

[simon7000]

Do you have a data sheet for the core?

 

[sawreyrw]

No, my old rule of thumb is 400 A/cm^2. This is for a few (or 1) layer(s) where the heat can easily be removed from the windings. If you have multiple layers, then you need to reduce the current density.

 

[Serge66]

sawreyrw is right.
I use 3-4 A/mm^2 in a normal transformer and the same for ferrite.

 

[eem2am]

A transformer manufacturer just recomended 12A/mm^2 for an etd44 ferrite core in a 90-265vac full bridge at 84khz.

he told me not to worry about skin effect cuzz the freq was too low

 

[Th3 uN1Qu3]

You'll find this document useful: www.coremaster.com/appnotes/an107.pdf

Even though it's for a supply with a different topology and purpose, the transformer calculations stay the same. You will see that knowing the dimensions of the transformer they simply use the mean length of a turn to calculate dissipations in the wire. I use a similar method. Now a bunch of people will chime in and say there's proximity effect, skin effect, core losses, and so on. Well, core loss is another parameter that you can get from the manufacturer, and either way, calculating the copper loss and taking that into account in your design is still better than just using a constant.

 

[johansen]

I would think that 12A/mm^2 will give you copper loss two orders of magnitude higher than core loss.