Inductor/transformer size vs frequency?

[crapalanche]

Hello all,

I've got a question about inductor size (in inductance, but also thus physical size, generally) vs frequency for DC-DC converters and likewise transformer size vs frequency. I know that a faster switching frequency allows you to have smaller inductors and transformers. I don't know why though--that's the question.

Does the faster frequency allow the components to transfer more power? Is this due to the non linearity of amount of magnetizing current over the time constant for an LR circuit? Do practical circuits keep the frequency so that they transfer current only within some time constant of a LR series circuit? With a fixed frequency, would using less inductance (when you have an excess) improve efficiency?

How does reactance fit into this? With a higher frequency, there is more reactance, thus a decrease in magnetizing current and less power transfer per cycle. Does the increased frequency make up for this? Is there a point where (minus the losses directly resulting from high frequency) frequency is so high that there is very little power transfer?

Thanks

 

[Grossel]

One point that should be pretty obvious - Then more amount of iron in the core, the more magnetic energy can be saved before the iron core saturates. That will have an effect on frequensis. For low frequensis, a little transformer will saturate faster than a big one (assume other conditions is similar).

[edit]
Other (old) relevant thread: measuring inductor saturation

 

[RCinFLA]

The reason the core size goes down with higher frequency is because the amount of energy storaged per cycle goes down for same power. More cycles of lower energy per cycle equals equivalent power.

The down side of running higher frequency is the input capacitance of MOSFET's requires more gate drive power. Power for MOSFET gate drive is C x V^2 x F, where C is the equivalent input capacitance and V is gate drive voltage of MOSFET. Peak drive current to move capacitance is Qg/risetime, where Qg is the bundled equivalent gate input charge which is combination of Cgs + Cdg x Miller effect.

You also have to select core for acceptable losses at the higher switching frequency. The magnetizing inductance is required to be less to account for higher frequency.

 

[crutschow]

It's basically a matter of the magnetizing inductive reactance which is inversely proportional to frequency. You want to keep the magnetizing current low to minimize power loss and avoid saturating the core. The higher the frequency, the lower the required inductance for a given inductive reactance and magnetizing current, thus the smaller the required core and/or number of turns on the windings.

Magnetizing current is a normal parasitic byproduct of the transformer inductance and the applied voltage level and frequency. The amount of power that can be transferred through a transformer is usually limited by the transformer winding resistances and is unrelated to the magnetizing current. Thus core size goes up at higher power levels due to larger required wire size, not due to any core limitations.