High-frequency transformers - how are they different?

[Plat]

I apologize for being somewhat scatter-brained here, I just have too many questions on this topic.

I know high-frequency transformers use ferrite cores instead of laminated silicon-steel, but what other differences are there?

Does their reactance still follow the standard formula: 2*Pi*F*L ?

What defines the upper operational frequency for a particular HF transformer?

How would the design of a 30kHz unit differ from, say, a 300kHz unit?

What frequency is most common is SMPS, and why?

 

[davenn]

but what other differences are there?

generally they can be substantially smaller
size is in general inversely proportional to frequency
so knowing that, the only thing that then would make a higher frequency core/transformer larger,
would be power handling requirement

What frequency is most common is SMPS, and why?

that would be so easily answered with a quick google searchDave

 

[Plat]

How does the core material affect the frequency it can be used at?

Is the advantage of ferrite its lower magnetic permeability? And is this an advantage because it results in lower inductance and therefore lower inductive reactance?

I don't know how that could make sense because even with a little less inductance, the reactance would still be massive at high frequency. Is this not really a problem? I'm just trying to wrap my head around all this.

 

[Baluncore]

Higher frequency transformers have relatively smaller core sections and less turns.

The core material dimension is primarily decided by frequency of operation. Skin effect decides how far the magnetic field can penetrate the core before the AC field is reversed. There is no advantage in having grains or laminations that are greater than twice the skin depth. If the material is thicker than needed then the winding wire will need to be longer, the resistance will be greater, so mass, cost and heat will all rise.

To get magnetic field energy in and out of the magnetic core material requires access to the internal magnetic surfaces, which is why there needs to be insulation between the magnetic components. An EM wave propagates through the insulation over the surface of the magnetic material. That insulation dimension is usually about 10% of the magnetic material dimension.

Grains or laminations are NOT there primarily to prevent eddy currents. They are there to give access to the magnetic core material. It is the sensible orientation of laminations that reduces eddy currents, at the same time as providing magnetic access to the core.

 

[Plat]

Higher frequency transformers have relatively smaller core sections and less turns.

The core material dimension is primarily decided by frequency of operation. Skin effect decides how far the magnetic field can penetrate the core before the AC field is reversed. There is no advantage in having grains or laminations that are greater than twice the skin depth. If the material is thicker than needed then the winding wire will need to be longer, the resistance will be greater, so mass, cost and heat will all rise.

To get magnetic field energy in and out of the magnetic core material requires access to the internal magnetic surfaces, which is why there needs to be insulation between the magnetic components. An EM wave propagates through the insulation over the surface of the magnetic material. That insulation dimension is usually about 10% of the magnetic material dimension.

Grains or laminations are NOT there primarily to prevent eddy currents. They are there to give access to the magnetic core material. It is the sensible orientation of laminations that reduces eddy currents, at the same time as providing magnetic access to the core.

That's very very helpful, thank you.

Are there any inherent/unavoidable losses as frequency increases that cannot be 'tuned' away in the transformer's construction? For example, could a transformer that was properly designed to operate at 200kHz be just as efficient as one that was designed to operate at 20kHz?

Just to be totally sure, the inductance of the core material plays no role in the performance of the transformer at a particular frequency, right?

 

[Uriah_Heep]

There are inherent losses, yes. These can be basically divided into 1) electrical losses and 2) magnetic losses.
1) The electrical losses can not be avoided since the windings will always have a series resistance, Rs. You can minimize it by improving conductor properties or by increasing wire diameter. But you will never achieve zero resistance. Thus, there will always be some ohmic losses.
2) You can improve the quality of the magnetic core, but since the core is not ideal, alternating current will result in alternating magnetic flux and hysteresis loss. The higher the frequency, expect higher losses. Same for eddy-current losses. Losses can't be tuned away.

Yes a transformer at 200 kHz can be as efficient as a 20 kHz transformer. But you need better materials (with less losses at 200 kHz) and improved design procedures to account for additional effects that may rise from the frequency increase (such as parasitic capacitances).

About one of your initial questions, some commercial SMPS work at 500 kHz nowadays. A typical Dell or HP laptop charger operates in the range of 100 kHz - 300 kHz.

I am not sure what do you mean by "inductance of the core material plaus no role..." A core doesn't have inductance. Windings do. Cores have magnetic permeability.