I have designed fly-backs and inductors and understand the basics about flux density, saturation and core losses. What I have never been very clear on is design of SMPS transformers. I get the concept that ideally you want to transformer to not hold any magnetic energy and instead to transfer it all, but what material properties allow this?

I’m assuming you want to have a high permittivity \$\mu_o\$ to transfer flux better but also a linear \$\mu_o\$ so there is a linear transfer of energy, what else? I have spent time over the years looking for good tutorials or information on this topic and never found anything very complete, there is a lot of information on inductors, or sometimes the application is not really clear but someone with a bit of experience can discern between true transformer design vs inductor based on if the design holds a lot of flux?

For instance I want to design an impedance matching transformer (1:3 step-up) for driving a load at ~300khz. I’m having a hard time selecting core materials, I was going to use a toroid because they are readily available but most of them are metal power instead of true ferrite. I know that metal powder “can work” but what properties of it should one look for it to be more ideal then other metal powder.

For reference I have posted images of some metal powder cores provided by Micrometals, For a high current inductor I would pick the -2 material, but for a transformer what would I want the \$\mu_o\$ curve to look like, and what is a good B-H curve for a transformer vs. inductor?

enter image description here

enter image description here

I guess my real question is, does anyone know of a book or good resource for high frequency transformer design that also goes into material property selection?


closed as primarily opinion-based by Leon Heller, Daniel Grillo, Fizz, W5VO Dec 15 '15 at 5:38

Many good questions generate some degree of opinion based on expert experience, but answers to this question will tend to be almost entirely based on opinions, rather than facts, references, or specific expertise. If this question can be reworded to fit the rules in the help center, please edit the question.

  • \$\begingroup\$ Why down-vote this question, it seems acceptable, it is not option based, and shows research? I would gladly modify it if someone thinks it could be improved... \$\endgroup\$ – MadHatter Dec 2 '15 at 19:01
  • \$\begingroup\$ I have been looking into this haphazardly also. The only resources I have found that are useful are videos and manufacturers guides. youtube.com/watch?v=3nfqBzPMknY&index=1&list=WL These guys have a lot of information which may be of use to you. mag-inc.com \$\endgroup\$ – mkeith Dec 2 '15 at 20:13
  • \$\begingroup\$ @mkeith Thanks! I'll take a look, hopefully it is worthwhile new material. \$\endgroup\$ – MadHatter Dec 2 '15 at 20:17
  • \$\begingroup\$ The video starts off with very tedious basics. But stick with it for a few minutes before you judge. \$\endgroup\$ – mkeith Dec 2 '15 at 20:36

I get the concept that ideally you want to transformer to not hold any magnetic energy and instead to transfer it all, but what material properties allow this?

This concept doesn't fit into my way of thinking. For a straight transformer with an AC voltage applied to the primary and a secondary on load, the ampere turns of the secondary (magneto motice force) is totally cancelled by the ampere turns in the primary that resulted from that secondary load. Should you disconnect the secondary load, the primary is just an inductor having an inductance determined by the core material, shape, gaps (if any) and number of turns.

To that end, for a HF transformer, you pick a ferrite material that has low losses at the operating frequency (read the material data sheets for this) and then you begin the process of determining the number of primary turns so as not to cause excessive saturation.

This means trialling an estimate of inductance, calculating the number of turns and therefore calculating the MMF (ampere turns) for the primary under no load conditions. You then factor in the mean magnetic field length (it's a core parameter contained in the data sheet) to calculate H: -

H = ampere turns per metre

Amps is based on inductance, frequency and AC voltage applied to primary just as any inductor would be. Take the peak current and multiply by primary turns and divide by effective core length.

Then go to the core parameters in the data sheet and see if the value of H is going to cause excessive saturation - i.e. use the BH curve.

If it looks like too much saturation then you'll need to increase the turns and possibly implement a gap. Same method as a flyback transformer.

  • \$\begingroup\$ Thanks, I made a spreadsheet that does everything you described! I figured I am doing it right as I have had success in the past but never have I tried a transformer, your summary is probably better then anything I have found on the internet yet. I have been calculating B right off the bat based on f,N turns, Voltage and \$A_e\$. Then H by \$(I * N)/(mean path)\$... And using the manufacturers curve fitting equation to calculate core loss. So I guess I am doing it right... I just felt a bit unsure. \$\endgroup\$ – MadHatter Dec 2 '15 at 20:12
  • \$\begingroup\$ Should I be worried about using something like a type 2 material as shown above for a transformer? Won't the magnetizing inductance be quite huge? Also What do you consider acceptable core losses (mW/cm^3)? \$\endgroup\$ – MadHatter Dec 2 '15 at 20:15
  • \$\begingroup\$ A bigger mag inductance reduces onset of saturation. Think about it like this. If turns double then inductance quadruples. If L quadruples then current (at f) is quartered therefore ampere turns half! For the type 2 material you need to look at complex permeability versus frequency. There should be a graph in the material data sheet. It will show two curves; permeability and derivative of permeability. The derivative is basically eddy current losses versus frequency. \$\endgroup\$ – Andy aka Dec 2 '15 at 21:27

For inductors and flybacks, where you want the magnetic core to store energy, you want low permeability, the -2 material.

For transformers, where you want the most flux for your magnetising ampere turns, you want high permeability, so thoer things being equal, the -8 material would appear to be the better choice.

Obviously you also need to choose based on the losses at the frequency you want to operate at, and these will also be strongly dependent on the flux. You may find that a low permeability low loss material gives you a better power throughput once you have trimmed your flux to meet your core heating requirements.


Not the answer you're looking for? Browse other questions tagged or ask your own question.