1
\$\begingroup\$

I am trying to design a transformer with 1 turn at the primary and 10 turns at the secondary. For application purposes, I need a long wire at the primary and a large conductors at the secondary, which makes it difficult to consider a higher number of turn at the secondary. I do not care about losses as the transformer is not use during a significant time. It works at 500 kHz.

So I have a large leakage inductance at the primary and to not loose a lot of voltage due to the voltage divider between the magnetizing inductance and the leakage inductance, I need a large magnetizing inductance, so a high permeability core. It will also reduce the peak current at the primary if I have a high magnetizing inductor value and then it will reduce the effect of the leakage inductance. I need also to have a high density flux saturation Bsat

I am looking for a graph which shows the permeability of different magnetic materials in function of the frequency. Does anyone know a material that have a good permeability at those frequency and a high Bsat? Better than ferrite?

It is said that ferrites is good material for high frequency, i.e. that the permeability is still high for high frequency. But I know that the Bsat of ferrites is very low and the permeability is not very high compared to supermendur.

enter image description here

Have a nice day!

\$\endgroup\$
6
  • 1
    \$\begingroup\$ For a transformer, you need a narrow (low loss) BH curve. Supermendur would be exactly the wrong thing, as would magnesil. What is the frequency you want to operate at? The choices of materials is very limited. There are ways to reduce leakage inductance with multiple bifilar windings, primaries in parallel, secondaries in series if you want a 1:10 transformer. but is a 1:10 transformer the best solution to your real problem? What are you actually trying to do? \$\endgroup\$ – Neil_UK May 18 at 10:08
  • \$\begingroup\$ Just a detail: the permeability of the material (the BxH curve or the hysteresis loop) does not vary considerably with frequency. The point is that as the frequency increases, the penetration depth of magnetic field becomes smaller and smaller and the flux ends up using only the periphery of the conductors, it is the famous skin effect decreasing the effective cross section. \$\endgroup\$ – Luiz Oliveira May 18 at 10:30
  • \$\begingroup\$ Skin effect is related to how the current is distribued into a conductor. I do think that the skin effect is really present on magnetic core as there is no significant current flowing into core compared to the current flowing into the windings, isn't it ? \$\endgroup\$ – Jess May 18 at 14:30
  • \$\begingroup\$ Thank you for your comment :) @Neil_UK, I will take a look on how to reduce leakage inductance ! \$\endgroup\$ – Jess May 18 at 14:43
  • \$\begingroup\$ @Jess... the reason the cores are laminated is to reduce the induction of eddy currents. The induction of these currents is totally related to the penetration depth of magnetic field and, consequently, to the skin effect. \$\endgroup\$ – Luiz Oliveira May 18 at 16:24
2
\$\begingroup\$

There's a good reason for the many different materials in transformer cores. Core loss is one of the many parameters involved in the trade-offs required, and at high frequencies (>500Khz) ferrite would be chosen over the other materials in your chart because core losses would tend to dominate in the other materials you have in your graph.

For a tape-wound material like most of the ones in your graph, the tape is kept as thin as possible to reduce eddy currents in the core. The construction allows for high flux density, but at high frequencies even 1/2-mil tape would have unacceptably high core losses. And as @Neil_UK pointed out, there are significant magnetizing losses due to the large hysteresis - remember that this type of loss occurs each time the core magnetization is reversed, so higher frequency means more magnetizing loss.

enter image description here

Source: Magnetics Inc. Tape Wound Core Catalog

Ferrite has acceptable core losses as seen below, but at a price: the flux density at which you can operate will be lower. Here is a curve of Magnetics "L" Ferrite:

enter image description here

Source: Magnetics Ferrite Cores Catalog

So you will be able to operate at high frequency. But this is only one of the many considerations you must make when choosing an approach. It is unusual to choose the frequency first in a power transformer design, but if you have to run at the high end, you are on the right track with ferrite.

Good luck!

\$\endgroup\$
3
  • \$\begingroup\$ Thank you for your comment :) Really interesting. I will edit my post for being more precise ! \$\endgroup\$ – Jess May 18 at 14:42
  • \$\begingroup\$ If I consider what you say, it could be interesting to use amorphous transformer ? \$\endgroup\$ – Jess May 18 at 15:10
  • \$\begingroup\$ I don't have any experience with amorphous core transformers - maybe someone else could comment? \$\endgroup\$ – John Birckhead May 18 at 18:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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