Residual air-gap in ER cores of ferroxcube
Actually i design a pulse transformer with an ER9.5 cores of Ferroxcube, the problem is how much the residual air-gap i have to take into account in the design, because it significantly affects on magnetizing inductance.
Anybody knows about this value in reality? i think it is less than 10 μm.
Thanks,
You can assume that the Al value specified in the datasheet will be achieved with correct processing of the cores. The related effective airgap will be mostly created by surface roughness, I think. It's surely below 10 μm.
P.S.: Some information in a Siemens (Epcos) application note.
hello,
If (total magnetic path length)/(air gap) >> relative permeability, then you can ignore the effect of the airgap. Note that you have the airgap twice in the magnetic path, so using the formula you need to use twice the single gap width.
Example: gap per leg = 2um, total magnetic path length = 15mm, relative permeability = 1000.
total magnetic length / total gap length = 15m/4u = 3750. This is almost 4 times the permeability, so actual permeability will be about 100%*3750/(1000+3750) = 80% of the material permeability. Note that the Al values are given for a core set, including the reduction due to non-perfect surface mating. Core sets with high permeability have better finishing to reduce the effect of residual air gap.
Thanks for your useful informations
Dear thevan48
Residual air gap for ferrite cores depends on processing of the cores.
For polished ones it can be as low as 0.6μm. Polishing is used only for high permeability materials.
For your ER9.5 total residual air gap is approximately 2.5 μm.
Regards
Ferroxcube Application Engineer
ER9.5 is actually available in a "high permeability" material quality, Ferroxcube 3E5 and 3E6 respectively Epcos T38. Specified effective permeability μe is only about half the actual material μi, which is apparently mostly caused by the residual air gap and to a small extent by the geometry.
Furthermore the Ferroxcube specification is assuming a clamping force of 10 N (quite strong). The most important point to achieve the specified AL numbers is however clean processing.