Identifying ferrite materials

This Jan 2012 article has been copied from my VK1OD.net web site which is no longer online. It is for reference from other related articles. The article may contain links to articles on that site and which are no longer available.

One often wants to identify the type of material used in a ferrite core for use at radio frequencies. This article captures advice that the author has offered in online fora stretching back more than a decade, yet it seems uncommon knowledge.

The most common method is to make some measurements to determine the initial permeability µi, usually at audio frequencies, and to compare that to a table of µi for common core materials. This method might well indicate several mixes that have similar µi, but each may be quite different at higher frequencies.

The suitability for use at RF usually depends much more on complex permeability at radio frequencies than it does on µi at say 10kHz.

cf01Above is a plot from the Fair-rite data book showing the complex permeability characteristic of #43 ferrite material.

In fact, the data presented in is measured on a toroidal core with a small winding, the datasheet states that in this case it was [m]easured on a 17/10/6mm toroid using the HP 4284A and the HP 4291A, and probably using just one turn.

This characteristic means that a small inductor wound on a core of #43 material so that there is very little flux leakage will have a complex impedance with X/R=µ'/µ”. The frequency at which R=X is a good ‘signature' for the material type, #43 crosses over at about 14MHz in the above diagram.

It is vital that this measurement is done with as few turns as necessary to give a reliable and accurate reading on the measuring instrument. More turns means increased stray capacitance and the impedance will not longer obey the simple model X/R=µ'/µ” due to the capacitance. On the other hand, measuring and impedance of the order of ohms with the ubiquitous MFJ-259B does not give accurate results.

If the inductor has appreciable flux leakage (eg a rod), then the flux leakage results in a departure from X/R=µ'/µ”.

(The technique is not applicable to powdered iron materials as they usually have much lower loss than ferrites, and grading by µi remains the best option.)

Note that ferrite materials are subject to manufacturing tolerances and variation with temperature, so do not expect 1% accuracy in applying datasheets to real cores.

Table 1: Selected Fair-rite data
Mix Frequency where R=X (MHz) µi
31 3.6 1500
33 4.5 600
43 14 800
52 30 250
61 43 125
73 2.3 2500
77 1.7 2000

Table 1 shows the cross over frequency and µi for some common Fair-rite materials. It can be seen that although #33 and #43 have similar µi, their RF performance is quite different, due in part to the fact that #33 is MnZn ferrite and #43 is NiZn ferrite. If one was to classify an unknown core of type #43 or #33 based on µi alone, allowing for manufacturing tolerances, temperature and measurement error, it would be very easy to wrongly classify it.

Likewise, other manufacturers may have cores of fairly similar materials, and measuring µi alone gives no indication of the RF performance, or a valid comparison with a known core.

Links

Changes

Version Date Description
1.01 29/05/2012 Initial.
1.02 09/02/2016 Copied from VK1OD.net.
1.03
1.04
1.05