AlNiCo magnet B/H 1

[Clyde38]

I’ve been making the assumption that the normal and intrinsic curves for AlNiCo magnets are (virtually) the same. I have never seen a second quadrant B/H curve for AlNiCo that has both (unless they are indeed one in the same). Properties for Hc and Hci are the same as listed in the data sheets. Is my assumption correct? And if so, why? When designing a motor in FEA with AlNiCo I’ve described the non-linear curve in the properties rather than just an Hc and relative permeability. Does this sound right? Not much call for AlNiCo in motors these days, coreless low inertia DC motors were always AlNiCo (very thick due to easy Demag).

 

[RyreInc]

The intrinsic and normal curves are pretty close, but never identical--normal curve is B vs. H, while intrinsic is B-H vs. H. They look similar because H is much less than B. Hc and Hci will be nearly identical because Hc is the value of H where B = 0, while Hci is the value of H where B-H = 0.

See attached excel file for normal and intrinsic curves for 3 grades of AlNiCo (data taken from FEA software I use.)

 

[Clyde38]

Hi RyreInc,
Thanks for the information and the spread sheet. I see that you used the Normal curve to construct the Intrinsic curve, and yes they are almost identical (see your modified spread sheet). The Normal curve is all than I have ever seen in data sheets. I assume since you went to the effort to construct the Intrinsic curve you feel that they are different enough to do so.

 

[RyreInc]

I'm not sure whether they're "different enough" to warrant drawing separate curves--since they're so similar the intrinsic curve may be excluded simply as a choice for readability when creating datasheets.

I was mainly trying to emphasize that the normal and intrinsic curves are necessarily different; same goes for Hc vs Hci--Hci will always be bigger than Hc, but maybe only slightly, and in this case nearly identical and perhaps within the margin of error.

As far as FEA goes, using a (normal) curve that follows the nonlinear properties of the material in question will always produce more accurate results. Most other magnet materials have normal curves that are linear (or nearly so), which allows one to specify Hc and µr to generate an accurate curve.

 

[Clyde38]

Thanks again. Most of the FEA work I'm involved with is with magnet types that allow specifying the Hc and µr. For the non-linear material (i.e. AlNiCo) I have scanned the data sheet and used graphic digitizing software to reconstruct the curve. This seems to work for me.