[femm] Re: Strange looking nonlinear curve

["David Meeker" <dmeeker@xxxxxxxxxxxxxxxxx>]

On 28 Sep 99, at 15:44, James Peng wrote:

> The following BH curve looks weir in Femm. I checked with excel and 
> everything looks fine. Even the solution seems ok. Does anyone have similar 
> experience with FEMM? What should I do to get it fixed?
> 
> 0.000000	0.000000
> 0.699870	39.788736
> 0.700011	79.577472
[...]

You have a really abrupt transition in your B-H curve, basically 
going from an incremetal relative permeability of about 14,000 to 1 
in a single step. Femm trys to fit this data with a cubic spline, and 
the resulting fit isn't very good for this particular case.

The cure is to add more points in the neighborhood of the transition-
-for this case, a lot of points probably on the range of H from 30 to 
60 A/m, so that, locally, the transition looks like a smooth curve.

Dave

--
David Meeker, Ph.D.
Senior Engineer
Foster-Miller, Inc.
350 Second Avenue
Waltham, MA 02451-1196
781-684-4070
781-890-3489 (fax)
dmeeker@xxxxxxxxxxxxxxxxx
personal webpage: http://members.aol.com/dcm3c