Re: [femm] Direction of Magnetization, Other Flux patterns

[David Meeker <dmeeker@xxxxxxxx>]

davet82000 wrote:

> I am trying to model the process of magnetizing a ferrite magnet. Say
> by running current carrying wires immediately next to its surface.
> This is a simplification but will suffice.
>
> The magnetized ferrite block is then removed from the wires and its
> effect (Force and Torque) on a second magnet is observed.
>
> I cannot see a way to ascribe the resulting field pattern to the now
> magnetized ferrite block. Does anyone know how to set the magnitude
> of the magnetization as a function of position across the block.
> eg less further away from where the wires were. Setting the
> magnetization angle is fairly straight forward and I can calculate
> torques using coenergy methods but cannot see how to set the above or
> a radial field distribution.

The problem here is that you potentially have to do a time-domain 
simulation with hysteresis (a simulation which femm doesn't really do) 
to figure out where each part of the magnet ends up after the 
magnetizing process is over. However, being lazy, I'd just assume that 
the magnetizing currents were strong enough so that the entire ferrite 
block was thoroughly saturated. Then, I might assume that the amplitude 
of the magnetization (Hc) is the same everywhere in the ferrite and just 
the direction is unknown. I'd break the magnet into some number of 
pieces and assign the direction for each piece individually. For some 
guidance on the "right" direction to assign, one could make a 
magnetostatic simulation of the magnetizing fixture at full current, 
representing the ferrite by its virgin magnetization curve. The 
direction of the magnetization should be pretty much along the direction 
of the flux in the full magnetization current simulation.

There are some people who read this mailing list who work at companies 
that make permanent magnets. One of these people may have a better 
strategy than mine for your problem, since this must be the kind of 
thing that they do all the time.

Dave.
--
David Meeker <http://femm.berlios.de/dmeeker>