Re: [femm] Voice Coil Motor with Aluminum bobbin

[David Meeker <dmeeker@xxxxxxxx>]

tano1938 wrote:
>   I'm
> trying to model a voice coil motor such that I can check the 
> performance at different frequencies.  The voice coil is wound on a 
> thin aluminum bobbin for support which I suspect will effect the 
> performance at high frequency due to eddy currents in the aluminum.  
> The coil is modeled as a current sheet with zero conductivity. This 
> is the first time I've used a frequency other than zero with this 
> design and I'm not too sure how femm operates in this mode.  The 
> only "field" in this model is the current in the coil, so I'm 
> assuming that this current is oscillating when I set a frequency 
> other than zero.   The results I get even when running very low 
> frequencies are not possible. 
>
> The file in the files section is "Eddy_VC.FEM"
>
> Thanks for your help
I'd guess that the node at (0.01,-0.325) should really be at
(0,-0.325), but other than than, the solution looks like I'd expect.
You'll also need to specify a nonzero conductivity for the 1018 steel
to get the eddy currents induced in those parts.

The result is just the AC part of the field due to the coil. Remember
that the program isn't solving the combined AC + DC offset
problem--it can only solve one part at a time.  Additionally, the AC
problem considers only major hysteresis loops (i.e. no DC offset).  I
occasionally threaten to write a combined AC + DC offset program,
because this scenario keeps popping up in the context of a lot of
practical problems, but I haven't gotten around to it yet.  

Anyhow, since your DC problem has significant nonlinearity, you'll have
to be careful how you set up the AC problem--you can't just superimpose
the results in the way that one can for linear problems. For AC
analyses, you'd want to replace the alnico with a linear material where
the permeability of the magnet is chosen to be representative of the
incremental permeability at the nominal operating point.  You may have
to do a similar trick with the 1018 in the areas with high fields in
the DC problem.

Dave.
-- 
David Meeker
dmeeker@xxxxxxxx
http://femm.berlios.de/dmeeker