Re: [femm] Re: Current

[David Meeker <dmeeker@xxxxxxxx>]

Keith Gregory wrote:

> At 20:29 22/07/2002 -0400, you wrote:
>
> > brian_hawes wrote:
> >
> > > Keith,
> > > Thanks for the reply. My problem is thus: I have a simple simulation
> > > of a single curent carrying ring. I have defined the current density
> > > in the model for the copper. If I simulate at very low frequency,
> > > such that skin and proximity effects are negligable, then the total
> > > current, found by using the block integral feature in FemView, is
> > > what I would expect - J x Area. Now, if I increase the frequency such
> > > that skin and proximity effects occurr, then the total current,
> > > measured with FemView appears to fall. I would expect, as you say,
> > > that the total current would remain constant, and just be
> > > redistributed.
> > > Brian
> >
> > If you define a "source current density," this is defines the current 
> > that you would get at 0 Hz. At nonzero frequencies, in conductive 
> > regions, some of the current gets bucked by inductance.
>
>
> I think I had begun to work this out looking at a similar situation to 
> that Brian has been using. The current density seems to act in a 
> manner similar to what would occur if a constant voltage was applied 
> to the region rather than a constant current. Am I on the right track 
> or is it not this simple?
>
> Keith.

Yes, that's right--the current density that you assign is a "source 
current density" which would be the same as conductivity*voltage 
gradient (at least for regions with a nonzero conductivity...) To force 
a particular net current on a conductive region regardless of frequency, 
you have to use the "circuit" properties to do it.

Dave.