Re: Plots of Je

["David Meeker" <dmeeker@xxxxxxxx>]

The time-harmonic problem that FEMM is solving is:

curl((1/mu)curl(A)) = -j*omega*sigma*A + Jsrc

where mu is permeabilityin H/m, omega is frequency in rad/sec, sigma
is conductivity, and A is vector potential in T*m

Jsrc is source current density in MA/m^2--that is, there current that
would be in the winding either at zero frequency or if conductivity is
equal to zero.

When you plot "Jeddy", you are plotting just the -j*omega*sigma*A
term--that is, the induced currents that tend to buck the source currents.

When you plot "Jeddy+Jsrc", you are plotting the complete RHS, or all
contributions to current.

In the case where the total current in a region is specified via a
"circuit" property, the program is automatically picking the Jsrc term
so that the integral of "Jeddy+Jsrc" adds up the the specified value.

To get a prescribed to total current in a conductive region for
frequency not equal to zero, you have to prescribe the total current
via a "circuit" property.

Dave.
--
http://femm.berlios.de/dmeeker

--- In femm@xxxxxxxxxxxxxxx, "Robert Macy" <macy@xxxx> wrote:
> Here is more on my problems plotting Je....
> 
> *IF* the conductor has zero conductivity (as in a wire bundle), then the
> plots of Je and Js+J_eddy all make sense whether the current is
defined by
> Js in the materials section or whether it is defined as a total
current in
> the circuit section.
> 
> *IF* the conductor has nonzero conductivity (as in a solid wire)
*and* the
> current is defined by total current in the circuit section, the
answers all
> make sense with the plot of Js+J_eddy being appropriate. However, Je by
> itself goes way wonky inside the conductor. Outside the conductor where
> there is no source current flowing the plots are still correct. 
Conclusion:
> Only use Js+J_eddy.
> 
> *IF* the conductor has nonzero conductivity (as in a solid wire)
*and* the
> current is attempted to be defined as Jsource in the materials
section, then
> the analysis really shifts the value of the source current and the
results
> are useless. Can't remember about the plots, but the errors make them
> worthless. Conclusion: Always use circuits to define current in a
solid
> wire.
> 
> 
> 
> Did I misuse the femm program without knowing it? Is a Jsource for
> conductive wires just incorrect?
> 
> - Robert -
> 
> 
> 
> ----- Original Message -----
> From: robert Macy
> To: femm@xxxxxxxxxxxxxxx
> Sent: Tuesday, July 08, 2003 8:17 PM
> Subject: [femm] Plots of Je
> 
> 
> I'm having difficulty interpreting the significance of
> plotting eddy currents.
> 
> I'm plotting eddy currents both in the magnetic material
> and in the conductor's area.
> 
> In the magnetic material where there is no circuit current
> if I plot Je or Js+J_eddy, the plots are identical and look
> as expected.
> 
> In the conductor area if I plot Js+J_eddy, the plots look
> as expected. But when I plot Je by itself, I get
> outlandish values.
> 
> If I plot from the magnetic area into the conductor area
> using only Je, I get identical values to earlier plots
> until the line gets to the conductor, suddenly everything
> goes wonky - making no sense.
> 
> Right now the safe thing to do is to simply only use
> Js+J_eddy for all locations, plot to file, then manually
> remove Js in Excel to see what Je looks like.
> 
> Am I doing something wrong here?
> 
> There is no mention of eddy currents in the manual.
> 
> - Robert -