[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [TCML] Maxwell 3D Electrostatic Simulation of a Tesla Coil



Mike,

Generally 3-D modeling is very hardware intensive, often memory is too small
to handle full system problems with any reasonable amount of nodes in the
mesh.  Because of this, problems are often reduced to some fraction of the
entire system (like modeling one slice of pie, rather than the whole thing,
because its symmetric).  You can likely do with a 2D axis-symmetric
simulation instead where now your slice became infinitely narrow.  If you
want to model something like the breakout point, or any other non-symmetric
thing, you could probably still break the problem down into 1/4 or 1/2 of
the thing to save some memory. Defining boundary conditions on the problem
becomes more complex in this case, you have to fake the solver into
believing the rest of the "pie" is really there, even though you are only
looking at one "slice". So its sort of like "model this pie, given
information about this slice, and assuming the rest of the pie is just like
this slice". The basic idea behind boundary conditions is to tell the
software if the electric or magnetic fields should be tangential or normal
to the surface at the boundary.  We can discuss this further if you like,
though i bet you already know how this works even if you havent formally
learned it yet.  Improper boundary definitions can make your solution total
junk, so you gotta get this part right.  Usually the field plots are
obviously wrong in these cases when compared to what you expected the fields
to do.

Specifying the currents or voltages is really another boundary condition.
Generally, though, most of these solvers can only handle either electric
fields or magnetic fields as separate problems, and its really up to you to
decide how to go about solving each side of the problem to get the
information you seek.  If you are completely new to this, id suggest doing
an electrostatic simulation first, as its often easier to set up the problem
and make it work.  The reason is that in electrostatics, all conductors can
be reduced from a volume to just a surface, that is, there is no charge
*inside* the conductors.  This reduces the number of nodes tremendously, and
its fairly obvious where to assign the voltages.  Getting the secondary
correct is harder since maxwell wont really know how the voltage is
distributed, you would either have to just assume some linear voltage per
length along the coil (and possibly have to define it as 1000 segments or
something), or a trick i used was to assign the top of the coil with some
voltage, and the bottom of the coil as ground, and fake the problem by
letting the secondary be a high permittivity resistor, which will force a
linear voltage distribution along its surface.

Anyway, its probably going to be a good learning experience (knowing how to
use such software is valuable on its own) as i expect you will run into
trouble and probably find ways around it.  But in a practical sense, its
probably not very useful because tesla coils are really nicely symmetric
most of the time, so 2D modeling is much easier and should be able to give
just as useful information with less work.

Steve

On Wed, Mar 2, 2011 at 4:50 PM, Michael Carosino <m.carosino@xxxxxxxxx>wrote:

> Hi, I am working to do a simple project that involves the simulation of a
> tesla coil with Maxwell 3D. I have little experience with the program and
> so
> I have just begun making some coils and other pieces to see how it works. I
> am curious if anyone has attempted to model a Tesla Coil in Maxwell and if
> it is indeed possible/reasonable? I am a bit curious as well
> about specifying relevant voltages and currents on the coil... I'd
> appreciate some ideas if anyone is familiar with this idea.
>
> Best regards,
>
> Michael
> _______________________________________________
> Tesla mailing list
> Tesla@xxxxxxxxxx
> http://www.pupman.com/mailman/listinfo/tesla
>
_______________________________________________
Tesla mailing list
Tesla@xxxxxxxxxx
http://www.pupman.com/mailman/listinfo/tesla