[gme-users] Advice on electronic circuit modeling (was Help)

[Jonathan Sprinkle]

To add some extra wisdom to Matt's response, we turn to Jonathan Sprinkle,
GME User, and known consorter of inventors of SPICE...

Jimmy,

When you are building any models in GME, they conform to a defined modeling
environment (in GME this is frequently referred to as a paradigm). This
provides you with the structure and components (e.g., a voltage source,
which can connect to different kinds of components, such as other sources,
or sinks like resistors, inductors, etc., etc., etc.).

The definition of the physical laws that govern these elements is what is
called in GME-speak an "interpreter"; it interprets the components you
built, and attaches physical laws based on (1) the *kind* of the component,
and (2) the way in which they are connected together. GME works best when
configuring an existing environment (in your case, an electrical ckt
simulator) to execute based on the structure in GME. However, it is possible
for you to also build your own simulator, and user your GME models as the
basis for simulation....

Anyway, in order to build electrical component models, you would either use
an existing paradigm for electrical components, or design your own (this is
the suggestion that Matt made, although I am pretty sure that a Bond Graphs
modeling environment already exists, and is used by Gautam Biswas, a
professor affiliated with ISIS: http://www.vuse.vanderbilt.edu/~biswas/)

I encourage you to further define what you mean by 'dynamic', since this can
have many different connotations. Presumably, you want to do something other
than what SPICE can give you, or you would not be investing time looking for
other environments. ;)

Not to encourage too much competition (grin), but if you want to experiment
with quickly built models and simulate their output, you could do worse than
the Ptolemy tool from Berkeley (http://ptolemy.eecs.berkeley.edu/) which is
more like MATLAB's Simulink, in that its simulation environment is built
into the visual tool. However, types like "circuit" and "resistor" are not
immediately available, so you would be reduced to defining them in terms of
their physical phenomena, and getting that out of the simulation.

If you joined these two (e.g., built an environment in GME that had
resistors and circuits...) and an interpreter that generated XML-files which
would configure the physical definitions to be simulated in Ptolemy, it
might give you the kind of behavior that I *think* you are looking for. This
would enable you to experiment with the layout of your circuit in GME, and
at the touch of a button, generate (from GME) and simulate (in Ptolemy) that
model...

Good luck, and read through the GME user's manual (now that you have this
example to look at) to get a flavor for the best uses of the GME
architecture.

Best,
Jonathan

> Could anyone guide me about configuring and building dynamic "electrical"
> models in GME? For e.g. What do I do if I need to put together a simple
> electrical circuit consisting of a Voltage source and a resistor in GME?
> Hiw do
> I define the physical laws that govern these electrical elements?
> 
> Hoping for some vital clues from you.
>
____________________________________
____________________________________

> If I were you, I would develop a modeling environment for bond graphs.
> If you aren't familiar with bond graphs, they are a graphical formalism
> for describing the dynamics of physical systems. Given a bond graph,
> there is a well-defined way to derive a corresponding system of
> differential equations. There's plenty of information about bond graphs
> on the web.
> 
> As to how to develop this environment and model interpreter in GME, see
> the user's guide and tutorial.