[SI-LIST] Re: E-M SOLVERS - HOW DO THEY WORK?
- From: "Jian X. Zheng" <jian@xxxxxxxxxx>
- To: <SRogers@xxxxxxxxxxxxx>, <si-list@xxxxxxxxxxxxx>
- Date: Mon, 4 Mar 2002 09:06:07 -0800
Hi, Steve:
I would like to give some comments on it.
1. Most general purpose electromagnetic simulators are like you say. They
will divide the areas of interest into small cells. Then, they will solve
the current distribution and field distribution on the small cells
represented linearly on the cells.
2. All field solvers are solving the the differential equations with some
specified boundary conditions. Static solvers are solving the laplace
equations: D*D phi = 0, where D is the (d/dx X + d/dy Y + d/dz Z). Full wave
solvers are solving some equations similar to: D*D phi + k*k * phi = 0,
where k is the wave number. This is the most fundamental equation. However,
the real ones may be much more sophisticated.
3. The fundamental of the full wave solvers are, in fact, the Maxwell's
equations. Finite Difference Time domain method tries to solve the Maxwell's
equations directly. It uses finite difference to repace the differential and
it solves the field in the space of interest. Normally, its solution domain
is the 3D space and there are many unknowns associated with the small cells.
However, some methods do not solve the Maxwell's equations directly. For
example, method of moment tries to convert the Maxwell's equations into
integral equation forms. The unknowns are the current distribution on the
surface of the metallic trace only. Its solution domain is much smaller than
those solving the field distribution. However, a general method of moment
code is much more difficult to implement and validate than a FDTD code.
Normally, method of moment codes are more accurate than FDTD codes because
of the nature of integration. The integration process always tries to reduce
the errors.
4. All the full wave simulators start from the microwave and antenna
technology. When they are applied to microwave circuits or other circuits,
they use the so-called de-embedding schemes to extract circuit parameters.
The simulation starts from some stimulus from the source. For example, if it
is a microwave circuit, the circuit is fed by a uniform transmission line
(TLN). The source is somewhere at the starting point of the TLN. The source
will excite the incident wave, and the incident wave will hit the cirucit or
the discontinuities. Part of the power will be reflected back and part of it
will go through the discontinuities. When we have multiple ports,
eventually, we want to describe the multiple port network using the [S]
matrix which determines the linear relationship between the incident and
reflected waves at each port.
5. In real implementation of full wave solvers, the sources are almost
hidden from the useres. Users normally just need to construct the goemetry
and define the ports on them. Construction of geometry normally require some
specific rules, depending upon the kinds of simulator it is.
I hope that above brief explanation is clear to you. Thanks!
Best regards,
-----------------------------------------------------------------------
Jian-X. Zheng, Ph.D
Zeland Software, Inc., 48890 Milmont Drive, 105D, Fremont, CA 94538, U.S.A.
Tel: 510-623-7162, Fax: 510-623-7135, Web: http://www.zeland.com
---------------------------------------------------------------------
> -----Original Message-----
> From: si-list-bounce@xxxxxxxxxxxxx
> [mailto:si-list-bounce@xxxxxxxxxxxxx]On Behalf Of Steve Rogers
> Sent: Monday, March 04, 2002 6:42 AM
> To: si-list@xxxxxxxxxxxxx
> Subject: [SI-LIST] E-M SOLVERS - HOW DO THEY WORK?
>
>
>
> Would anyone care to explain how electromagnetic solvers get from
> a geometry
> and some voltages at places on the geometry to the fields etc??
>
> I understand that for static field problems you can do the following (I
> think):-
>
> (a) Split the area containing the conductors of interest up into smaller
> areas.
> (b) Assume a linear variation (most of the time) of voltage between each
> smaller area and its neigbour.
> (c) The voltages for all areas is then calculated from a knowledge of
> surrounding voltages in an iterative manner
>
> Clearly some voltages are required to start with (those surrounding the
> problem). I believe this is the basis of the Finite Difference
> Method. Which
> is basically a numerical method to solve the Laplace equation. I think you
> can also cheat and put imaginary boundaries around the problem
> far away from
> any conductors of interest. In this way you can still use the Finite Diff
> Method for problems that dont really have a closed area.
>
> Now, Once the voltages at the nodes are known, the currents about the
> geometry can be found (basically ohms law using the conductivity of the
> material being used). From Here you have V and I, the E and H fields,
> capacitance, inductance etc are trivial.
>
> What about once things are not static any more??? This is where my head
> starts to hurt. Clearly with time varying voltage and current one
> must have
> to take into consideration the magnetic and electric field coupling. Heres
> my thoughts on whats going on .....
>
> Consider a square loop of wire passing current due to applied sinusiodal
> voltage. As before (above) slice up the wire into small elements. Now as I
> see it the problem is one of finding the right amount of current to put on
> each of the bits of wire at the same instant in time such that everything
> all ties up nicely (current in any short wire element is as it should be
> when all coupling is taken into consideration). However the solution needs
> to take into consideration the coupling between each element and all the
> remaining elements (Nasty). I assume that solutions for all the other
> instants in time until 1 cycle of signal later must then be computed. This
> could well be utter rubbish!!!!!!
>
> I have tried reading a few books on this subject but have given up most
> times with the sheer pile of maths without any explanation. Most
> books seems
> to cover the solutions for static fields.
>
> Would anyone care to have a go at an explanation or point me in the
> direction of one.
>
> Regards
>
> PS
>
> My reasons for asking the question....
>
> (1) I have an interest
> (2) I belive that you should have some idea of how the tools you are using
> work in order to reduce the 'egg on face' factor.
>
>
>
>
> Steve Rogers B.Eng(Hons) AMIEE
> RF Design Engineer
> Micromill Electronics Limited
>
> email: srogers@xxxxxxxxxxxxx
> Tel : +44 (0) 2392 366600
> Fax : +44 (0) 2392 366673
> Website: http://www.micromill.com
>
> Micromill Electronics Limited
> Leydene House
> Waterberry Drive
> Waterlooville
> Hampshire
> PO7 7XX
> United Kingdom
>
> Registered No. 1456922 (England). Registered Address: Brook
> Road, Wimborne,
> Dorset, BH21 2BJ.
>
> THIS EMAIL AND ANY ATTACHED FILES ARE CONFIDENTIAL AND MAY BE LEGALLY
> PRIVILEGED
> If you are not the addressee, any disclosure, reproduction, copying,
> distribution, or other dissemination or use of this communication is
> strictly prohibited. If you have received this transmission in
> error please
> notify the sender immediately and then delete this e-mail.
>
> It is the policy of Micromill Electronics Limited that no legally binding
> statements, representations or commitments (collectively 'statements') may
> be made by e-mail. Any such statements must be confirmed either by
> facsimile transmission or by post before they will have legal effect. The
> sender of this e-mail is not authorised to commit the company in
> any way and
> the addressee is hereby formally notified of that fact.
>
>
> ------------------------------------------------------------------
> To unsubscribe from si-list:
> si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
>
> or to administer your membership from a web page, go to:
> //www.freelists.org/webpage/si-list
>
> For help:
> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>
> List archives are viewable at:
> //www.freelists.org/archives/si-list
> or at our remote archives:
> http://groups.yahoo.com/group/si-list/messages
> Old (prior to June 6, 2001) list archives are viewable at:
> http://www.qsl.net/wb6tpu
>
>
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List archives are viewable at:
//www.freelists.org/archives/si-list
or at our remote archives:
http://groups.yahoo.com/group/si-list/messages
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
Other related posts:
