FDTD software: useful or not ?

[Kit-the-great]

In this forum I'm found the many contradictory expressions about the softwares, which uses the FDTD method to 3D EM simulation and most of the expressions has been not in favour of this progs. But anybody have the good results in EM simulation with this software (Empire, IE3D, XFDTD etc.) or not ? What problems type the FDTD-software can solve efficiently ? Or this softwares absolutely not useful ? Maybe the FDTD method have many self limitation to simulate the real structures ? Share Your experience and opinions.

Welcome to discuss !

Kit-the-great

 

[cheng]

FDTD

This is because we put emotions over it I guess
After all FDTD is a great and damn simple algorithm and many excellent devices have been designed with it. Great goody. Like others, there are some restrictions but everything depends on one'st applications and pocket. Last but not least - the legacy software.
Before turning to Ansoft HFSS we had 22 designers working with Agilent and was not an easy decision for one to make to switch over night. New designers may take it easier, but old fellas get sticked to their early gadgets much stronger and is a helluvu job to port to a new tool and resimulate again.

cheers

 

[loucy]

Why not try it on some problems and find it out by yourself.

If the method is not useful, why you see so many people are developing it?

 

[personal]

fdtd

fdtd is good for open-boundary problem and less accurate for closed boundary problem...

 

[mesfet]

I think both MoM and FDTD have their strong and weak aspects.

For MoM, it is fast and accurate for small structure. We can even optimize the structure which is not feasible in FDTD.

For FDTD, it can be applied to the problem of the scattering response of plane or other large objects. It is believe that MoM is not applicable to handle large objects because the enormous amount of memory required for manupluting the matrix.

For me, as a circuit designer, momentum and hfss can satisfy all my job.

 

[yuyu]

FDTD is very efficient for radiation problem but is good also for close problem; of course mode matching is very good for particural problems as waveguide transition but is not a general purpose method. In that case mode matching allows very fast optimization.

Further FDTD commercial softwares are very different; one of most important thing is the conformal feature to solve small details. Often, it is the difference.

Regards.

 

[goxy]

But anybody have the good results in EM simulation with this software (Empire, IE3D, XFDTD etc.) or not ?

Zeland's IE3D is MOM tool not FDTD. Fidelity by Zeland is FDFD and I can believe that same company made both programs.
IE3D is best EM simulator for me, but Fidelity is @$@#$%!&.
With FDTD tool you can't made plates and wired structures without several GB of RAM mem. and for solver you will wait several days.
If you want sim. simple stuctures like caps or resistor with FDTD you can do that.

 

[midnite]

I believe that FDTD method is a powerful techinique to solve 3D EM problems. It is a method in time domain that allows for a frequency band investigation with only one simulation ( in frequency domain method you need a simulation for each frequency point); moreover it allows for transient analysis. Thanks to variable mesh tool you are able to carefully describe details of the structure without a huge need of memory.
It is very powerful in:

- arbitrary geometry structures
- radiation problems
- multilayers structures
- closed structures that involve discontinuities

Main disadvantages:

- long simulation time for resonant structures
- the basic algorithm do not take in to account for losses (especially in thin layers)

 

[pewang]

fdtd have a good algorithm.it derived from Maxwell differential equitions and no need for matrix inverse operation.but it mesh divide is hard to set up than FEM I think.since the later is a more general method althoug its low efficiency for radiate problem.

 

[yuyu]

Hi Pewang,

you wrote "... but it mesh divide is hard to set up than FEM". That is true for classical FDTD but now are available several codes with CFDTD, Conformal version of FDTD.

Regards.

 

[pewang]

Hi,yuyu
Can you suggest me some useful citation
about CFDTD.I have bought one book on computation electromagnetic wrriten in chinese,there are some chapters on CFDTD,but I haven't undestand them all.
Thanks!

 

[yuyu]

Hi Pewang,

you can find an useful forum on Elektroda named "IE3D 10 and Fidelity 4" with several references.

Regards.

 

[mwpro]

FDTD is so good, why doesn't the EM software giant like @nsoft release a fdtd software? They got everything. Just curious.

 

[loucy]

MWPro,

To answer your question in a way, software released by "Giant" Ansoft is not necessarily good (compare Ensemble with IE3D or FEKO). They have to think about whether their FDTD product (if they have one) can compete with others.

 

[microstrip]

i found this forum recently! I like it at once !

i think the fdtd software will develop very soon! Now ,have some problem ! But the trends will not change!

Say hello to every one!

 

[HxJi]

eng,some time

 

[plasmon]

lecture

Is there some one can gice a systemic lecture for this problem?

 

[Sigtest]

HI,

I'm also learning ftdt recently. I found a headache to me is fdtd requires huge memory if I simply use uniform grid. There are some approaches to help(sub-grid, local cell...) but I have no idea which one is better. Any one can give me a hint ?

 

[plasmon]

I also want to know these answers. thanks

 

[romanpan]

I think fdtd is very useful!

I think fdtd is very useful!