3d field solver

In general, It is well known ansofthfss and cstmws are two great 3D sover. cstmws is more easy to build model.

There is no good solver as it is highly dependent on the type of problem you are solving.....


For example, the use of HFSS in analyzing CPW structure may not be very fruitful as it is very time-consuming......


Perhaps, em guru may like to add a few words on this?

Hi:

Actually, the kernel of the solver depends on
the Partial Differential Equation (PDE) solver.

As you all know, Maxwell Equation is the PDE
to be solved in all these EM simulator.
To compare the EM solver is basically another
way to compare PDE solver.
As far as I know, there are several ways to
solve PDE

1) Finite Difference Time Domain (FDTD)
2) Finite Element Method (FEM)
3) Method of Moment (MOM)
4) Finite Integration Method (I donot know much
about this one, which is used by CST MWS)

You may add more, please, if I miss some.
Well, you can also use parallel computing, which I
am interested in but have no chance to try yet.

Every PDE solver has is plus and minus.
My point is that unless the PDE solver has a big
progress, the EM solver will stay steady for a while
and you can NOT see a super-fast, super-accurate,
EM solver happening.

Regards
Div

Finite Integration Method is very similar to FDTD; in practise, Maxwell equations are solved in integral form instead of differential form and that is more suitable to develop conformal schemes.

Regards.

OOPS. Sorry for blank post. I have demo ed EMPIRE and SEMCAD which are FDTD software. EMPIRE hase a very laborious GU interface but is one of the lower cost softwares. SEMCAD seemed to have a simpler drawing command structure than MWS but it was slower in analysis (much slower). I prefer not having to be an expert in the theory of FDTD to get good engineering analysis fast.

I've tested EMPIRE 3.0. It's absolutely unfriendly for me, after one day of work I didn't built any of examples...
One advantage is that it can display electric currents.
Do you have any better experiences with Emire?
Regards,
Eirp

I've been learning hfSS lately. Its hard to enter the geometry but once that's done, it works well. Convergence hasn't been a problem. I've just done stripline stuff so far but the view of the E/H fields allowed me to see where the imbalance was in the structure and to correct it. Very nice indeed.

I always use @gLI@ent H1F55 in my design. I think it is the bett one.

hi

I am using ansoft HFSS and XFDTD, I think for a frequency simulation, HFSS is better, but for a sweep application, I normally choose XFDTD.

regards

I think Feko is nice....

Generally, the future trend wil be towards Time Domain - this is what we are modelling in real life - you can also use the results from time domain to calculate parameters required in time domain reflectometry etc.

A major weakness of Ansoft for example is its poor scalability - bigger model (n*lambda), non-linear requirement of memeory - that's why ansoft were desperate for 64 bit. Can HFSS (easily) simulate antenna systems of the 10 lambda range? This is important - modelling and simulating small structures is possible with all the simulators but the trend is towards larger models/systems.

Frequency domain methods are generally not ideal for HF/MW because you cannot always know in advance where resonances are going to occur.
Even fast frequency sweep methods are inefficient. TD allows an accurate range of S-Parameter in one simulation.

Of course, there are other aspects that should be considered - materials, post-processing etc.

What about the Zeland FDTD program "Fidelity"...does anyone have any experience with it? Is it user friendly?

Regards,
Joe

It looks like Zeland hasn't put much effort on improving Fidelity.

XFDTD hasn't upgraded for quite some time.

about empire:

any comments?
because i had similar feeling to the one eirp had :>
(2 hours and i gave up)

Rgz,

Excuse me if I'll write my opinion on Fidelity - it's a piece of sh.. !
I've worked with it time ago and results were wrong in most cases. And don't let me talk about stability and GUI...
Don't use it, save your time and resources!
Regards,
Eirp

[quote="loucy"]It looks like Zeland hasn't put much effort on improving Fidelity.

Wrong,they improoved Fidelity a lot and as of ver.4.0 it will have conformal as well.
Some of the students that use the 3.5 version say it is very fast and easy to learn and use.
Don't forget that Zeland Software became also object oriented like MWO, CST, and Ansoft Designer(when it will be available?).

Hi there:

In stead of talking about indivitual tools,
how about give a ranking for the following list:

IE3D & Fidelity
Ans0ft HFSS
Micr0wave office
ADS momentum
Empire
CST microwave studio
XFDTD
FEKO

The list doesnot include all the solvers, and
I am NOT ranking it.
You could add more if you want.

Regards
Div

Hi Div,
first of all Microwave Office is a circuital software. This CAD has an e.m. option, with extra payment but is a planar solver.
Also Momentum is a planar solver.
IE3D is not a true 3D e.m. simulator.
Empire is not an expensive FDTD simulator but is not friendly and it is not accurate.
HFSS is good generally but not for radiation problem (memory resources).
FEKo is a moment based code, very fast but not for general purpose application.
I don't know XFDTD.
Microwave Studio is very good for general purpose and especially for radiation problems.

Regards

I think MWS is a better choice for 3D solution.