EM Simulation packages list

I see here a lot of questions about how is better for what. anyway, here is a list of EM simulators, check them and we may then have a better discussion about'em.

http://directory.google.com/Top/Scie...tism/Software/

urs, murti

Here is a paper, benchmarking some simulation software packages

They aren't listed all but appreciate good idea.

OK, I have a question if someone have experience with comparing results of various p a t c h antennas (better say compact) given by IE3D and CST-MWS.

Usually I'm getting frequency shift between those two.
Note that I'm trying to set similar conditions when possible - like (in)finite ground plane.

I also have to say that the measurement is close to IE3D's results when having typical "2.5D" structures.

In MWS - mesh settings, distance of PML, accuracy (energy drop) don't have significant influence to the fq. shift..

Any opinion/experience?

Cheers

eirp

Hi Eirp,

why don't you post some of your typical designs here and we all have a look. I do not have access to IE3D but maybe I can give my coment to your MWS simulations.

F.

Hi, RFsimulator!

Okey, I'll post smt later.
But now I think I found the origin of problem.

Imagine classical patch antenna.

- When using infinite groung in MWS (zmin=PEC) and infinite ground in IE3D, there's frequency shift

- When using finite ground in MWS and inifinite ground in IE3D, the results are in very good agreement!!

- finite ground in MWS and IE3D is of course also OK

So from my experiencies now, the problem is related to behavior differences in using infinite grounds in MWS and IE3D

Dear RFsims :),
I have simulated a set of Multisegment Rectangular Dielectric Antenna (MSDRA) with Fidelity and HFSS (agilent, 4.3 i think), and have compared the results with each other and with measurements. the amazing thing that always FEM(HFSS) had a left frequency shift, and FDTD(fidelity) had a right frequency shift! so if the structure simulate with both codes, one can surely find the right frequncies (average res. freq.). Now, eirp, please see if this is the situation for infinite ground plane cases in MWS and IE3D, or not. if you can, please try a formerely measured sample.

murti.

Hi, Mamali, RF and others..

I didn't perform measurement yet but here's the comparison MWS x IE3D.

In IE3D both cases are with infinite ground and then:

tslot_infinite - infinite ground in MWS
tslot_finite - finite ground in MWS

Structure is rectangular patch with two T slots, on 1.5mm air substrate, probe feed..

Your comments please? :)

eirp

Dear eirp and others,
I sow responds, due to thin substrate you used, the IE3D results seems to be more preciser, and looking to finite and infinite results, it seems IE3D is more robust to ground effects (they have embeded a built in surface wave calculator in it). thus, it seems we could rely heavily on IE3D results, and CST MWS should be used as a checker, and must used with finite ground planes. I have some other experiences with IE3D and other simulation methods and package, and I think It is the case for all of 'em. so, VIVA Zeland :).

your friend, murti

Hi eirp

try to extrude the region of the port only for few mm
and then calculate the s11. tell if it works.

pl

Hi, plasma!

Of course, I have the port extension as small as possible..
The problem is there still.
I have also note, that my project uses air dielectric, so possible truncation of substrate is also out of game.
TNX

eirp

Hi Eirp,

could you just post your MWS project here?

Thanks,

F.

Hi all,

I have encountered the same problem with CST MWS but for my application most results were within measurement tolerance . I think it has something to do with how the port is actually modelled. IE3D uses a probe that is optimised to a patch antenna but MWS uses a waveguide/discrete port so it depends on how accurately the "bits" of the port/antenna is modelled. A reactance check shows a difference between IE3D and MWS and therefore the frequency shift. Changing port dimensions might help (i hope). I have noticed this happening in Ansoft Ensemble as well and reactance are consistently slightly lower than measurments but within tolerance. So I think same problem here.

Do you all think this is a valid reason?

Cheers,
Element7k

what is 2.5D mean?

2.5D states for metalic/dielectric layers (2D) in XY and vias ("0.5D") in Z through

why use the term 0.5D.
and not use 3D.
and this mean it is diference from 3D?But what deference?

Hi,

considering layered strusture simulators like Sonnet, Momentum, Ensemble, IE3D etc. (layers are paralell to x-y plane, z plane is perpendicular to interfaces), I think the terminology will be like this:

2D-you can have layered media and metal traces on interfaces only. Not any z orieted metal element can be introduced

2.5D-you can have some z oriented elements (e.g. vias) in your structure. 0.5 comes from the fact that some (early) simulators would only ensure the current continuity, not any field produced by z oriented current. Some simulators will account for field produced by z oriented current but not for the fields produced by circular current (flowing around via walls). There is an approximative way out for the later as you can divide the single layer with via to more layers so the circular current will be present at the interfaces and fields will be accounted for.

Finally, 0.5 stands as a reminder that structure cannot be fully 3D (metal sphere, as the most obvious example) as the methods for 2.5 simulators are based on deriving Greens functions for layered media, so the media should consists of layers, cannot be completely arbitrary.

I hope IICCEE that this can answer your question.

Regards

flyhigh