cst floguet
Regards.
I saw the new features described on recent issue of Microwave Journal.
Cheers,
Element7k
Dear Element7k!
Please post the paper here, as I'm also very interested!!
TNX,
Eirp
Hi EIRP,
It isn't a paper but it is very tiny article (about 15-20 lines)in the product update section. At the moment, I can't find the magazine and was hoping that someone with a subscription will be able to post it here.
Cheers,
Element7k
Scanning MJ issues, I cannot find any news about CST MStudio 5. In the News Products section of the 9/2003 issue you can read
---------------------From MJ---------------------------------
This service release of CST Microwave Studio ? (CST MWS) and CST DESIGN STUDIO? (CST DS) includes several new functionalities since the release of version 4.0. The most significant features include: numerous improvements in the calculation of farfields including Ludwig Transformations, nearfield calculations, time domain probes, phase plots and phase centre calculations. Modal waveguide ports can be defined inside the calculation domain. The new general voxel import allows engineers to input or create their own models for SAR calculations in addition to using the commercially available anatomical data set "HUGO." Enhanced post-processing capabilities include new automatic schemes enabling easy, versatile evaluation of parameter studies, field integration over arbitrary faces, and the application of the AR filter to probe signals, implementation of a direct, multi-frontal frequency domain solver with fast frequency sweep, improved periodic boundary conditions in the frequency domain and improved 2D and 3D import options.
---------------------End-------------------------------------
But no mention to the "service release" they are talking about. I think they must be talking about 5.0 because in the 6/2003 issue they announce 4.2 release
---------------------From MJ----------------------------------
The latest release of Microwave Studio ? version 4.2 provides new functionality in its service packs. These functionalities include: numerous improvements in the calculation of farfields, including Ludwig Transformations and Time Domain probes for antenna designers; model waveguide ports that can be defined inside the calculation domain; new general Voxel import that allows engineers to input or create their own models for SAR calculations in addition to using the commercially available Anatomical Data Set "HUGO"; and a new automatic post-processing scheme to enable easy, versatile evaluation of parameter studies.
------------------------End----------------------------------
As usual, their web page says nothing
Seems like they commenting 4.3 version.
OK, let's wait for info from their website
Apologies, I have made a mistake. It is indeed 4.3 that they are talking about.
Element7k
Some info about ver. 5.0 You can find on:
http://www.cst.de/company/events/tra...m_Oct20_da.pdf
It's a registration form, but contains this:
As I understand, the ver. 5.0 features will be demonstrated only in 20 Oct. 2003, Darmstadt.
Best regards,
Kit-the-great
Now, we use CST 4.3, it is very excellent!
October 7th 2003, CST- Computer Simulation Technology, Darmstadt announces breakthrough:
CST MICROWAVE STUDIO? (MWS), the industry?s leading 3D EM time domain code, will be
extended by a frequency domain solver based on tetrahedral meshing.
CST MWS will be the first commercial code to offer the advantages of both orthogonal and tetrahedral
meshing in one 3D EM simulator. Customers will now be able to choose the method (Method on
demand?) and the mesh (Mesh on demand?) best suited to a particular structure.
This extension joins innovations such as PBA (Perfect Boundary Approximation?, 1998), TST? (2001),
and, most recently, MSS (Multilevel Subgridding Scheme?, 2003). Steps of such significance in
simulation technology have ensured CST MWS?s position as market leader in 3D EM time domain
simulation. Though the time domain module will remain CST MWS?s flagship module thanks to its great
all-round performance, the new frequency domain module will be an invaluable extension in certain
areas including periodic phased-array antennas and electrically small structures. Users will also draw
the additional benefit of cross-checking results with different simulation technologies if desired.
************************************************** **********
To me, it seem's they have gone for FE in frequency domain, very similar to HFSS. First they mentioned "multi-frontal Frequency Domain solver with fast frequency sweep" (the method used in ANSOFT HFSS), and now they talk about tetrahedral meshing. Don't you think so?
True. I have expressed that "fear" long ago in our postings - There are at least 3 reasons I see with that move -
First - Showcases where FDTD does not perform well. I know for a fact a number of items not well addressed by MWS and the support folks could not figure out a solution in a reasonable time.
Second - Many occasions folk just need few freq. responses and not a band. Particularly true for initial design phase to converge to a decent first step.
Third - They want to shorten out HFSS in terms of market share. Over here I am pretty confident. What amazes me is that Ansoft fellas are very cool. Every time I have talked to them and raised up the question they were ice cool. Never heard a bad word about competitors, unlike the staff at CST - arogant and....you know. To thi spoint it seems the same discussion Agilent and AWR are having - whatever claims AWR has, the latest research still says ADS has nearly 80% market share and it makes AWR folks a lot nervous. Over here I see the same:)
To me the most relevant is item 1 and 2 - there had been complaints about CST inconsistencies in a number of structures. Plus, their Frequency solver was ....well I refrain from calling it "crap", but certainly not decent.
I am also curious why are they afraid of HFSS brand names - "Solver on demand" - they just invented their slogan "Method on demand", which is by definition a patent nonsense. And not a good taste under the best of circumstances.
Certainly they are playing on the Ansoft folks nerves:) but if I have to count on a freq. domain solver I will certainly (like at present) resort to HFSS. Any argument about that?
One other point that seems to have eluded: Why CST owns such limited customer base? -I mean what it says on the Web site is only 5 "big" names. ANy suggestions? :)
P.S This post is not intended to enter into discussion FDTD vs. FEM - Hell not. We had ample of that. FDTD has a bunch of its own merits - very well documented in many posts throughout our forum.
HFSS is a terrible SW to predict effective spurios frequencies!
In fact, HFSS very often adds many false spurios frequencies;
further, if you try to use fast frequency sweep feature, you can find other false spurios frequencies.
Regards.
someone said CST5.0 will be out Oct,2003.
hi, i visited their booth at europen microwav conf. i was told that it will officially release early next year. as for the most significant features:
- new conformal multilevel sub-gridding scheme. this, together with the well known precise structure representation offered by the PBA and TST enables high accurate simulations with unprecendented efficiency
-distributed computing for multi-port structures, parameter studies and optimisation.
-new broadband spice model extration.
-frequency domain solver with periodic boundary conditions. floguet mode ports and multifrontal solver.
-improve communication between cst mws and agilent cst
cray
hi,why not they try to improve their optimizers.both mws and hf$$'s optimizers seems ugly.in mws i prefer userdefine optimizers. but in HF$$ VBA interface is so terrible.
Hi
CST5.0 will be out Jan,2004. said the salesman.
the big advantage in VER. 5 is the Mesh on demand
along the structure you can change the mesh to make
refinement where it's needed and large cell where it's not needed
in example that I saw of helix antenna in the ver.4.3 was 1.5milion cells
in ver.5 200.000 cells so now you can gain in time or move to big
structure. for now the mesh manually in the next year it will be
automatically
also the Design Studio Second program that workes with Microwave Studio help to save time in big structure try it and then you will see
the benefit of working with both Programs.
RGZ
PLASMA
Another thing in waveguide like magic tee in the 4 ver mesh was in the all volume including the air around the waveguide now in 5 the mesh is only in the waveguide
plasma
Dear plasma,
Thanks for the info, particularly for the first comparison on number of cells on the two versions. However, I believe the second case that you mentioned is not a problem of CST, but the way you define the background material. If you define it as PEC, and your magic tee is defined using, for example, air, then you will not have the mesh outside the magic tee.
Regards
Dear wave-maniac
I thought like you but I checked it and, I saw that if you define the background as PEC, and your magic tee is defined using air, then you have a boundary box and mesh outside the magic tee.and inside this box and the solver even there is no electric field in the PEC it still mash and calculate there checke it with CST.COM so now in ver 5 it's not calculated
and its reduce the mesh and save time.
plasma
Hi Plasma,
Thanks for the info.
I think the sub-grid or mesh on demand along structure is one of those features similar to the AR-filter. It will not work on every structure. Also, I think although the mesh is reduced, there is no reduction in memory usage as behind this mesh engine, it needs the basic mesh around the object under investigation to solve the fields. Maybe only reduction in time but not in memory. In other words, it is still less effective than optics methods to solve large object problems (object >> lamda). I am also wondering if this scheme works with objects with complicated geomentry? It will be very much appreciated if you can checke with CST.com. :)
Overall, I am pleased with all the changes that has been made. Hopefully, there won't be many bugs and I shall await eagerly for either the beta version that is coming out end Nov or the final v5 that is coming out begining of next year.
Cheers,
Element7k