Help needed for CST MWS eigenmode solver!!

[keating]

solver run aborted cst

Is there any magic step when using CST MWS eigenmode solver??
My simulation is just to find eigenmode frequencies of a tungsten woodpile EBG structure. I ran it on a 12G ram computer and just couldn't get it through! Always error info comes out like this :"Maximum iteration number reached. Simulation aborted."

I followed every possible step I think it should take:

1. Set units
2. Set background material (normal -> air)
3. Define structure (set up my unit cell and use periodic boundary conditions on x, y and z directions)
4. Set frequency range (I first used 200 - 650 GHz, doesnt work, then I tried fmin=2GHz and only calculate 3 modes, still aborted)
5. Set boundary conditions
6. Start eigenmode solver

Help needed! Am I doing everything right?

Added after 2 minutes:

And BTW I choosed JKM algorithum already, which according to CST Help should be able to deal with lossy materials including Tungsten.

Added after 14 minutes:

Memory should not be a problem since I can run the same model through on HFSS and HFSS is usually more memory consuming than CST.

 

[irfan1]

Could you please post your projet file? I have done quite a lot of eigenmode calculations for photonic crystals by using CST and also RSOFT Bandsolve.

 

[itzikhaim]

Hello keating !

You need to understand that the EM solver is usefull for resonant structures !
So you'll need to have resonance - it will show you the fields of the resonance and their frequency as well.

I saw your post and I noticed that you set the background material to Air.
Its wrong because resonator device (which are most suitable to the EM solver [EigenMode Solver] ) need to be filled with air but need to be in a PEC surrounding(and to have Electric Boundary).

Check this - good luck !

 

[irfan1]

Eigenmode solver can easily solve the modes of a periodic structure. So, I would not agree with your comment. As a matter of fact, I myself used CST eigenmode solver to calculate the band structure of a 2D photonic crystal. And the results agree quite well with the results of well known band solver BANDSOLVE.

 

[keating]

Could you please post your projet file? I have done quite a lot of eigenmode calculations for photonic crystals by using CST and also RSOFT Bandsolve.

Here is my design model. Thank you very much for help Irfan1!

Actually I have Bandsolve in hand but have trouble with it. Its manual is really dummy and not helpful so I still dont know how to setup metallic materials with it. Besides it doesnt give its basis vectors so if I want to sweep k-points in 1st BZ I would have difficulty setting up my k-point list. Any advice would be welcome!

 

[irfan1]

I would suggest using AKS method. I would not expect that considering lossy metals as PEC would be much of a consequence when the frequency is so low. Please check my attached calculation.

 

[keating]

I would suggest using AKS method. I would not expect that considering lossy metals as PEC would be much of a consequence when the frequency is so low. Please check my attached calculation.

I checked your attachment... Are you saying at frequency of several hundred GHz the JDM solver simply doesnt work for my model? Is it because of my model or CST MWS?

If you use AKS then CST would simply ignore the loss right? That would make no sense to me b/c what I want indeed is tungsten WPS band diagram from 200GHz to 600GHz.

Irfan can you come accross any other way to do my problem? Thank you for the time and patience anyway~

 

[itzikhaim]

Dear Keating !

I ran your model on my comuter (dual core 2.8GHz with 2GB of RAM) and
I used the JDM method of the EM solver and It ran about half a minute !!!
The CST MWS calculated the frequwncy of the eigenmode to be:
olver Results:

--------------------------------------------------------------------------------
Mode Frequency | Accuracy
| |(Ax-x)/x| max(e) div(e)
--------------------------------------------------------------------------------
1 237.436437741 | 3.90e-011 8.50e-004 1.32e-008
2 323.457360136 | 5.38e-010 6.26e-004 1.63e-008
3 326.028714476 | 3.09e-006 2.90e+000 1.62e-008
4 364.112337903 | 6.29e-008 2.23e-002 1.94e-008
5 364.112429269 | 1.30e-009 3.47e-003 1.78e-008
6 * 453.387883466 | 2.22e-003 3.80e-001 3.23e-004
7 453.404743042 | 1.05e-005 5.91e-002 1.72e-008
8 * 472.348320887 | 1.34e-003 5.60e-001 1.59e-006
9 * 492.629186611 | 2.29e-003 8.82e-001 1.78e-007
10 * 515.215939147 | 3.05e-002 1.26e+000 4.78e-004
--------------------------------------------------------------------------------

* = This mode has a bad accuracy.

Optimum guess for the highest eigenfrequency would be: 515.216.

--------------------------------------------------------------------------------

--------------------------------------------------------------------------------
Solver Statistics:

----------------------------------------------------------------------------
Peak memory used (kB) Free physical memory (kB)
Physical Virtual At begin Minimum
----------------------------------------------------------------------------
Matrices calc. 37068 32228 1363052 1356360
Solver run total 13168 10364 1387856 1356424
----------------------------------------------------------------------------

Mesh generation time : 5 s
Solver time : 30 s

Total time : 35 s

--------------------------------------------------------------------------------

I used your model!

 

[wandering heart]

I just learnt a little about CST and want to calculate band structure of PBG.
But now I have no idea how to get band structure with CST.
Is a parameter sweep necessary?

Thank you!

Could you please post your projet file? I have done quite a lot of eigenmode calculations for photonic crystals by using CST and also RSOFT Bandsolve.