Troubles with dense mesh in CST2006

I have a model in CST2006 feed by the lamped port. When mesh become more complex and dense the solution is wrong, stucture reflect all energy, i.e. output and input signals are equal!
When the grid is small (~20000 cells) the solotion is normal...
With the same grid options in CST 5.1.3 the solution was good, at least it was quite reasonable.
Where is the problem?

PS CST 2006, SP4.

Thank You !!

I met the same question.It offen happen in calculating of microline or stripline.

Yes, I met this effect in the same cases! How overcome this problem?

Can you guys be more specific in your setup? I have just recently started to use CST, and have not yet encountered this problem (although I have tried).

Are you running a transient or frequency simulation when this occurs?
What kind of problem are you simulating? Is it just a microstrip line or something more complex?

Thanks!

-john kitt.

Please explain more detail of your project... from my experien first thing port must be put in a global coordinate, grid should be specify at boundary between two material(manual set up).

If posible post your project picture or draft.


Somkit.

Just such a project:
a 50 Ohm stripline with high b=1mm , t=0.017mm ,width=0.6155 mm,
use transient solver ,adaptive mesh refinement.
the error will appear,
s11=1 s21=0 , but port calculation is correct.

use SP5 maybe this will fix the problem.

pl

just as Attachment

I met this kind of problem when I simulated a simple half wavelength of dipole. When the mesh is about several 10K, the results is correct. When mesh reached several 100K, the S11 become 0dB.

Yep, I ran your model, allysun, and got the same incorrect result. I wonder if anyone tried it yet with SP5?

Somebody please tell us where to get SP5.

Dear allysun,

I ran the file you have provided but didn't see any problem.

S11 better than -80 dB
S21 nearly 0 dB

I used both your dense mesh setting and also a very coarse one. Both results are fine. Mine is SP3.

Regards

I use the fixed version of CST 2006,and met the same problem,when mesh is smaller,the result is incorrect.

Someone said you may find special->Solver->Use parallization up to 2 CPUs, and uncheck it. Then you may get the correct results.

your license is wrong

I tested. It vcan get the correct results.
Thanks !

it can work!,Thanks!