CST:Different results with Time and Frequency domain solver

[prafuldeo]

cst models

Dear friends,
I was comparing the time and frequency domain solver in CST MWS 2009 for the accuracy of the results and I found that the results in both of the solvers are very different. I've been reading about these solvers and everywhere they are showing nearly identical results so I am curious why my results are not matching. I use time domain solver for my antenna simulation but wanted to check the frequency domain solver too. I am using adaptive meshing technique to get the correct meshing. Reults are not same for two solvers and thats strange. I've attached with this post two cst models. One is dipole in free space and another is same dipole pierced through the dielectric slab. The total length of the dipole is 500 mm with wire radius r= 0.25 and it is fed via discrete port at 125 mm. Same dipole is pierced at 45 degrees in dielectric slab of Er = 16 with slab thickness of 100 mm . Now when I simulate the only dipole (First model) in time and frequency domain the results are different and same is happening to my second model i.e. dipole with dielectric. It will be great if someone just look at the model and suggest me where I am making a mistake and how to get the same results in both the solvers. Thank you.

Regards
Praful

 

[shahid78]

cst frequency range different results

hi prafuldeo

i was trying opening your model in cst 2008 but i am having problems

i think you are using cst 2009, isnt it?

let me know how can i open your model in cst 2008


regards

 

[bkproteus]

cst frequency domain

I'm using CST 2006 to simulate the spiral antenna with EBG. But the time for running is too long. It take about few days running.
I wonder if the newer version of CST such as 2007,2008,2009 can do it faster. Can some one tell me about that and please give me some advices for faster simulation. THanks

 

[jpsmith123]

time domain solver cst

Hello Praful,

I experimented with your free-space dipole problem a little bit.

When I refined the mesh (transient solver) somewhat, the results (S1,1) of the two solvers got closer. I get about a 0.5 db difference with about 4 or 5% frequency shift.

I also get about this much frequency shift when I compare the two solvers using CST's dipole example.

(Bkproteus, since you have a different subject in mind, you should start a different thread).

Regards,
jpsmith123

 

[prafuldeo]

how to get time domain from frequency domain

Hello jpsmith,
Thanks for your reply. When I was increasing the mesh using adaptive meshing in Transient solver there is always a frequency shift of 5-8%. But yes the key is to increase a mesh cells enough to cover all the finer details of the structure. Thanks for giving your time.


Thanks & Regards
Praful

Added after 3 minutes:

Bkproteus]: As jpsmith123 suggested that opening a new thread would be better as much more people will read it. There are many things which affects the simulation time. Its not too much difference between CST 2006, 2008 and 2009 but if you have a choice then go with latest edition. If you have any particular model you want to get checked then post it here and I'll try to figure out the problem.

Thanks & Regards
Praful

 

[mirshahram]

frequency domain solver

I think when you use time domain solver you can get the reflection phase for TEM but in frequency domain solver TE and TM. I cant underestand : how can you check your structure in !F solver would you please explain more? I use CST 2008 and I also can not open your file.