Re： EBG structure modeling in CST

hi all
yes i agree with piboo. i have seen some articles which model microstrip line over EBG structure to characterize the surface wave bandgap of EBGs,
Prafuldeo can you post your model land s21 result so we can use whether you are making any mistake in modeling or suggest something else to resolve your problem
regards
shahid

Hello piboo and shahid.
Thank you for replying my query. I've attached my cst model which is the same as Samii's Microstrip antenna with EBG paper. I've modelled 8x5 EBG with 50 ohm microstrip line over it. But with this method I can't match the results with paper results. Atleast I should be getting the same band-gap as in paper. I've also attached the result page from paper. Also in the same paper samii did his simulations using infinitesimal dipole source with ebg in between. Can we do same in CST? Please let me know how to obtain the correct results using CST. Thanks and Regards.
Praful

hi prafuldeo
i am not able to open your model. which version of CST you are using?
can you try to send the model again
regads

hi shahid
to open the prafuldeo's project:
1/ open the project --> bugg
2/ insert variable p, g and h
3/ Open and update the history list
Then it will work
regards
Piboo

Hello Shahid.
I've used CST 2009 for this project. Please let me know if you can't open the project.
Regards
Praful

hi prafuldeo
well i am still not able to open the project
i tried defining the variables p,g and h and then tried to update the history list
but it says variable values are too small.
so what values should i use for p, g and h?
regards
note: also i cant see this model of micro strip line over EBG structure in the Samii paper which you mentioned in your post?
so can you tell me to which paper results you are comparing to ur results?
also how u decided about the microstrip line width ? it should be 50 ohm line for matching with wave port. so which thickness and Er u used for calculating the microstrip line width for 50 ohm characteristic impedance?

Hello Shahid..
Thanks for reply. The paper i m trying to simulate is Fan Yang and Yahya Rahmat-Samii, ?Microstrip Antennas Integrated with Electromagnetic Band-Gap (EBG) Structures: A Low Mutual Coupling Design for Array" . The model I've uploaded is the simple EBG design based on paper. I want the band-gap for that but using finite EBG structure say 6x6 cells . The periodic boundary conditions method simulates the infinite EGB but its no good for real conditions. The microstrip I used is 50 ohm matched. But results I got is not matching the published results. So if you have simulated this paper please let me know. Also I'll upload CST 2008 model for the same so that you can open it.
Regards
Praful

hi prafuldeo
well thats why i asked you in my last post. i can t see any model in your refered paper which is simulating microstrip line over EBG structure.
the figure which you mentioned is for infinitisimal diple surrounded by finite EBG cells.
so how can you get the comparable result for your model and this paper model?
can you explain?
also i asked you how u model 50 ohm microstrip line?
i can't get 50 ohm for your modeled line. thats why i asked you which thickness and permitivity value u choose for calculating the width of the line
i hope now u understand whay i am asking
regards

Hello Shahid,
I'll try to explain my query here. I've posted new thread " EBG surface wave characterisation using Samii setup in CST " please have a look at that also .
Now for complete EBG characterization we need two results (1) Reflection phase (2) Surface wave band-gap. methods of finding reflection phase ( macro method & wave guide method) is discussed in many places in this forum and there is no confusion associated with that . Now I am looking for the best possible method (in terms of accuracy) for obtaining surface wave band-gap. There are many setups to do that. One is as you said Full or terminated 50ohm microstrip line. But it gives the filter characteristics of the EBG and I am interested in the surface wave attenuation of EBG.
Second method is to model one column of EBG and put waveguide ports on both ends and make boundaries periodic to mimic a semi-infinite EBG and measure the S21 to get the band-gap. Now this method is close but it doesn't give band-gap information for finite EBG, say 6x6. The tird alternative is using samii set up which uses the infinitesimal dipole source and reference plane on the edges. For antenna purposes I think this one gives most accurate results as it gives the band-gap results depending on number of EBG rows. So I was wondering is it possible to simulate same setup in CST? Because if we can generate the same results in CST then we'll know that other EBG's which we'll design is giving correct results and it saves the time and effort of fabrication and measurement. I've uploaded samii model in my new thread.
Have a look and give some thought about it.
Thanks & Regards

Hello prafuldeo
well i appreciate your detailed explanation on EBG characterization. i try to answer some of ur queries.
well i think surface wave characterization using 50 ohm microstrip line is the best for finding surface wave bandgap of finite EBG structures . i dont know why you are saying its giving filter characteristics of EBG and not surface wave bandgap.
in my view there is not much difference between the two. after all bandgap means filtering of waves and by looking at S21 u find the waves that have been filtered by EBG.
regarding the Samii setup , yes i am also interested to model the same setup in CST.
i was looking at your model. i dont understand why you have used wave guide port at the edge?
becoz in this way you are exciting two ports , the discrete port and waveguide port which i think u dont want. correct me if i am wrong.
i will also try to think and model the same setup in CST. and let you know if i get convincing results and i hope u will do at ur end and share your observation.
however , i think u need to use probes to capture the electric field at the edges for the all the frequency range. but in Samii paper they have plotted E2. so u need to do some processing in excel to get the same results.
i hope u will succed in getting good results.
hope to be in contact with u on this topic
cheers
shahid

Hello Shahid,
Thanks for your reply. I was thinking about using the probes but because samii used reference plane so I used the full face waveguide port . But I'll surely try probes too. Have you simulated the same Samii EBG by using microstrip line? Did you able to get the band-gap similar to Fig. 2 of the paper? Please upload the microstrip model of the samii paper if you have one, so that I can check it too because for some reason I am not able to get the correct results for it.
Thanks
Praful

Hello
"-frequency domain solver:
1/using unit cell and do a de-embedding "
I tried simulate EBG with unit cell method, how says piboo, but not succesful. Please can you tell me someone, how is exatly settings for one cell, I used "unit cell" for "x" and "y" walls and "Zmin=electric" and "Zmax open(add space)", default port in Zmax, but I do? no, if I must use now some port too...next I used Theta = 90 ... I add my model on the screen...
Thanks very much

hello meerek,
make y=magnetic, rest electric and de-embade the port to top of ebg patch. then use transient solver and look for arg of s parameters. you should get the reflection phase.
Regards
Praful

prafuldeo
I tried your settings, but when i place port on top of ebg patch CST write error - Wavequide is to short -,or do you mean place the port at some distance from patch?
thanks

Hello meerek,
yes put the port at some distance and in port option de-embbed the port to the top of the patch. I think its under distance for reference plane option. Try it otherwise I'll set it for you just upload your model.
Regards

I have a question.
I put the waveport at 10, and the top of the unit cell is at 0. Does it mean I should put the distance away at -10?
If it is right, I want to know whether the top of the unit cell having a PEC condition will influence the result?
Thx.

hi piboo
well i was trying to use your method of deembeding the port for unit cell template in frequency domain solver.
but i am getting an error when i evaluate the post procesing result.
let me ask you some questions.
in you post you write this in the 3rd step;
/ You need to have a file which contains all the frequency points. For that Copy a *.sig file in your CST folder and rename it as Fband.sig (for example). Then In CST copy one of your result, for example, your phase result.
Open Fband.sig with excel, delete all and paste it.
i dont understand what do u mean when you say "then In CST copy one of your result, for example, your phase result."
we have already laoded the phase results using laod 1D results. and in this step all we are doing is create a file which contain all the frequency points. so i dont understand when u saind in cst copy one of your result?
also when i open the *.sig file in excel. all the data is contained in one column.
so how could you say delete all the data but not the first column which contains the frequency points?
i will appreciate if you could answer these questions
regards

Hi Prafuldeo
long time no new updates from your side.
how you are getting along with EBG simulation?
any success in simulating EBG surface wave bandgap in CST
looking forward to your comments
regards

Hello Shahid,
I've finished the EBG simulations. Regarding surface wave bandgaps the model I sent you is the best and giving me satisfactory results so I am sticking with it. Every software gives slightly different results so until they are very different I m ok with it. You tell howz your work going? Did you do anything after simulating that double square FSS?
Regards
Praful

Hi piboo,
Have you finished writing your application note on different methods to characterized EBG in CST? Pls upload the document. It would be really helpful for me since i am looking for some tutorials about EBG in CST.
Thank you very much.
Le