[Help] CPW-fed planar monopole in C$T MW$

I am just wondering if anyone has got close to the results that was presented in the paper? I am awaiting for any comments that will come through.

Thanks

Element7k

Hi there...

I am getting an error everytime I try to download your upped files...
Is something wrong ? The download counter works fine but no success(xx times downloaded) hehe :)

Maybe could re-upload them?

mogwai.

I've also had difficulties when trying to simulate antennas provided by different authors (mainly from K.L. Wong book: Planar antennas for wireless communication).

The problem is to model return loss correctly. I've had errors around +10dB or even more. Of course, me neither have an idea if the measures in the book are inaccurate. I've also tried to increase mesh density. In one particular problem there was, however not significant effect if model had 10'000 or 600'000 cells.

I wonder if anyone have managed to model return loss trustable with C$T? Maybe I'll try Ansoft HFSS because it's also available to me.

Regards,

Markus

Hi mogwai,

I have updated the files. Please try downloading again. I will very much appreciate any help.

Thanks in advance.

Element7k

The dim provided in the book are usually incomplete so guessing is needed for some of the values. Maybe this may led to inaccuracy. Moreover, some of the antennas are tuneable. K L W0ng does not usually mention what values to twitch so that makes life difficult for someone trying to replicate the antenna.

Most designs in K L Wong books need slight alterations and they are mostly simulated with 1E3D or HF$$. (Maybe dim altered because of commerical reasons) However, I feel if it is a submission to a education journal, such as the electronics letters, I think I should be able to obtain similar results at the first goal with having to alter any dim. :? :(

We have made numerous antennas including CPW ones and got good results with C$T.

Regards,
Element7k

The best results I could run from your project in MWS are attached...

Increased accuracy level (-50db - AR off) gives better results for return loss levels, but resonances are still offset in frequency...

Probably the dimensions are altered in the proposed design...
There is an apparent double-resonance near 2GHz but is not as clear as in the paper...

Maybe you have to run some parametrics to see how it behaves ... :?

Hope this helps a little bit ;)

cheers!
mogwai.

Hi mogwai!!

thanks for the simulation run. It does help as it confirms that CST can only get so close and the dims does need to be changed. It will be brilliant if anyone using IE3D can verify this result.

Cheers,
Elment7k

results of ie3d
and you can take ie3d10 for trial now , it's free for 60 days.

Dear Divan,

thanks for posting the ie3d results. They do seem closer to what is presented in the paper now. Did you change any dimensions? Maybe there is some port mode problems with my model in CST. I will have to investigate. However, as I am not a regular user of IE3D, I have one more question, about how to feed a CPW in IE3D? Is is fed in the usual edge fed manner? Or is there a special way in which you can define a CPW port?

Any help is appreciated.

Element7k

Hello!
i had tryed to use all dimensions from article, dxf file is included in archive ( full ie3d model set as well ).
substrate FR4 #1.6 mm, Er=4.3 , tand=0.02

mesh parameters was as
Highest freq = 7 GHz
20 cells / wavelenghth, 2007 cells in total
edge cells subdivision was OFF , so this was not so good, but i had no time to solve better .

SMSi Solver was used.

Coplanar port definition you can make as : Port-> Port for Edge group ,
after that Right mouse click you define "+Port" , Left mouse click will define "-Port" . Cells for deembedding arms = 3 (need more i guess , this by default)

Model is not good and result at ~5GHz is defferent from reference because. You can improve it if make automatic edge cells and may be migrate to 3d model, but i'm not sure in last sentence.

Have anyone tried it with HFSS. I have tried it without any success. Anyone can help? It seems that CPW feed is a probem with most simulator. I got a better sresult (still does not agree well with the paper) with lump feed but no luck with waveport. there is also a result from CST with lump port.

Hi, friend!

I'm also trying to simulate that CPW monopole in MWS. First I've noted that your FR4 has εr=4.4 instead 4.9 (?)
Anyway the port impedance is ~70Ω, is that correct? I guessed that the line was designed for 50Ω.

BTW, I will have a new machine with two AMD 64bit processors and 4GB of RAM :D
Still not real-time tunning of antennas but hope speed improvement will be significant..

Hi eirp,

the FR4 in the paper was 4.4. Thanks for highlighting the port impedance I will change the CPW to 50 ohm and see if it coincide with results published. I believe this antenna can achieve UWB bandwidth by changing the geometry slightly. Any suggestions?

lol dual processor AMD 64! Did you build the machine yourself? Forgive my ignorance but have you tried to see if the 4 GB of RAM is actually addressable in term of expanding EM problem size or is it just making the computer run faster?

Cheers,
Element7k

The 4 GB of RAM won't help unless you are using Windows XP64 (64 bit windows) and the 64 bit version of CST MWS. With the standard, 32-bit Windows XP, that largest amount of RAM a single allocation in a program can address is 2 GB. UNLESS you find the special boot switch in XP Pro that will let you address up to 3 GB. The 4th GB is still useful though, so your OS can use it and not force you to swap to disk when you work on a problem that takes up to 3 GB though.

--Max

Hi,

of course I'm talking about 64bit system + 64bit MWS.
My machine allows max amount 12GB of RAM, anyway I think that 4GB is fine for the beginning :D

Added after 2 minutes:

Hi,

of course I'm talking about 64bit system + 64bit MWS.
My machine allows max amount 12GB of RAM, anyway I think that 4GB is fine for the beginning :D