Coax fed antenna - problem is the return loss diagram

Coax fed antenna antenna

Hi friends;

I'm simulating a Zeroth Order Antenna with HFSS excited by a coaxial probe feed (See article & project in the compressed file),my problem is the return loss diagram ,

1-I can't find the same diagram as the author's one;even i've read a lot of comments about the coax fed design in this forum.

2-I can't determine the input impedance of the antenna?the author has not signaled any specifications about the caox feed radii.

3-Does the length of the coax feed change the S-parameters and the resonance frequencies?

4-Which model to use driven or terminal?

5-Does mesh refinement at the coax (outer and inner conductors) matter?

Could you plz help and thanks.

Re: Coax fed antenna antenna

You have problem with the coax.

If you have the HFSS manual, have a look on the first example about prob feed antena.


If you dont have it, let me know

Re: Coax fed antenna

Hi Mustapha,

Thank you for the reply,i have the HFSS tutorial where vacuum is used as substrate.My problems are how to determine the input impedance of an antenna fed with a coaxial cable? ,and which substarte should i use (Teflon for real coax cable with outer and inner conductors-(copper or PerfE???) or vacuum).

Could you run the project,it want take a lot of time, have a look to the S-parameters,when compared with author's one ,a shift at 8 GHz exists for the zero order mode and high return loss -9 dB for the first negative mode.!!!!

I'm so disappointed with this project because i've tried everything but without results and as you said Mustapha there's a problem with the coax cable.
Thanks.

Re: Coax fed antenna antenna

Hi MrElec,

Wether you use teflon or vacuum as the dielectric for your coax will certainly change the characteristic impedance. I don't know what impedance the authors design there coax cable for but generally it's 50 ohms. I suggest you do the same. Also when looking at your design I noticed a perfect H plane within the coax feed, did you mean to do this and if so for what purpose? Usually when I'm doing designs with HFSS I use the driven modal solver.

Re: Coax fed antenna antenna

Hi YoungEng,

Thanks for your help,i've used the perfect H plane having the same radius as the coax for the boundary condition so that the tangential composite of the E field doesn't change when the wave travels from dielectric 1 (teflon) to dielectric 2 (Rogers Duroid).Generally an antenna designer uses 50 Ω as coax cable impedance and then he try to find the feed location on the patch for matching.

My trouble comes from this point,i should know the imput impedance of the antenna.

Do i have to design a coaxial cable with 50 Ω impedance and then preceed with deembedding?.

If you have seen my return loss diagram, it is different from author's one.Really i'm stucked.
plzzzzzzzzzzzzz i need help from edaboard members.

Re: Coax fed antenna antenna

I believe you should design the coax for 50 ohms and then parameterized your coax so that you can moved it to different locations on the patch. You can setup an optimizer to find which location gives you the best return loss.

You should note that if you have designed the coax for an arbitrary Zo then that is the reason for the poor return loss. Also deembedding is used for phasing your ports and won't have much affect on the return loss.

If you plan on building this antenna later on you might want to include to effects on the fields within the coax due to the change in dielectric. This change in dielectric inside the coax will also play a role in the overall return loss of your circuit. You might want to simulate just the coax with the 2 different dielectrics and see what kind of return loss you get. If you get a return loss thats worse than 25-30dB you might want to consider keeping the coax cable as a single dielectric and drilling a hole in the substrate.

Re: Coax fed antenna antenna

Hi YoungEng,

Very thanks for your help

I believe you should design the coax for 50 ohms and then parameterized your coax so that you can moved it to different locations on the patch. You can setup an optimizer to find which location gives you the best return loss.

Click to expand...

The feed location is fixed by the author ( 1.5 mm from the center of the antenna in the x direction),it was published in an article of IEEE see the winrar file.

You should note that if you have designed the coax for an arbitrary Zo then that is the reason for the poor return loss. Also deembedding is used for phasing your ports and won't have much affect on the return loss.

Click to expand...

Effectively the poor return loss is due to an arbitrary Zo of the coax,but deembedding is used to get the input impedance of the antenna as i've read in HFSS tutorials,so if i simulate the structure with a 50Ω coaxial cable i can get Zin=50*(1+S11)/(1-S11) is this correct?!

If you plan on building this antenna later on you might want to include to effects on the fields within the coax due to the change in dielectric. This change in dielectric inside the coax will also play a role in the overall return loss of your circuit. You might want to simulate just the coax with the 2 different dielectrics and see what kind of return loss you get. If you get a return loss thats worse than 25-30dB you might want to consider keeping the coax cable as a single dielectric and drilling a hole in the substrate.

Click to expand...

What do you mean by "keeping the coax cable as a single dielectric", does this mean same dielectric for the substrate and the coax?!!

How can you explain the shift in resonance frequency at 7.8 GHz (simulated by author) and 8.25 GHz (my simulation).?

Re: Coax fed antenna antenna

Hi MrElec,

You can get the impedance looking into the junction where the coax meets the substrate by calculating the return loss as long as your coax is 50 ohms, or at least you know the Zo of the coax. However if you are interested in the Zo looking into the patch itself you must remember that as the probe of the coax extends into the substrate up to the patch the Zo will change. This will make the calculations a little more difficult.

The shift in the resonant frequency can be due to a number of possibilities. I didn't read the paper but if your goal is to just replicate their results then you should simply make a replica of their circuit. Of course that depends on how much info is given to you by the authors. What could be done is to take their idea and make your own circuit. You can then use HFSS and try to make improvements on their design.

What I meant about my comment is to investigate the effects of changing a dielectric in a coax. I haven't done this myself but perhaps its better to have the dielectric constant of the coax close to the dielectric constant of the substrate, or maybe not. In any event it would be interesting to know.

Good luck with your project.