Anular ring patch simulation with HFSS

floquet port hfss

Hi everybody,
I'm trying to analyze an anular ring patch on ground plane that will be used as a reflectarray element with HFSS ver.11. In order to define the ring size I need to know the reflection coefficient phase curves toward frequency, ring size and field incident angle variation.
The first attempt is analyze it as a frequency selective surface so I set an incident wave, master and slave boundary, PML or radiation boundary and I find the first problem:
- if I put the second radiating/PML surface far (quarter wavelenght) from the ground plane HFSS gives a warning:
The background does not fully touch the following boundaries: 'PerfE1_ground_plane'. This may produce an incorrect result.
- If I put it adjacent to the ground plane it gives me the obvious warning:
Warning - Boundary 'PerfE1_ground_plane' and Boundary 'Rad2' overlap.

Neglecting the warnings I analyze the structure and plot S parameters table finding the real problem:
the table contains "Nan" values so it seems that HFSS couldn't get the correct solutions. The strange thing is that in the solution data (Results->Solution data-> Matrix Data) I find the S11 value while the others S parameters are all zeros.
I can't undestand if the problem comes from PML settings or convergence, or something else. I also thought of plotting field data and getting reflection coefficient dividing manually reflected field by incident field but I can't plot any kind of data relative to fields because I can't set any geometry in the field plot menu (and sometimes reflected field strenght is two or tree times greater than incident field, this is quite strange...)
I also tryed to use the floquet's port but this excitation is listed but not settable in the Excitation-> Assign menù.
Can anybody help me please?
Thank you
Alessandro

radar cross section plane wave hfss

Hi Ale, or I would have to say 'Ciao' since I'm Italian as well. I'm gonna use English since the forum must be enjoyable for all members. I work at the Microwave Laboratory of UniCal, CS. Where are you from exactly? I'm into reflectarrays as well, in particular active quasi-optical arrays and reflectarrays. I checked your files. I didn't have so much time for correcting the PML's one, but I was succesful with the 1st one. The only thing you have to do is to use a driven modal solution: HFSS-->Solution Type, Driven Modal. Results look good and wihtout errors!!!
As of Floquet's ports' use, the necessary conditions for using them are: Driven Modal Solution and Master-Slave boundaries. For plane waves use TE00 and TM00 modes, which are called 'specular modes'. Phase delays may permit you to define array scanning properties and should be in aggreement with periodic boundaries phase delays.
I will try tomorrow to find a few minutes and have a look at the other file (the one with PML).

Let me know how it goes.

Ciao e a presto.

I.

floquent port hfss

Hi Alessandro,
I checked the second file. In order to avoid the number of warnings during validation check, you should apply Master-Slave condition not only on aibox faces, but on PML box faces as well. So, draw 4 rectangular sheets, for each group of PMLs' and airboxes' faces, and apply M-S. Use Driven Modal Solution type. Look at the attachment.

Hope this helps.

Ciao,

Ivan

Added after 7 minutes:

PML file

anular ring

Hi guys!
I'm working to the same problem (look at my recent post)
Now i'm trying to use floquet ports in hfss 11.
i'll tell the news as soon as i get some useful results
bye

rcs floquet hfss

Hi Ivan, thanks a lot for your advices, you showed me what was wrong! I agree with you about the forum language, we should adopt english rather than italian!
I still have some questions:
I saw that you didn't set infinite ground plane option in the perfect E boundary assignment of the ground plane,is there a specific reason?
I noticed that using PML and radiation boundary I can't get the S-parameters phase (always zero) while I obtain it using floquet's port. I red the documentation but I still don't understand clearly what HFSS means for floquet's port 1:2 in result plotting.
As you well know the reflectarray design needs the knowledge of reflection coefficient phase behaviour. To achieve these kind of data I activate the deembedding post processing option in order to obtain the phase of S(FloquetPort1:1,FloquetPort1:1) at the reflecting surface (ring metalization) but I'm not sure of the results. I expected at resonant frequency magnitude near to 1 and phase around 180° (Γ=-1, metallic reflection) but for the second time (same results with period structure dedicated software) I get Γ=1.
Which result is correct?
Thanks again for your help!

Clodxp I'm interested about your studies, I'm waiting for your news.

Best regards,
Alessandro

floquet polarization hfss

Hi Ale,
if you use infinite gnd plane the background plane doesn't fully touch ground plane. Indeed, the background material inside the airbox is vacuum or air, which is enclosed into a finite volume, has a limited area section. If you set an infinite ground plane, the remaining part of the gnd, outside the box, doesn't touch the enforced background... I hope you understand what I mean... So use a normal PEC: using master-slave it'll be like having an infinite gnd plane since this boundary reproduce infinite periodic structures and fields.
For whom it concerns the reflected phase. For this kind of problems you gotta use Radar Cross Section (From Far Field Results), in particular 'complex monostatic RCS', which retains information about phase. Use monostatic RCS (normal) for getting information about reflected amplitude (choose the polarization you are interested in).
I think this model is good enough for designing the single unit cell of the reflectarray, since you can get amplitude and phase info with RCS.
As for Floquet's stuff: I think you have to study Floquet's theory at first, tomorrow I will suggest you a couple of good books. It is basically an expansion of an infinite array Green's function and thus represents a complete set of modes which can describe the fields related to the infinite array itself. In HFSS, the two specular modes describe all the possible polarizations for a plane wave in FF, infact they're orthogonal with each other. For getting reflection information, use the incident plane wave model anyway, I think it's easier to use.
The reflection coefficient from a PEC wall must be -1, in theory... I'm not sure, but try to get complex RCS re-setting the airbox height, 2*lambda0, lambda0 from each side of the unit cell surface and fix the position of the incident wave lambda0 far from the ring. This is not important, since you should build a phase curve with respect to ring dimensions and then normalize the curve to a reference value.
Try with Ansoft Designer too, it provides periodic boundaries as well.

I hope this helps.

reference for fss radiation

Hi!
I looked at your simulation and i found some strange (for me) things.
Please can you explain me why the air box extends also below the ground plane?
Moreover i looked at the radiated field and i found that the maximum field intensity if below the ground plane, where i expected to have no field! (i looked at the total field, not the incident only).
So i guess there is some basic concept about the hfss i don't know.
Can you explain it to me?
thanks

phi theta refrence floquet hfss

Infact there's no need of extending the airbox besides the gnd plane, because no wave transmission occurs behind it (the wave is completely reflected). For a most effective (less time consuming) and precise simulation, of course, reduce the airbox as much as you can, so terminate it at GND plane. I just didn't modified this aspect in Alessandro's files 'cause I mainly concentrated my attention on his precise questions. Anyway, your consideration is right.
I will post the corrected file, just let me modify it entirely.

Regards.

Added after 53 minutes:

File

anular patch

Useful book for floquet's theory: 'Theory and Analysis of Phased Array Antennas', Amitay, Galindo, Wu, Ed. Wiley

the background does not fully touch edaboard

Hi!
i sugguest another useful book i'm using:
Phased Array Antennas: Floquet Analysis, Synthesis, BFNs and Active Array Systems
Arun K. Bhattacharyya

I'll analyze your new patch.
Bye
C

Added after 20 minutes:

For Ivan_:
I saw your example (v2) and i observed you used a radiation condition instead of a PML. Moreover you set (in the advanced condition) incidence field - reference for FSS. What is this option for????
Moreover, why u did not used PML?? since this is normal incidence??
thank you
C

convergence floquet ports

Hi clodxp,
have a look at the following topic:


PMLs are used to simulate materials that absorb outgoing waves, minimizing numerical reflection due to the discontinuity between the air material in the airbox and the non-material background outside. In general, PML are useful with a non-normal incidence, respect to radiation, since they define an anysotropic material. Thus reflections from PML are much less than Rad.boundaries for non normal incidence. There are two types of PML applications: free space termination and reflection-free termination. With free space termination, PMLs are associated with a surface that radiates into free space equally in every direction. PMLs are more appropriate than radiation boundaries in this case because PMLs enable radiation surfaces to be located closer to radiating objects, reducing the problem domain. I simply didn't have so much time for modifying the PML's file, but you can try and compare the results with the one with rad.boundaries. As you enforce a non normal incidence, differences should start to become evident.
As it concerns FSSs' reference option, it is useful when you want to analyze higly reflective structures and specifies the radiation to be set as reference for the input wave and for calculating the reflected field. Since you can define only one input and only one output, you can define just one reference.

Now I wanna ask you something. I was looking for the book 'Phased Array Antennas: Floquet Analysis, Synthesis, BFNs and Active Array Systems
'. Last month I looked at the review and I believe it is very very good. But it is quite expensive... I was trying to find some cheap vendor!!! How much did you pay it? Any direction of where to buy it?

phase delay floquet hfss

Thank you Ivan_!
Regarding the book, i'm sorry, it's not mine, it's from a library. I've used for some days.
byez C

reference for fss hfss incident wave

Hi guys!
I've spent a lot of time trying different setups with parametric simulation and finally (maybe) I find what I was looking for. To get phase curves toward inner radius variations I plot complex bistatic RCS (along the right polarization vector) phase with PML/radiation boundaries and incident plane wave, just as my helper Ivan said!
I find strange that the curves show very fast variations (too fast, steps greater than 100°) at resonant frequency and some local minimum value before it. RCS (σ) is not linearly tied to reflection coefficient (Γ), I find that:
σ=(4*pi*R^2)*[abs(Γ)]² (R= distance)
May I apply an inverse formula calculation?
Thanks,
Alessandro

Thanks for advices regarding the books, I will sure borrow one!

background does not touch hfss

Ale,
use S11 for amplitude and for phase complex bistatic RCS. That's right... or get the phase from the complex monostatic RCS. It's just a matter of phase reference. For bistatic, remember that transmitter an receiver are not at the same location. If you are interested in normal reflection get the central value (theta=0). after you build the phase curve for increasing values of radius, fixing the frequency and the angle of reflection, normalize all to the central phase value (the one for the intermediate size in the range of radius variation). That's how it's done, don't worry...

Bye

floquet rcs hfss

hi cicocicobum,i am abdoeng,i already make some models under ansoft designer,it so easy than hfss that i don't make any macro files,also i use cst in my simulation ,they give good results.i hope u contact me
jamywood4000@yahoo.com
abdoeng78@gmail.com
i look for contact any person in reflectarray interesting

using pml and infinite ground plane hfss

Hi!
I've analyzed the ring patch (from cicocicobum) under HFSS10.
I've considered six different incidence directions:
1)θ=15°,φ=20°
2)θ=30°,φ=20°
3)θ=45°,φ=20°
4)θ=15°,φ=160°
5)θ=30°,φ=160°
6)θ=45°,φ=160°
In the graphs i report what i got. Moreover i attached the hfss file.
Can anyone confir me you got the same results?
Do you think these results are reasonable? I think there's something strange since the maximum is not exactly in the specular direction, and the difference grows as i increase θ;. Moreover, since i considerd a 0.5λ; cell
Do you think i should improve the starting mesh??
Thank you

hfss 11 ebook

I forgot a consideration:
since i considered a 0.5λ cell the high order floquet modes are in the invisible space. As consequence they cannot influence the obtained pattern.
Thankx C

anelar patchs

Hi ClodXP, I obtained the same results with your files.
Maybe are caused by the reflection coefficient of the RA, elements resonate at a lower frequency respect to 4.4 GHz (take a look to the field pattern, it's greater @ 4.36 GHz in field animation) the surface impedance is not purely resistive and the array works scanning a bit the beam direction (it's a RA skill).
I attach some pictures from a old thesis that show how the reflected field phase rise if you increase the incidence angle, this can explain why the maximum doesn't aim the specular direction and the difference grows with theta.
This is just an idea, I don't know how to verify if it's true or false. Any suggestion??
I have a question: why do you set (phi + 180°) in slave phase delay settings? I thought that the values required were just the incidence angles theta and phi.
Regards, Alessandro

hfss floquet infinite array

Hi alessandro!
I used (180°+phi) in order to get to specular reflection. A similar approach is used in the attached ansoft training presentation (page 17).

Regarding the simulation, if i use hfss 11 with the Ivan_ file, for some of the angles (th=15,phi=20°) considered in my simulation i get the maximum in the specular reflection. It is very strange, but since i'm using a beta version i dismissed the version 11.
If you get other results with the v11, please let me know
Thankx C

hfss reception of incident wave

Hi guys,
which boundary are you using for non-normal incidence and reflection? I told you PMLs are more suitable than rad.s in such a case. Anyway i suggest, to proceed with a first simulation in Ansoft Designer, for those frequencies it is quite good, as abdoeng said. Then repeat simulations with PMLs and compare results to have a better comprehension of what's happening.

Bye