HFSS -Antenna Boundary conditions

May someone please help me in this regard


1. What should be the AIRBOX size dimensions be. ...(generally considered to be at a distance of lambda/4) for an antenna?

Is it from the radiating surface alone, Should I consider from edge of gnd plane (patch antenna)éé

2. Should I consider the box size( varying continously) wherein the return loss doesnt change any further?

4.How about the meshing for the AIRBOX ? does meshing it finer improves ur accuracy ?

5. How does PML and Radiation boundaries differ in HFSS?

6. Will I get same results if i use PML and Radiation boundary.. for a simple patch antenna?

1) airbox should be 1/4 lambda for radiation boundary and 1/10 lambda for PML. if your antenna is on infinite ground plane, then airbox corresponds to ground plane in lateral dimension
2) No, if you are looking for accuracy, use PML
3) No, you can try this yourself, but use either multiorder tets, 1st order, or 2nd order tets and you will be fine on accuracy. Just remember that S parameters converge before field quantities, you may want to set convergence on your far field quantity of interest to ensure converged results. See expression cache in the help for more info.
4) Radiation boundary is a second order termination function that operates only on matching the tangential component of E...ergo the airbox size suggestion as 1/4 lambda, most E fields are nearly tangential to the airbox (away from reactive zone). PML is a 3D object that is composed of a lossy anisotropic material that HFSS determines that will perfectly match to the freespace termination up to incident angles of near 60 - 70 degrees.
5) Your results will be near identical for a single radiating structure. The PML is almost always suggested as the computation time is not that much greater due to the smaller airbox dimension needed.

have fun :)

Thank you for the reply..

Do u have any PMLsetup document?

HI Venu,
1- Airbox size dimension should be Lamda/4 as you have mentioned yourself and dun forget to use that Lamda which is being calculated using LOWER FREQUENCY of your interest.
This Lamda/4 distance should be from the substrate on which you have your radiating structure.

2-I did not understand your 2nd point very well. Explain it a bit more what you mean by it.


3- I do not think the idea of meshing the AIRBOX is a nice one. Meshing is always done on those areas where you expect to have high field concentration,so usually Radiating structure are manually meshed in order to refine the Mesh which is being generated by HFSS by default to get more accuarate results.

4- A radiation boundary is used to simulate an open problem that allows waves to radiate infinitely far into space, such as antenna designs. HFSS absorbs the wave at the radiation boundary, essentially ballooning the boundary infinitely far away from the structure. The distance of Radiation boundary from your structure should be Lamda/4.

Perfectly matched layers (PMLs) are fictitious materials that fully absorb the electromagnetic fields impinging upon them. PML are usually placed very closed to the radiating structures like LAmda/10 distance away and they fully absorb any waves impinging on them. PML boundaries are preferable in those cases where you want to reduce the domain of your problem as you can place them much closer to your radiating structure and you do not care for the effects at a distance far away from your radiating structure.

In case of antennas always use raidation boundary conditions and trust the result you are getting from them. If you ll use PML and put them at Lamda/10 distance from your antenna structure then you will get messed up radiation pattern results which you have to see in the far-field region.

According to my experience Radiation boundary conditions are perfect in case of antennas and PML are used for designing of periodic structures such as high-impedance surfaces with forbidden frequency gap because in those structures you do not care about the effects in the far-field region.

5- I think the answer of your last point is also in the above explanation. Do try to read HFSS help as it is very useful and it will remove most of your doubts automatically.


Hope what i have explained above,helps you somehow.

/SC

The PML setup is through a wizard...just try it before asking so many questions please. You can also look up mixed order tets in the help for more info.

Hi

On 2nd point if are considering an patch antenna on a finite gnd plane should the Airbox be considered at a distance of lambda/4 from edge of the plane or from patch(radiating surface)..?

But HFSS suggests to use PML instead of radiation boundary for antennas......

May u clarify which one is supposed to be used... for antennas wherein i am interested in far field pattern...

Hi,

Atleast i have never seen any single example in HFSS tutorials where they have used PML instead of Radiation boundaries in case of antennas.I have always used Radiation boundary for antenna case and got very good match between measured and simulated results. I have used PML boundary condition only in case of Periodic structures.

Take Lamda/4 from edge of plane.

/SC

The easiest way to confirm what sweetchoto says is to set up the same antenna and simulate it with both radiation boundaries and PMLs...they will be near identical...

Hey hi

Thanks for ur help guys..
As tallface65 suggested, its almost identical with PML and radiation boundary..for an antenna. But for a simple case of patch antenna PML takes a longer time rather than radiation boundary.
When I used the same for multiband antenna PML takes less time compared to radiation boundary..

@tellface65-- I was trying to go through multi order (mixed basis functions in the slution setup), when should i use it.. It states based on lambda refinement...Can u make it clearer...

Advacne thanks

I have experienced the same situation as quoted. Therefore, I guess for simple antennas PML tends to consume more time than radiation boundary due to the additional tets in the PML boundary itself. Besides, similar impedance results are obtained both from PML and radiation boundary.

What is the difference between first order, second order and mixed order basis functions. Ther's a difference in lambda (for meshing).
I find difference in results when i do first order, second order and mixed order .please help

Advance thanks

You can address these questions in the HFSS help... Search "mixed order" and the third option down is a selection "Basis Functions" Select this and there is a nice explanation about the interpolation schemes of each order.

later :)

I think PML is very useful for Antenna design because it gives accurate matching between fabricated and simulated results, i have checked this.

- - - Updated - - -

I think PML is very useful for Antenna design because it gives accurate matching between fabricated and simulated results, i have checked this.

1. Radiation boundary condition does not extrude your exterior boundary of the air box into further meshes. Thus the calculation is relatively faster than PML for sure. Its first order is simply make sure that the TEM wave have impedance match on the boundary. Which means only normal incidence waves TEM can be perfectly abosorbed. For any oblique angle, especially gaze incidence near the corner region of the air box, Radiation boundary is NOT good, and sometimes cause even instability in transient simulations, this has been stated in literature for years. (http://www.engr.uky.edu/~gedney/courses/ee624/) taught in class. NOTE, high-order radiation boundary condition can improve absorption a lot. However, for gaze angle it can never really reach the strict math convergence. Because the ideal analytical calculation is to use the Summerfield integral, which solve the matched fields by inversing a dense matrix from the entire exterior boundary integrals. This is very very expensive (much more expensive than PML).

2. PML is in theory "perfect", which means it is independent of incident angle, polarization, medium property. It is actually transparent at the PML interface, and squeeze the wave and decay it exponentially inside the PML. Thus it does not rely on any field inside your problem domain, in simple word, it "eats up" anything comes into it. However, PML has its optimal performance at different frequencies, since it is a dispersive media, because s=kappa+sigma/j*omega. where omega is the frequency. So at different frequencies, PML does have different performance. the PML parameters need to be slightly adjusted when sweeping the frequency. I do not know if HFSS is doing this. Of course, PML is more expensive since there are extra FEM elements needed to model the PML.

3. For real applications, if you know your radiated field is simple and smooth (most evanacent mode decayed out) and its main energy is not radiating to the corner, 2nd or 3rd order radiation boundary condition is sufficient, use it since its fast and does not bother you much between measurement and simulation. However, if you are designing something that you yourself even do not know what is the field going to be look like. PML is a secure way. Keep in mind by enlarge your air box, you should always get better results for both.