hfss radiation box
I have to simulate a very easy antenna (a dipole) surrounded by a radome (a cilindrical or spherical geometry) of many different thickness.
I've just find out that if I change the dimensions of my Air box (I've done a cilindrical air box) my radiation patterns change a little bit.
Do you have any suggestion fot the size of the air box?
I have to compare the results of the radiation pattern with the only dipole (no radome around it) with the radiation pattern of the same dipole with different thickness radome around the antenna.
Should I use the same air box dimensions in all the simulation to have comparable results?
I've read in other post I should use an air box from 2 to 10 times the lambda value. But for 2.4 Ghz (that's my freq) it seems to much for the simulator..
Any suggestion?
Thanks a lot.
i guess quarter wave distance from each boundary would be enough. For your case 35 mm far away from each side.
see my response here
https://www.edaboard.com/ftopic310551.html
I would concider several things. Spacing the radiation boundary about a quarter of a wavelength out is probably enough.
You might want to seed the boundary and the radome to improve accuracy. I would also consider making the radome "vacuum" as the reference and keep the radiation boundary constant in all your comparisons.
You might also want to fool with the convergence to let it run more iterations and settle down before stopping.
This is a natural for a parametric model and multiple runs with the appropriate changes for your particular situation.
If the model starts getting too big you can use symmetry boundaries to reduce the size to a more managable level. Or just get a bigger computer.
yes i am having the same problem. I put airbox λ/4 away but it didnt solve; and i put airbox λ/2 away this time s11 parameters are changed; i made rectangular patch and put airbox λ/4 away it did well , but it doesnt give clear result for dipole,
2D2/λ= Far field ,
D=longest part of antenna , so if i make dipole λ/2 at that length, the longest part will be λ/4 so that means; 2.(λ/2)2/λ=λ/2 far field zone , am i wrong?
u can use PML boundaries to avoid this problem than u can use airbox at lambda/4 height also
From my experience, λ/4 is often not enough for good radiation boundary calculation, although it is advised in manuals.
The only way to know for sure if the software gives convergent results is to increase the accuracy of calculation and check with the previous results. If they don't change, they've converged and you can be confident about them.
Also, if HFSS says that adaptive analysis converged, it means that the algorithm achieved the maximum allowed error during the mesh refinement process at the solution frequency. If you are doing a sweep, this may not mean that your S-parameters in the whole range are convergent, so probably the sure thing to do is try also on a few solution frequencies and compare.
I would advise solution frequency definitely to be above the center of the simulated frequency range.
In all my HFSS simulations I always start with a cubic air box.
Set Box Parameters:
Position - (-x_box/2,-y_box/2,-z_box/2)
X = x_box
Y = y_box
Z = z_box
Make all dimensions of the box equal to your wavelength
If you haven't built up your model to be iso-centered on the origin of the axis, move it there.
Now you will have lambda/2 distance in all directions from the center of your structure.
This has always worked great for me. Don't forget you have to select the box and assign it a Radiation Boundary Condition...
This basically means that you are dealing with very small antennas. Otherwise your structure would penetrate the cubic radiation box.
So, I guess this advice is good just for very small antennas.
A few more questions: Have you ever tried increasing the size of the cubic box and check if the results change? How do you know that this worked for you every time? Are you talking based on comparison with measurements? What was the level of agreement of simulation/measurement that you were satisfied with?
OK, but if I have an antenna (and in my case even a radome around the antenna) bigger than lambda/2, what could I do?
Isn't it correct to have the air box at a distance of lambda/2 or lambda/4 (like said someone) from the border of the structure we want to simulate? So if the structure increse the air box will always be lambda/4 more than the structure.
Or has to be lamda/2 or /4 from the center of the structure we want to simulate?
-Another things:
Does matter the shape of the air box? For example cilindrical or cubic.. In my case I made it cilindrical because the object I want to simulate is nearly cilindrical.
Thanks for your help.
I was gonna suggest using a different type of geometry for the air box, other than cylindrical. The absorbing boundary conditions in numerical EM methods are usually formulated for the normal plane wave incidence case. So they are most efficient when the incident wave is quasi-normal to the absorbing/radiation plane. I guess a dipole (especially if small) would radiate spherically. Therefore putting a spherical airbox would be the best choice. You can also use rectangular boxes, but these are usually inefficient around their corners. If you can predict the far field radiation pattern with educated guesses or error prone simulations, then you can play with the orientation or the shape of the air box so that the faces are normal to the most radiated directions.
There was a question which is not answered:
Whether the airbox distance from the structure is important or from the structure center?
The distance from the structure boundaries to the airbox should be more than lambda/4, but even more, according to my experience. The shape of the airbox also plays a role, so play with it to see what happens.
But the distance from the center of structure to the airbox is not important. It works well for etothe456t because obviously he/she is simulating only electrically small structures.
in my point of view lemda/4 is enough. and also try to make the same distance from all sides. the s11 are changing slightly with the cahnge in airbox size because by doing this actually you are altering the electrical size of your antenna.
i am ususally use lemda/4 to simulation antenna
and i try increase the air box , feeling s11 the same
maybe you can use lemda/2 try it