Element7k,the radiation effieciency is larger than one ,why?

时间：03-23 整理：3721RD 点击：

Hello,Element7k,
I have found the radiation efficiency was larger than one after I got
the farfield result,can you tell me why?
In additon,How to set the mesh properties in order to ensure precision?
beiga

I'm not Element7k but I'll answer: it's a bUg.

Hi Beiga/goxy,

sorry 4 d slow reply. Was busy with work. u r right :D radiation efficiency cannot b >1 :) Maybe try to set the solver to a higher accuracy. E.g. 50/60 dB instead of 20/30 dB. There are truncation errors if you just use -20 or -30 dB. :( Sometimes it is a compromise between accuracy and speed but if you get 1.00xxx for -20/-30 dB accuracy, then I think still ok.

I think meshing requires experience. Try to use adaptive mesh or look at the mesh and use local mesh to try and increase the mesh density around important regions of the antenna. The advance manual introduces some tactics to meshing too. I think EIRP will be better help in this department

Cheers,
Element7k

p.s. there are loads of C/S/T users in this forum anyone will be able to give u an answer :)

You help me as possible ,I appreciate very much! :D

beiga

Here is one way I examine such problems. Try a very simple case that you can calculate by hand and see if it matches the simulation. Try a short dipole first and see if it matches classical theory. This will tell if there is something wrong in the way you input the parameters or in the way the simulator works. Then try a more complex thing that still can be solved by hand. A dipole in front of an infinite sheet conductor.

Another source of error is if the simulator gives peak values and you assume rms.

Well, beiga, itsn't easy question...

I can give you comments how I'm working with mesh..

1) first I choose adequate template (planar, free space and so on).
2) in most cases this mesh is quite rough (OK, you must decide by eyes..)

I estimate a limit of meshcells for given structure accounting RAM. Say 1 million of cells.

Then I go into mesh properties and play with settings.
Recommended is to have more than 15 cells per wavelength. Almost always I apply Specials --> Refinement --> Refine at pec... I use factor=3 for standard structures and 4 for planar.

When your structure have big differences in dimensions, you need to ensure good meshing at those small parts. This can be done by varying lower mesh limit and mesh line ratio limit. You need experiences, for example microstrip line should have 3-4 transversal lines.

Another way is to locally change mesh properties for defined part of structure, is inner conductor in connector.
To make this turn mesh on, select this part under layers and go to Edit-->Mesh properties. Then you can change Maximum mesh step width in all directions (i.e. from 0 to 0.5).

I think mesh wizard isn't very good since it only changes lines per wavelength and refine at pec edges factor.

Best way is to use your eyes

I'm sorry I think it wasn't very helpful..

Best,

Eirp