cst radiation efficiency
I'm facing problem with understanding radiation efficiency in MWS.
When simulating dish antenna with horn feeder, the radiation efficiency is around 55% only.
I think that the rad.eff is in MWS defined as Power in farfield/Power at input port
Thus I suspect the surface waves on parabola surface which decreases faster than 1/r and can't be accounted into calculation in far field.
From the Efield animation I'm quite sure I see some wave not reflecting but travelling along the parabola.
However, I'm true? Please tell me your opinion!
Thanks
eirp
Hi Eirp,
In MWS , the rad eff. is the ratio of the radiated to accepted (input) power of the antenna.
For a lossless structure, the rad eff. should be 1.
However, sometimes MWS calculates rad eff. for a lossles structure which is not 1. This might have several reasons:
1: The structure does (almost) not radiate at the freq. of the farfield monitor. In this case the accepted and the radiated Power are almost zero. Small errors might lead to a efficiency <> 1 because of the 0/0 effect.
2: Some PEC material is touching the boundary and you have some kind of surface wave travelling through the PML.
3: Inhomogeneous ports-> Some energy is absorbed by the ports but not included in the S-Parameter.
To name just a few.
Maybe you can post your design and we all can have a look?
Hope this helps,
F.
Another reason about the difference from the exact value (i.e. 1) of radiation efficency (or ratio Gain/Directivity) in a lossless structure is sometime related to abortion of a transient calculation at a too early stage.
In this case we have a meaning remaining energy inside the structure and that affects the correct evaluation of Gain (in fact the evaluation of this quantity is related also to the losses and a remaining energy is similar to a loss); then it is possible to have a radiation efficiency not equal to 1 also in a lossless structure.
Regards.
I saw two things in CST, radiation efficiency and total radiation efficiency. I think that for the radiation effiientcy, CST calculate for the stucture only no ohmic loss. e.g. an antenna that is not resonant, some power will be store in the imaginary part. and the total efficiency combine both. I am not sure about this.
Hi,
The total efficiency is defined as the ratio of radiated to stimulated power of the antenna:
-> Takes into account the S-Parameter (matching)
The radiation efficiency does not take into accout the S-Paramerter. Only the accepted Power counts.
F.
Hi, RFsimulator :) and all
Rading RFsimulator possible reasons:
ad 1) no
ad 2) no
ad 3) no
Now I'm experimenting with PML distance from the structure.
But nobody gave me clear explanation on surface waves..
Can one prove if their excitation is possible on parabolic dish antenna?
Tomorrow I'll pack my structure and post it here..
Thanks all for your interest!!
eirp
Hi eirp!
I think everything's OK, expected is about 0.6..
We have several sideeffects- aperature, aperature taper, phase error and spillover efficiencies.
Look about at: http://www.ece.rutgers.edu/~orfanidi/ewa/ch17.pdf , pgs from 583 to 586.
Please don't forget post your structure here.
PROBLEM SOLVED!!!
Coaxial port was too short (4 meshlines not enough..)
After extending the port length from 1cm to 5cm, radiation eff. is close to 1.
Thanks all for help!!
regards
eirp
Sorry, my mistake, 0.6 is for aperature efficiency :)
Hi!
I found another interesting problem when playing with feeder lenght!
It significantly affects even module of S_11 too.. (deep rather than resonance position)
To be more clear..
There's big difference in S_11 of coaxial feeder 10mm(~4 meshlines) -> 50mm (@2.45 GHz)
more extending 50mm -> 100mm don't play role
First, I'd like to thank to RFsimulator for his valuable comments via PMs :P
Now it's clear.. The problem is connected with subgridding introduced in ver. 5!!!
It predicted quite bad farfield pattern (caused strange backlobe) and of course low rad. eff.
My result: don't use new features until few service pack is released :)
kind regards
eirp
Hi Eirp,
can you explain (more details) what do you mean about "the problem is connected with subgridding introduced in ver. 5"?
Thank you in advance for your attention
Regards
Hi!
bellow please find my mail to Eirp.
.....
Hi Eirp,
I had a look at your structure. Please allow me to give my 50cents to your simulation.
The Farfield problem is most likely due to the fact that you use sub-griddig. I know from the CST support, that there is a known inaccuracy for the farfield efficiency calculation if you use sub gridding.
Its not really a bug but the farfield calculation needs some special method for the sub gridding case. This has not been implemented so far. Maybe you can check, I guess, you switched off the sub-gridding as you used a longer coax?
...
Don't get me wrong: I think the sub gridding is a very good feature but it is not suited for all problem types. I mainly use subgridding for WG devices with very small details and large metal parts. For this type of problem sub meshing helps you to save a lot of calculation time.
I normally do not use it for antenna problems (no small details, no large metal parts). If you use sub gridding for this type of problems, the calculation time might even go up, because you save only a very small number of nodes.
Hope this helps.
F.