Discrete port or waveguide port ?

I am dealing with an antenna which is fed by a tranmission line.
First, I used a waveguide port with a standard size, that is guided in the manual.
Then, I have tried to use discrete port .
After that, I compared two these simulation results and found that:
1/ If you don't check the FIX 50 OHM in the transient solver setup, the results will be different and vary each simualtion case.
2/ If you check this option, or use a renormalized S-parameter with 50 Ohm, the results will be same, but it is still different with the Dsicrete port case.
In both cases, waveguide and discrete port, the efficiency is almost 100%

I am wondering that which case (waveguide port or discrete port) is more reality?
Actually, each port is always defined properly, however, the results is very differents! And even though the normalization S-parameter for both case is 50 Ohm!

(In HFSS, the port type is always discrete port, but in MWS, we have many choices.)

Please give me a trade off for my case!
Thank you!

if you are using CST MWS then you have to option for feeding the patch antenna :

1: discrete port
2: waveguide port

if you are interested in radiation pattern of the antenna then use discrete port,
the s parameters result given by discrete port are not accurate.

if you are interested in accurate s paramters then u must use waveguide port.

i hope this helps you

thank you shahid78
another problem is the normalization of S-parameter
Do I always have to fix the 50-ohm normalized S-parameter in the Transient Solver setup?
When we use the network Analyser to measure the return loss of the antenna, the input impedance of the coaxial cable connected to the antenna is always 50-ohm, am I right?

yes if you want close agreement between mearsued and calculated s parameters then you must use the same impedance for normalizing.
if your network analyzer cable impedance is 50 ohms which usually is then you have to normalize the s parameter using 50 ohm in transient solver as well

hope this helps

regards

I am using the 2006B version of MWS in which there is a subgridding in the meshing generation.
That 's mean the time consuming will be reduced significantly if we use this new option.
However, if I use this option in addition with the waveguide port, the input EM is decayed very fast, and the EM wave is almost no propagating.
Then, I try to un-check this option, the results is very good.

I have also tried with some other excitation source by changing the lowest frequency , but the problem is same.

Do you know what this problem is?

Thank you!