Group delay of UWB antenna using HFSS or CST

Hi all,
I am researching about UWB antenna with band-notched characteristic. My question regards simulating an antenna and calculating the group delay/group delay variation from the simulation results.
How do I do to obtain exactly group delay graph? (from S21? or S11?)
If from S21, I must build 2 identical antennas and simulate?
So, How much the distance should be between these two identical antennas?
If possible, can you give me your project about this.
Please help me!
Thanks so much!

hi
yes you must locate 2 identical antennas in farfield of each other ( if you have a conventional antenna 30-50cm is sufficent) you must calculate phase of S21 betwenn them.
good luck.

Thank myazdi83!
I had simulated two identical antennas with the distance between antennas about 30 cm. Then, I took directly result from GroupDelay(p2,p1). However, I am wondering about this result, it is correctly or not?
Let me know the how to take correctly the Groupdelay result.
Thanks so much!

Hi, All of them.
I saw this board, and I realize everybody confuse about the group delay.
the group delay from S11 and S12 is used based on the application.
If you are radar engineer, you need the group delay from S12. Sometimes, from S11 (mono case)
If you are antenna desinger, you need show the group delay from S11.
the group delay from S12 is varied in terms of the environment, distance, angle, and so on.
Therefore, Even I usually concern the group delay from S12 due to my project in my lab, I first look at the group delay from S11. It is the reference.
Please, don't confuse the group delay.

hi
it is not true. can you use CST?
if yes you can find group delay in curve parameters. if you have not CST you can derivate from phase of s21.
good luck.
myazdi83@gmail.com

Putting a frequency notch in an antenna will also put a big group delay variation in the passband at the notch frequency. Can an electromagnetic simulator program accurately predict the group delay? Maybe. You have to use enough frequency steps around the notch frequency for the equations to converge. You could run a different manufacturer's emag program and see if you get similar results. Or you could build the antennas and do an actual measurement, which is really the only way to know!

hi biff44
yes. CST can perdict group delay in your case.

myazdi83, I think he has the prediction, but it looks whacky and he is wondering just how often cst actually works right! The only way to really tell is the verify with actual measurements!