size of air box for uwb antenna with notched bands

i am trying to simulate an ultra wide band(3.1 GHz-10.6 GHz) antenna with notched bands centered at 3.6 GHz and 5.5 GHz as per an IEEE paper.could anyone please tell me the size of radiation box and the solution frequency that i should use .i have tried a separation (between substrate and radiating surface)of quarter wavelength corresponding to the center frequency and lowest frequency,but couldn't obtain the same results as mentioned in the paper.i found the notch frequencies and their VSWR changes with size of radiation box and solution frequency. kindly help me with this
with regards
aswathy

You need to be clearer about
i) which type of antenna you are using
ii) which software you are using.

If you do not mind the excessive simulation time, you can set the solution frequency to the highest frequency and size of radiation box to be lambda/4 at the lowest frequency.

The answer would also depend on the type of sweep you carry out to determine the performance, again coming back to point ii.

sir
I am using a printed monopole antenna ( patch antenna with reduced ground plane).I am simulating in HFSS version 14 and I have used fast sweep.
thanking you

In HFSS, it is sufficient to set your air box to be at a distance of lambda/4 ( at the lowest operating frequency ) away from the radiating surface

I would make the following suggestions to get accurate results:
1. Set your solution frequency to the highest frequency you want to solve for.
2. Set your sweep to discrete and create a list of the frequencies you want to solve for.
3. How are you exciting the antenna ? Instead of a lumped port, you can include a model of the SMA connector/feed pin and excite the other end of the SMA cable with a wave port.

The *most accurate* way of solving this problem is to use the HFSS-IE solver in conjunction with the FEM solver to perfectly truncate the air region surrounding your antenna. This is the so-called FEBI method (Finite Element-Boundary Integral). Using the IE solver to truncate the region affords the most accuracy with the smallest air region. The mechanics of this are to simply assign the Radiation BC to the airbox, and then enable the option to "Model Exterior as HFSS-IE domain."

I would also recommend against using the Fast Frequency Sweep for these broadband antennas. An Interpolating sweep for the S-parameters is best, and then a Discrete sweep for the individual points where you need fields is usually sufficient. There is also a new option in the Discrete sweep to save only radiating fields. This can help conserve disk space for large problems.

I have tried an interpolating sweep together with FEBI used a solution frequency of 11GHz .still I couldn't achieve the notches at center frequencies specified in the paper .
kindly help me
thanking you