Help CST MWS Warning Farfield is inaccurate

I'm simulating a antenna. Microstip S-shaped planar antenna.
Well, NO warnings in Trasient Simulation, but in Farfield appear this warning "Farfield is inaccurate!".
What is the problem?.
Thanks.
Version of CST MWS 4.3.

I can't see your model, but
1) Boundary conditions are too near
2) If you use the AR filter then s11 is good, but too much energy left in the system for a FF transform
3) Far field monitor size is too small and cuts the structure or port.
HTH

Hi,
- How much we suppose to add as space in the boundary ?
- How to use the AR filter? and what's the physical meaning of that?
-

The AR filter is basically a response predictor in the time domain. All this does is recognisae the damping and the response adn extrapolates so that when the FFT is performed there is no time domain truncation error (ripple on S parameters).
If you stop simulationg with too much energy in the domain, the AR will be OK, BUT the near to far tx will be incorrect.
If you use open add space boundaries it will calculate the correct distance

Sorry,
This AR-filter is something very new for me. Is there any links to read more about it and when to use it ?
There is no ripples in the above example. Som why to use AR-filter then ?
We have inaccurate far field results, could AR-filter solve this problem ?
Regards