change in axial ratio on changing frequency sweep

i simulate the circular polarize patch antenna at 2.45 GHz and frequency sweep is 2.2 to 2.6 GHz i get good result on fast sweep but when i change the frequency sweep from 2.2 to 6 GHz my results suddenly change. my axial ratio changes from 0.74db to 5.6db,s11 from -23db to -18db and gain from 1.5db to 2.06db.
please tell me why this happen , tried this in descrite sweep but my result got change.

Thanks

The discrete sweep will always be more accurate. If you still get two different data for separate discrete sweeps, your simulation is likely not converged enough.

i solve this model on both driven model and driven terminal. on increasing sweep i still get two different data , it dosen't show me any warning that your solution is not converge.
so please tell me what i do.
Thanks

Increase convergence. Just because your simulation converges, doesn't mean it is correct.

I would also stick with driven modal until you sort this issue out.