Problem with energy in CST for 24 GHz dual reflector antenna.

Hello there,
I'm trying to model dual-reflector antenna with eleptical sub-reflector. All dimensions set parametrically. Feed structure is a cutting of circular waveguide with waveguide port in start. I use Time domain solver.
At first I set material to PEC, and main lobe was ~34dB. All was good except error "The following port mode is at least partially excited below its cutoff frequency: 1(1)". I think it's associated with the choice of frequency range 0-25 GHz and dimension of circular waveguide. It is right?
But then, I have changed material to aluminium, but I've made a mistake (mixed up "Aluminium" and "Aluminia") and set aluminia. Main lobe went down to ~12 dB.
Finally, I set material back to PEC or aluminium (it doesn't important), but now energy remains constant = -0.025 dB and now I've an error "Maximum simulation time reached, solver stopped. Please note that the steady state energy criterion has not been satisfied". At the same time directional pattern have an expected shape (Main lobe ~34dB). Can trust this results? How I can fix energy criterion (I'm increasing "number of impulses" from 20 to 100 and it did not help)?
Thank you.

Problem was caused by mistake of frequency range for solver. I set narrow range(23-25 GHz), it's helped.