Hi loke,
can you post your design here? Again, the transient solver works best for broadband structures. If you have a narrow band filter the time signal for the broadband excitation takes a long time to go down.
Before CST "upgrade" the frequency solver (before version 5) the AR Filter was used as a possibility to shorten the calculation time somewhat.
However, for filters with many pols the AR Filter has its limitations as well.
The FD or Fast Resonate solver are a much better bet for this structures. If you can post the design here I can just give it a try. I have the FD solver available.
F.