Consider starting with the modal analysis only and completing it before you move to a vibration analysis. Ask your self:
1. What frequency domain does the structure operate in (requirement) 5-500 Hz, 20 - 2000 Hz, etc?
2. Run your modal analysis and understand those results (ie) what are the mode shapes of the structure?
3. What are your dominant modes (ie) mass particpatation factor?
Once you are comfortable with the results of the modal analysis then and only then should your consider the methodology associated with the vibration analysis. Post the results of your modal analysis with pics or an avi file and you will probably get more feed back.
Hope this helps. Good luck.
From my experience, when doing modal analysis, the only pertinent data is the frequency. The displacements are arbitrary in a modal analysis, they simply show you the shape of the mode. Displacements are typically scaled to 10 mm or 1 inch in a modal analysis and they mean nothing other than to drastically warp the model to show you how it would react to a given frequency. Same goes for stress. As you could imagine if you had displacement that was 1 inch or 10 mm....your stresses would be through the roof. Once you have the frequencies you want to analyze...you go back and start a fresh simulation. This time instead of selecting frequency analysis...you select a linear dynamic analysis. In linear dynamic you're going to want to set up a forcing function at the desired frequency. In this linear dynamic study (assuming you have your boundary conditions, forcing function, mesh, and constraints established correctly), you can analyze the displacements and stresses of a real-world vibration.