Frequency simulation: applying roller fixture

[triumphengr]

I am attempting to run a simulation that consists of the following: a heat exchanger mounted to a vibration test fixture, which is mounted by bolts to a flat vibration test stand.

 

 

I applied fixed geometry to the bolt holes, and a mass load to where the heat exchanger is to be mounted.

 

When I run the simulation, I get low resonant frequencies. I believe this is because the areas with the tower pieces are displacing a large amount. I would think that adding a roller would model the effects of the flat test stand which the fixture is mounted to, as it would be preventing the fixture deflecting downwards, but the roller in solidworks prevents upwards deflection as well. I am wondering if anyone has an idea to constrain the fixture to prevent downwards deflection on the bottom of the fixture, without preventing upwards deflection.