I like Jim's post here with regard to your question - in particular 400 lines of resolution being applied - definitely more than enough resolution and huge time saving.
For baseplates and larger surfaces I have use a small fractional hp variable rpm d/c motor. I have a fly-wheel type rotor fitted to the shaft with a known amount of unbalance weight to create vibration.
I place the motor in position - secure it and begin tuning (via attached rheostat controller) through different rpm ranges while measuring in the different locations and in each axis - don't be surprised to find at times only one direction is an issue. Tracking the rpm of the motor (forcing frequency)can be done via a photo tach or strobe light. The flywheel attachment I made has drilled and tapped holes equally spaced around the diameter for the addition of extra weight if needed. Of course this is dependent on the amount of mass involved as to exciting any natural frequency and in some cases the old "dead blow" sledge hammer is required.
I find this way I get consistent force vs. using a hammer or other method to excite natural frequencies. I have used this for identifying natural frequency of objects and for nodal testing on structures and piping. Word of caution is to be sure the unit is properly secured to whatever you are testing - in particular piping. If you question the safety of your mounting - go back to your "dead-blow" hammer process. If you like I can post a pic of the unit I put together quite some time ago.