In this report, we discussed several important issues to simulate MEMS devices: simulatability, suitability, efficiency and scale. Some or all those issue are also important to general simulation tasks. We presented how to tackle those issues into two successful simulations using Impulse: the linear electrostatic stepper motor and its articulated structure.
The experimental simulation presented in this report also suggests
that a general purpose dynamic simulator
such as Impulse is a powerful tool but may not be
appropriate for all kinds of MEMS
simulation due to the small dimension of devices and their motions
(typically in the 
range) and the high frequency of the
motion (1KHz for the linear electrostatic stepper motor).
The small dimension causes numerical problem in integration. In both the stepper motor and the level arm cases, we have to scale the dimensions up to get reasonable simulation. The high frequency causes under-sampling of the motion. For example, Impulse only generates 180 frames per second. Although it applies correct impulse to generate the correct ``net motion'', but details of the motion cannot be captured by the movie. To see this, just imagine an oscillation of 180 Hz. The movie will take a snapshot of the oscillation at almost the same displacement.
The experimental simulation presented in this report demonstrate the novel idea of dynamic simulation of MEMS devices. Using simulation, we can visualize the behavior of a design without solving any tedious equations. Using simulation, we can also easily study the effect of the change of parameters, such as change of frequency. We are encouraged to pursue dynamic simulation of MEMS structures in our future research. We will also improve the simulator based on the lessons we learned here.